7. ROM MSX-BASIC INTERPRETER

    Microsoft BASIC has evolved over the years to its present position as the industry standard. It was originally written for the 8080 Microprocessor and even the MSX version is held in 8080 Assembly Language form. This process of continuous development means that there are less Z80-specific instructions than would be expected in a more modern program. It also means that numerous changes have been made and the result is a rather convoluted program. The structure of the Interpreter makes it unlikely that an application program will be able to use its many facilities. However most programs will need to cooperate with it to some extent so this chapter gives a detailed description of its operation.

    There are four readily identifiable areas of importance within the Interpreter, the one most familiar to any user is the Mainloop (4134H). This collects numbered lines of text from the console and places them in order in the Program Text Area of memory until a direct statement is received.

    The Runloop (NEWSTT, 4601H) is responsible for the execution of a program. It examines the first token of each program line and calls the appropriate routine to process the remainder of the statement. This continues until no more program text remains, control then returns to the Mainloop.

    The analysis of numeric or string operands within a statement is performed by the Expression Evaluator (FRMEVL, 4C64H). Each expression is composed of factors, in turn analyzed by the Factor Evaluator (4DC7H), which are linked together by dyadic infix operators. As there are several types of operand, notably line numbers, which cannot form part of an expression in Microsoft BASIC the term "evaluated" is only used to refer to those that can. Otherwise a term such as "computed" will be used.

    One point to note when examining the Interpreter in detail is that it contains a lot of trick code. The writers seem particularly fond of jumping into the middle of instructions to provide multiple entry points to a routine. As an example take the instruction:

3E D1

Normal: LD A,0D1H

    When encountered in the usual way this will of course load the accumulator with the value D1H. However if it is entered at "Normal" then it will be executed as a POP DE instruction. The Interpreter has many similarly obscure sections.

    The mathematical core of the MSX-BASIC is called "MSX Math-Pack" and is composed of several arithmetic, trigonometric and logarithmic functions, which work on all MSX machines and turbo R. Its usefulness for Assembly programming, though, is severely spoiled because these functions operate on BCD numeric representation, wasting numeric range (that is, precision) in favour of some easiness to the BASIC interpreter, which uses it. The following entries are part of the MSX Math-Pack, which can be used like a BIOS:

    Address ... 268CH (MSX Math-Pack)
    Name ...... DECSUB (Decimal Subtraction)
    Input ..... DAC=Minuend, ARG=Subtrahend, VALTYP=DAC numeric type
    Exit ...... DAC=DAC-ARG
    Modifies .. All registers

    This routine is used by the Expression Evaluator to subtract two double precision operands. The first operand is contained in DAC and the second in ARG, the result is returned in DAC. The second operand's mantissa sign is inverted and control drops into the addition routine.

    Address ... 269AH (MSX Math-Pack)
    Name ...... DECADD (Decimal Addition)
    Input ..... DAC=First Parcel, ARG=Second Parcel, VALTYP=DAC numeric type
    Exit ...... DAC=DAC+ARG
    Modifies .. All registers

    This routine is used by the Expression Evaluator to add two double precision operands. The first operand is contained in DAC and the second in ARG, the result is returned in DAC. If the second operand is zero the routine terminates with no action, if the first operand is zero the second operand is copied to DAC (2F05H) and the routine terminates. The two exponents are compared, if they differ by more than 10^15 the routine terminates with the larger operand as the result. Otherwise the difference between the two exponents is used to align the mantissae by shifting the smaller one rightwards (27A3H), for example:

19.2100 = .1921*10^2 = .192100
+ .7436 = .7436*10^0 = .007436

    If the two mantissa signs are equal the mantissae are then added (2759H), if they are different the mantissae are subtracted (276BH). The exponent of the result is simply the larger of the two original exponents. If an overflow was produced by addition the result mantissa is shifted right one digit (27DBH) and the exponent incremented. If leading zeroes were produced by subtraction the result mantissa is renormalized by shifting left (2797H). The guard byte is then examined and the result rounded up if the fifteenth digit is equal to or greater than five.

    Address ... 26FAH (MSX Math-Pack)
    Name ...... DECNRM (Decimal Normalize)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=DAC normilized
    Modifies .. All registers

    This function formats the number to a "normal" standard. Examples:

0.0000123	⇒	0.123e-4
123	⇒	0.123e+3
-0.0123e+6	⇒	-0.123e+5
-123e-3	⇒	-0.123e+0

    Address ... 273CH (MSX Math-Pack)
    Name ...... DECROU (Decimal Round)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=DAC rounded
    Modifies .. All registers

    This function rounds numbers with more digits than the type allows. If the next digit after the last significant digit for the given precision ranges from 0 till 4, it is simply discarded, else it is discarded and the absolute value of the last significant digit, incremented.

1.23456789012345	⇒	1.2345678901235
1.23456789012344	⇒	1.2345678901234
-1.11111111111115	⇒	-1.1111111111112
-1.99999999999999	⇒	-2.0000000000000

Address

@ (MSX only)

    This routine adds the two double precision mantissae contained in DAC and ARG and returns the result in DAC. Addition commences at the least significant positions, DAC+7 and ARG+7, and proceeds two digits at a time for the seven bytes. Address ... 276BH

Address

@ (MSX only)

    This routine subtracts the two double precision mantissae contained in DAC and ARG and returns the result in DAC. Subtraction commences at the guard bytes, DAC+8 and ARG+8, and proceeds two digits at a time for the eight bytes. If the result underflows it is corrected by subtracting it from zero and inverting the mantissa sign, for example:

0.17-0.85 = 0.32 = -0.68

Address

@ (MSX only)

    This routine shifts the double precision mantissa contained in DAC one digit left. Address ... 27A3H

Address

@ (MSX only)

    This routine shifts a double precision mantissa right. The number of digits to shift is supplied in register A, the address of the mantissa's most significant byte is supplied in register pair HL. The digit count is first divided by two to separate the byte and digit counts. The required number of whole bytes are then shifted right and the most significant bytes zeroed. If an odd number of digits was specified the mantissa is then shifted a further one digit right.

    Address ... 27E6H (MSX Math-Pack)
    Name ...... DECMUL (Decimal Multiplication)
    Input ..... DAC=Multiplicand, ARG=Multiplier, VALTYP=DAC numeric type
    Exit ...... DAC=DAC*ARG (Product)
    Modifies .. All registers

    This routine is used by the Expression Evaluator to multiply two double precision operands. The first operand (multiplicand) is contained in DAC and the second (multiplier) in ARG, the result (product) is returned in DAC. If either operand is zero the routine terminates with a zero result (2E7DH). Otherwise the two exponents are added to produce the result exponent. If this is smaller than 10^-63 the routine terminates with a zero result, if it is greater than 10^63 an "Overflow error" is generated (4067H). The two mantissa signs are then processed to yield the sign of the result, if they are the same the result is positive, if they differ it is negative.

    Even though the mantissae are in BCD format they are multiplied using the normal binary add and shift method. To accomplish this the first operand is successively multiplied by two (288AH) to produce the constants X*80, X*40, X*20, X*10, X*8, X*4, X*2, and X in the HOLD8 buffer. The second operand remains in ARG and DAC is zeroed to function as the product accumulator. Multiplication proceeds by taking successive pairs of digits from the second operand starting with the least significant pair. For each 1 bit in the digit pair the appropriate multiple of the first operand is added to the product. As an example the single multiplication 1823*96 would produce:

1823*10010110=(1823*80)+(1823*10)+(1823*4)+(1823*2)

    As each digit pair is completed the product is shifted two digits right. When all seven digit pairs have been processed the routine terminates by renormalizing and rounding up the product (26FAH).

    The time required for a multiplication depends largely upon the number of 1 bits in the second operand. The worst case, when all the digits are sevens, can take up to 11ms compared to the average of approximately 7ms. - 91 - 5. ROM BASIC INTERPRETER Address ... 288AH

Address

@ (MSX only)

    This routine doubles a double precision mantissa three successive times to produce the products X*2, X*4 and X*8. The address of the mantissa's least significant byte is supplied in register pair DE. The products are stored at successively lower addresses commencing immediately below the operand.

    Address ... 289FH (MSX Math-Pack)
    Name ...... DECMUL (Decimal Division)
    Input ..... DAC=Dividend, ARG=Divisor, VALTYP=DAC numeric type
    Exit ...... DAC=DAC÷ARG
    Modifies .. All registers

    This routine is used by the Expression Evaluator to divide two double precision operands. The first operand is contained in DAC and the second in ARG, the result is returned in DAC. If the first operand is zero the routine terminates with a zero result if the second operand is zero a "Division by zero" error is generated (4058H). Otherwise the two exponents are subtracted to produce the result exponent and the two mantissa signs processed to yield the sign of the result. If they are the same the result is positive, if they differ it is negative.

    The mantissae are divided using the normal long division method. The second operand is repeatedly subtracted from the first until underflow to produce a single digit of the result. The second operand is then added back to restore the remainder (2761H), the digit is stored in HOLD and the first operand is shifted one digit left. When the first operand has been completely shifted out the result is copied from HOLD to DAC then renormalized and rounded up (2883H). The time required for a division reaches a maximum of approximately 25 ms when the first operand is composed largely of nines and the second operand of ones. This will require the greatest number of subtractions.

    Address ... 2993H (MSX Math-Pack)
    Name ...... COS (Cosine)
    Input ..... DAC=Angle, VALTYP=DAC numeric type
    Exit ...... DAC=Cosine(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "COS" function to a double precision operand contained in DAC. The operand is first multiplied (2C3BH) by 1/(2*PI) so that unity corresponds to a complete 360 degree cycle. The operand then has 0.25 (90 degrees) subtracted (2C32H), its mantissa sign is inverted (2E8DH) and control drops into the "SIN" routine.

    Address ... 29ACH (MSX Math-Pack)
    Name ...... SIN (Sine)
    Input ..... DAC=Angle, VALTYP=DAC numeric type
    Exit ...... DAC=Sine(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "SIN" function to a double precision operand contained in DAC. The operand is first multiplied (2C3BH) by 1/(2*PI) so that unity corresponds to a complete 360 degree cycle. As the function is periodic only the fractional part of the operand is now required. This is extracted by pushing the operand (2CCCH) obtaining the integer part (30CFH) and copying it to ARG (2C4DH), popping the whole operand to DAC (2CE1H) and then subtracting the integer part (268CH). - 92 - 5. ROM BASIC INTERPRETER

    The first digit of the mantissa is then examined to determine the operand's quadrant. If it is in the first quadrant it is unchanged. If it is in the second quadrant it is subtracted from 0.5 (180 degrees) to reflect it about the Y axis. If it is in the third quadrant it is subtracted from 0.5 (180 degrees) to reflect it about the X axis. If it is in the fourth quadrant 1.0 (360 degrees) is subtracted to reflect it about both axes. The function is then computed by polynomial approximation (2C88H) using the list of coefficients at 2DEFH. These are the first eight terms in the Taylor series X-(X^3/3!)+(X^5/5!)-(X^7/7!) ... with the coefficients multiplied by successive factors of 2*PI to compensate for the initial scaling.

    Address ... 29FBH (MSX Math-Pack)
    Name ...... TAN (Tangent)
    Input ..... DAC=Angle, VALTYP=DAC numeric type
    Exit ...... DAC=Tangent(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "TAN" function to a double precision operand contained in DAC. The function is computed using the trigonometric identity TAN(X) = SIN(X)/COS(X).

    Address ... 2A14H (MSX Math-Pack)
    Name ...... ATN (Arc-tangent)
    Input ..... DAC=Tangent, VALTYP=DAC numeric type
    Exit ...... DAC=Arc-tangent(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "ATN" function to a double precision operand contained in DAC. The function is computed by polynomial approximation (2C88H) using the list of coefficients at 2E30H. These are the first eight terms in the Taylor series X-(x^3/3)+(X^5/5)-(X^7/7) ... with the coefficients modified slightly to telescope the series.

    Address ... 2A72H (MSX Math-Pack)
    Name ...... LOG (Logarithm Natural)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=Logarithme(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "LOG" function to a double precision operand contained in DAC. The function is computed by polynomial approximation using the list of coefficients at 2DA5H.

    Address ... 2AFFH (MSX Math-Pack)
    Name ...... SQR (Square Root)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=Square root(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "SQR" function to a double precision operand contained in DAC. The function is computed using the Newton-Raphson process, an equivalent BASIC program is:

10 INPUT"NUMBER";X
20 GUESS=10
30 FOR N=1 To 7
40 GUESS=(GUESS+X/GUESS)/2
50 NEXT N
60 PRINT GUESS
70 PRINT SQR(X)


    The above program uses a fixed initial guess. While this is accurate over a limited range maximum accuracy will only be attained if the initial guess is near the root. The method used by the ROM is to halve the exponent, with rounding up, and then to divide the first two digits of the operand by four and increment the first digit.

    Address ... 2B4AH (MSX Math-Pack)
    Name ...... EXP (Exponential)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=eDAC
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "EXP" function to a double precision operand contained in DAC. The operand is first multiplied by 0.4342944819, which is LOG(e) to Base 10, so that the problem becomes computing 10^X rather than e^X. This results in considerable simplification as the integer part can be dealt with easily. The function is then computed by polynomial approximation using the list of coefficients at 2D6BH.

    Address ... 2BDFH (MSX Math-Pack)
    Name ...... RND (Randomic number)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... if DAC=0, then DAC=RNDX
                if DAC<0, then RNDX=DAC
                if DAC>0, then DAC=RNDX, RNDX=pseudo-random number
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "RND" function to a double precision operand contained in DAC. If the operand is zero the current random number is copied to DAC from RNDX and the routine terminates. If the operand is negative it is copied to RNDX to set the current random number. The new random number is produced by copying RNDX to HOLD, the constant at 2CF9H to ARG, the constant at 2CF1H to DAC and then multiplying (282EH). The fourteen least significant digits of the double length product are copied to RNDX to form the mantissa of the new random number. The exponent byte in DAC is set to 10^0 to return a value in the range 0 to 1. Address ... 2C24H

Address

@ (MSX only)

    This routine is used by the "NEW", "CLEAR" and "RUN" statement handlers to initialize RNDX with the constant at 2D01H. Address ... 2C2CH

Address

@ (MSX only)

    This routine adds the constant whose address is supplied in register pair HL to the double precision operand contained in DAC. Address ... 2C32H

Address

@ (MSX only)

    This routine subtracts the constant whose address is supplied in register pair HL from the double precision operand contained in DAC. Address ... 2C3BH

Address

@ (MSX only)

    This routine multiplies the double precision operand contained in DAC by the constant whose address is supplied in register pair HL. - 94 - 5. ROM BASIC INTERPRETER Address ... 2C41H

Address

@ (MSX only)

    This routine divides the double precision operand contained in DAC by the constant whose address is supplied in register pair HL. Address ... 2C47H

Address

@ (MSX only)

    This routine performs the relation operation on the double precision operand contained in DAC and the constant whose address is supplied in register pair HL.

    Address ... 2C4DH (MSX Math-Pack)
    Name ...... MAF (Move to Argument From DAC)
    Input ..... DAC
    Exit ...... ARG=DAC
    Modifies .. AF, B, DE, HL

    This routine copies an eight byte double precision operand from DAC to ARG. Absolutely no check about the contents of DAC is made, this routine simply copies eight bytes from it to ARG, regardless of the datum type.

    Address ... 2C50H (MSX Math-Pack)
    Name ...... MAM (Move to Argument from Memory)
    Input ..... HL=pointer to memory
    Exit ...... ARG=[HL] (8)
    Modifies .. AF, B, DE, HL

    This routine copies eight bytes pointed by HL to ARG. Absolutely no check about the contents of the memory is made, this routine simply copies eight bytes from it to ARG, sequentially.

    Address ... 2C53H (MSX Math-Pack)
    Name ...... MOV8DH (Move 8 bytes to [DE] from [HL])
    Input ..... HL=pointer to memory source, DE=pointer to memory destiny
    Exit ...... [DE]=[HL] (8)
    Modifies .. AF, B, DE, HL

    This routine copies eight bytes pointed by HL to memory pointed by DE.

    Address ... 2C59H (MSX Math-Pack)
    Name ...... MFA (Move to DAC From Argument)
    Input ..... ARG=Operand
    Exit ...... DAC=ARG
    Modifies .. AF, B, DE, HL

    This routine copies an eight byte double precision operand from ARG to DAC. Absolutely no check about the contents of ARG is made, this routine simply copies eight bytes from it to DAC, regardless of the datum type.

    Address ... 2C5CH (MSX Math-Pack)
    Name ...... MFM (Move to DAC From Memory)
    Input ..... HL=Pointer to memory
    Exit ...... DAC=[HL] (8)
    Modifies .. AF, B, DE, HL

    This routine copies eight bytes pointed by HL to ARG. Absolutely no check about the contents of the memory is made, this routine simply copies eight bytes from it to DAC, sequentially.

    Address ... 2C67H (MSX Math-Pack)
    Name ...... MMF (Move to Memory From DAC)
    Input ..... HL=Pointer to memory
    Exit ...... [HL]=DAC (8)
    Modifies .. AF, B, DE, HL

    This routine copies eight bytes pointed by HL to ARG. Absolutely no check about the contents of the memory is made, this routine simply copies eight bytes from it to DAC, sequentially.

    Address ... 2C6AH (MSX Math-Pack)
    Name ...... MOV8HD (Move 8 bytes to [HL] from [DE])
    Input ..... HL=pointer to memory destiny, DE=pointer to memory source
    Exit ...... [HL]=[DE] (8)
    Modifies .. AF, B, DE, HL

    This routine copies eight bytes pointed by DE to memory pointed by HL. Address ... 2C6FH

Address

@ (MSX only)

    This routine exchanges the eight bytes in DAC with the eight bytes currently on the bottom of the Z80 stack. Address ... 2C80H

Address

@ (MSX only)

    This routine inverts the mantissa sign of the operand contained in DAC (2E8DH). The same address is then pushed onto the stack to restore the sign when the caller terminates. Address ... 2C88H

Address

@ (MSX only)

    This routine generates an odd series based on the double precision operand contained in DAC. The series is of the form:

        X^1*(Kn)+X^3*(Kn-1)+x^5*(Kn-2)+X^5*(Kn-3) ...

    The address of the coefficient list is supplied in register pair HL. The first byte of the list contains the coefficient count, the double precision coefficients follow with K1 first and Kn last. The even series is generated (2C9AH) and multiplied (27E6H) by the original operand. Address ... 2C9AH

Address

@ (MSX only)

    This routine generates an even series based on the double precision operand contained in DAC. The series is of the form:

        X^0*(Kn)+x^2*(Kn-1)+x^4*(Kn-2)+x^6*(Kn-3) ...

    The address of the coefficient list is supplied in register pair HL. The first byte of the list contains the coefficient count, the double precision coefficients follow with K1 first and Kn last. The method used to compute the polynomial is known as Horner's method. It only requires one multiplication and one addition per term, the BASIC equivalent is:

10 X=X*X
20 PRODUCT=0
30 RESTORE 100
40 READ COUNT
50 FOR N=1 TO COUNT
60 READ K
70 PRODUCT= ( PRODUCT*X ) +K
80 NEXT N
90 END
100 DATA 8
110 DATA Kn-7
120 DATA Kn-6
130 DATA Kn-5
140 DATA Kn-4
150 DATA Kn-3
160 DATA Kn-2
170 DATA Kn-1
180 DATA Kn


    The polynomial is processed from the final coefficient through to the first coefficient so that the partial product can be used to save unnecessary operations. Address ... 2CC7H

Address

@ (MSX only)

    This routine pushes an eight byte double precision operand from ARG onto the Z80 stack. Address ... 2CCCH

Address

@ (MSX only)

    This routine pushes an eight byte double precision operand from DAC onto the Z80 stack. Address ... 2CDCH

Address

@ (MSX only)

    This routine pops an eight byte double precision operand from the Z80 stack into ARG. Address ... 2CE1H

Address

@ (MSX only)

    This routine pops an eight byte double precision operand from the Z80 stack into DAC. Address ... 2CF1H

Address

@ (MSX only)

    This table contains the double precision constants used by the math routines. The first three constants have zero in the exponent position as they are in a special intermediate form used by the random number generator.

Address		Constant		Function		Address		Constant		Function
2CF1H		.14389820420821		RND		2DAEH		6.2503651127908
2CF9H		.21132486540519		RND		2DB6H		-13.682370241503
2D01H		.40649651372358				2DBEH		8.5167319872389
2D09H		.43429448190324		LOGe		2DC6H		5		LOG
2D11H		.50000000000000				2DC7H		1.0000000000000
2D13H		.00000000000000				2DCFH		-13.210478350156
2D1BH		1.0000000000000				2DD7H		47.925256043873
2D23H		.25000000000000				2DDFH		-64.906682740943
2D2BH		3.1622776601684		SQR(10)		2DE7H		29.415750172323
2D33H		.86858896380650		2LOGe		2DEFH		8		SIN
2D3BH		2.3025850929940		1/LOGe		2DF0H		-.69215692291809
2D43H		1.5707963267949		π/2		2DF8H		3.8172886385771
2D4BH		.26794919243112		TAN(π/12)		2E00H		-15.094499474801
2D53H		1.7320508075689		TAN(π/3)		2E08H		42.058689667355
2D5BH		.52359877559830		π/6		2E10H		-76.705859683291
2D63H		.15915494309190		1/(2π)		2E18H		81.605249275513
2D6BH		4		EXP		2E20H		-41.341702240398
2D6CH		1.0000000000000				2E28H		6.2831853071796
2D74H		159.37415236031				2E30H		8		ATN
2D7CH		2709.3169408516				2E31H		-.05208693904000
2D84H		4497.6335574058				2E39H		.07530714913480
2D8CH		3		EXP		2E41H		-.09081343224705
2D8DH		18.312360159275				2E49H		.11110794184029
2D95H		831.40672129371				2E51H		-.14285708554884
2D9DH		5178.0919915162				2E59H		.19999999948967
2DA5H		4		LOG		2E61H		-.33333333333160
2DA6H		-.71433382153226				2E69H		1.0000000000000

    Address ... 2E71H (MSX Math-Pack)
    Name ...... SIGN (Sign)
    Input ..... DAC=number, VALTYP=DAC numeric type
    Exit ...... if DAC is Zero, then A=00H, flag Z, NC
               if DAC is Positive, then A=01H, flag NZ, NC
               if DAC is Negative, then A=FFH, flag NZ, C
    Modifies .. A, F

    This routine returns the mantissa sign of a Floating Point operand contained in DAC. The exponent byte is tested and the result returned in register A and the flags. Address ... 2E7DH

Address

@ (MSX only)

    This routine simply zeroes the exponent byte in DAC.

    Address ... 2E82H (MSX Math-Pack)
    Name ...... ABSFN (Absolute Function)
    Input ..... DAC=number, VALTYP=DAC numeric type
    Exit ...... DAC=absolute(DAC)
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "ABS" function to an operand contained in DAC. The operand's sign is first checked (2EA1H), if it is positive the routine simply terminates. The operand's type is then checked via the GETYPR standard routine. If it is a string a "Type mismatch" error is generated (406DH). If it is an integer it is negated (322BH). If it is a double precision or single precision operand the mantissa sign bit in DAC is inverted.

    Address ... 2E8DH (MSX Math-Pack)
    Name ...... NEG (Negative)
    Input ..... DAC=floating-point number
    Exit ...... DAC=-DAC
    Modifies .. AF, HL

    This routine changes the signal of Floating Point numbers. It does not check the DAC datum type in VALTYP, it assumes the number to be a floating-point, returns with no action if the value is zero, else changes the signal of the number by inverting the bit 7 of the first byte in DAC, which contains the number's signal.

    Address ... 2E97H (MSX Math-Pack)
    Name ...... SGN (Signal)
    Input ..... DAC=number, VALTYP=DAC numeric type
    Exit ...... if DAC is Zero, then DAC=0000H, VALTYP=2 (integer)
               if DAC is Positive, then DAC=0001H, VALTYP=2 (integer)
               if DAC is Negative, then DAC=FFFFH, VALTYP=2 (integer)
    Modifies .. AF, HL

    This routine is used by the Factor Evaluator to apply the "SGN" function to an operand contained in DAC. The operand's sign is checked (2EA1H), extended into register pair HL and then placed in DAC as an integer. Address ... 2EA1H

Address

@ (MSX only)

    This routine returns the sign of an operand contained in DAC. The operands type is first checked via the GETYPR standard routine. If it is a string a "Type mismatch" error is generated (406DH). If it is a single precision or double precision operand the mantissa sign is examined (2E71H). If it is an integer its value is taken from DAC+2 and translated into the flags shown at 2E71H. Address ... 2EB1H

Address

@ (MSX only)

    This routine pushes a four byte single precision operand from DAC onto the Z80 stack. Address ... 2EC1H

Address

@ (MSX only)

    This routine copies the contents of registers C, B, E and D to DAC. Address ... 2ECCH

Address

@ (MSX only)

    This routine copies the contents of DAC to registers C, B, E and D. Address ... 2ED6H

Address

@ (MSX only)

    This routine loads registers C, B, E and D from upwardly sequential locations starting at the address supplied in register pair HL. Address ... 2EDFH

Address

@ (MSX only)

    This routine loads registers E, D, C and B from upwardly sequential locations starting at the address supplied in register pair HL. Address ... 2EE8H

Address

@ (MSX only)

    This routine copies a single precision operand from DAC to the address supplied in register pair HL. Address ... 2EEFH

Address

@ (MSX only)

    This routine copies any operand from the address supplied in register pair HL to ARG. The length of the operand is contained in VALTYP: 2=Integer, 3=String, 4=Single Precision, 8=Double Precision. Address ... 2F05H

Address

@ (MSX only)

    This routine copies any operand from ARG to DAC. The length of the operand is contained in VALTYP: 2=Integer, 3=String, 4=Single Precision, 8=Double Precision. Address ... 2F0DH

Address

@ (MSX only)

    This routine copies any operand from DAC to ARG. The length of the operand is contained in VALTYP: 2=Integer, 3=String, 4=Single Precision, 8=Double Precision.

    Address ... 2F21H
    Name ...... FCOMP (Single Precision Floating Point Comparison)
    Input ..... CBED=Operand 1, DAC=Operand 2
    Exit ...... if CBED>DAC, then A=FFH (-1), Flag NZ, C
                if CBED=DAC, then A=00H (0), Flag Z, NC
                if CBED<DAC, then A=01H (1), Flag NZ, NC
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to find the relation (<>=) between two single precision operands. The first operand is contained in registers C, B, E and D and the second in DAC. The result is returned in register A and the flags.

    It should be noted that for relational operators the Expression Evaluator regards maximally negative numbers as small and maximally positive numbers as large.

    Address ... 2F4DH (MSX Math-Pack)
    Name ...... ICOMP (Integer Comparison)
    Input ..... DE=Operand 1, HL=Operand 2
    Exit ...... if DE<HL, then A=01H (1), Flags NZ, NC
                if DE=HL, then A=00H (0), Flags Z, NC
                if DE>HL, then A=FFH (-1), Flags NZ, C
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to find the relation (<>=) between two signed integer operands. The first operand is contained in register pair DE and the second in register pair HL. The results are as for the single precision version (2F21H).

    Address ... 2F5CH (MSX Math-Pack)
    Name ...... XDCOMP (Extended Double Precision Comparison)
    Input ..... ARG=Operand 1, DAC=Operand 2
    Exit ...... if ARG>DAC, then A=FFH (-1), Flags NZ, C
                if ARG=DAC, then A=00H (0), Flags Z, NC
                if ARG<DAC, then A=01H (1), Flags NZ, NC
    Modifies .. All

    This routine is similar to FCOMP and ICOMP, it compares two double precision Floating Point numbers, contained in ARG (first operand) and DAC (second operand) system variables.

    Address ... 2F83H (MSX only)
    Input ..... DAC=Operand 1, ARG=Operand 2
    Exit ...... if DAC>ARG, then A=FFH (-1), Flags NZ, C
                if DAC=ARG, then A=00H (0), Flags Z, NC
                if DAC<ARG, then A=01H (1), Flags NZ, NC
    Modifies .. All

    This routine is used by the Expression Evaluator to find the relation (<>=) between two double precision operands. The first operand is contained in DAC and the second in ARG. The results are the inverse of the XDCOMP routine: Address ... 2F8AH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CINT" function to an operand contained in DAC. The operand type is first checked via the GETYPR standard routine, if it is already integer the routine simply terminates. If it is a string a "Type mismatch" error is generated (406DH). If it is a single precision or double precision operand it is converted to a signed binary integer in register pair DE (305DH) and then placed in DAC as an integer. Out of range values result in an "Overflow" error (4067H). Address ... 2FA2H

Address

@ (MSX only)

    This routine checks whether DAC contains the single precision operand -32768, if so it replaces it with the integer equivalent 8000H. This step is required during numeric input conversion (3299H) because of the asymmetric integer number range. Address ... 2FB2H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CSNG" function to an operand contained in DAC. The operand's type is first checked via the GETYPR standard routine, if it is already single precision the routine simply terminates. If it is a string a "Type mismatch" error is generated (406DH). If it is double precision VALTYP is changed (3053H) and the mantissa rounded up from the seventh digit (2741H). If the operand is an integer it is converted from binary to a maximum of five BCD digits by successive divisions using the constants 10000, 1000, 100, 10, 1. These are placed in DAC to form the single precision mantissa. The exponent is equal to the number of significant digits in the mantissa. For example if there are five the exponent would be 10^5. Address ... 3030H

Address

@ (MSX only)

    This table contains the five constants used by the "CSNG" routine: -10000, -1000, -100, -10, -1 Address ... 303AH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CDBL" function to an operand contained in DAC. The operand's type is first checked via the GETYPR standard routine, if it is already double precision the routine simply terminates. If it is a string a "Type mismatch" error is generated (406DH). If it is an integer it is first converted to single precision (2FC8H), the eight least significant digits are then zeroed and VALTYP set to 8. Address ... 3058H

Address

@ (MSX only)

    This routine checks that the current operand is a string type, if not a "Type mismatch" error is generated (406DH). - 100 - 5. ROM BASIC INTERPRETER Address ... 305DH

Address

@ (MSX only)

    This routine is used by the "CINT" routine (2F8AH) to convert a BCD single precision or double precision operand into a signed binary integer in register pair DE, it returns Flag C if an overflow has occurred. Successive digits are taken from the mantissa and added to the product starting with the most significant one. After each addition the product is multiplied by ten. The number of digits to process is determined by the exponent, for example five digits would be taken with an exponent of 10^5. Finally the mantissa sign is checked and the product negated (3221H) if necessary. Address ... 30BEH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "FIX" function to an operand contained in DAC. The operand's type is first checked via the GETYPR standard routine, if it is an integer the routine simply terminates. The mantissa sign is then checked (2E71H), if it is positive control transfers to the "INT" routine (30CFH). Otherwise the sign is inverted to positive, the "INT" function is performed (30CFH) and the sign restored to negative.

    Address ... 30CFH (MSX Math-Pack)
    Name ...... INT (Integer)
    Input ..... DAC=Number, VALTYP=DAC numeric type
    Exit ...... DAC=Integer of DAC
    Modifies .. All registers

    This routine is used by the Factor Evaluator to apply the "INT" function to an operand contained in DAC. The operand's type is first checked via the GETYPR standard routine, if it is an integer the routine simply terminates. The number of fractional digits is determined by subtracting the exponent from the type's digit count, 6 for single precision, 14 for double precision.

    If the mantissa sign is positive these fractional digits are simply zeroed. If the mantissa sign is negative each fractional digit is examined before it is zeroed. If all the digits were previously zero the routine simply terminates. Otherwise -1.0 is added to the operand by the single precision addition routine (324EH) or the double precision addition routine (269AH). It should be noted that an operand's type is not normally changed by the "CINT" function.

    Address ... 314AH
    Name ...... UMULT (Unsigned Multiplication)
    Input ..... BC=Multiplicand, DE=Multiplier
    Exit ...... DE=BC*DE (Product)
    Modifies .. AF, BC, DE

    This routine multiplies the unsigned binary integers in register pairs BC (multiplicand) and DE (Multiplier), the result (Product) is returned in register pair DE. The standard shift and add method is used, the product is successively multiplied by two and register pair BC added to it for every 1 bit in register pair DE. The routine is used by the Variable search routine (PTRGET, 5EA4H) to compute an element's position within an Array, a "Subscript out of range" error is generated (601DH) if an overflow occurs. - 101 - 5. ROM BASIC INTERPRETER

    Address ... 3167H
    Name ...... ISUB (Integer Subtraction)
    Input ..... DE=Minuend, HL=Subtrahend
    Exit ...... DAC=HL=DE-HL (Difference)
    Modifies .. All

    This routine is used by the Expression Evaluator to subtract two integer operands. The first operand (minuend) is contained in register pair DE and the second (subtrahend) in register pair HL, the result (difference) is returned in HL and DAC. The second operand is negated (3221H) and control drops into the addition routine.

    Address ... 3172H
    Name ...... IADD (Integer Addition)
    Input ..... DE=Addend 1, HL=Addend 2
    Exit ...... DAC=HL=DE+HL (Sum)
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to add two integer operands. The first operand (addend 1) is contained in register pair DE and the second (addend 2) in register pair HL, the result is returned in DAC. The signed binary operands are normally just added and placed in DAC. However, if an overflow has occurred both operands are converted to single precision (2FCBH) and control transfers to the single precision adder (324EH). An overflow has occurred when both operands are of the same sign and the result is of the opposite sign, for example:

        30000+15000=-20536

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to multiply two integer operands. The first operand is contained in register pair DE and the second in register pair HL, the result is returned in DAC. The two operand signs are saved temporarily and both operands made positive (3215H). Multiplication proceeds using the standard binary shift and add method with register pair HL as the product accumulator, register pair BC containing the first operand and register pair DE the second. If the product exceeds 7FFFH at any time during multiplication both operands are converted to single precision (2FCBH) and control transfers to the single precision multiplier (325CH). Otherwise the initial signs are restored and, if they differ, the product negated before being placed in DAC as an integer (321DH). Address ... 31E6H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to integer divide (\) two integer operands. The first operand is contained in register pair DE and the second in register pair HL, the result is returned in DAC. If the second operand is zero a "Division by zero" error is generated (4058H), otherwise the two operand signs are saved and both operands made positive (3215H). Division proceeds using the standard binary shift and subtract method with register pair HL containing the remainder, register pair BC the second operand and register pair DE the first operand and the product. When division is complete the initial signs are restored and, if they differ, the product is negated before being placed in DAC as an integer (321DH). - 102 - 5. ROM BASIC INTERPRETER Address ... 3215H

Address

@ (MSX only)

    This routine is used to make two signed binary integers, in register pairs HL and DE, positive. Both the initial operand signs are returned as a flag in bit 7 of register B: 0=Same, 1=Different. Each operand is then examined and, if it is negative, made positive by subtracting it from zero. Address ... 322BH

Address

@ (MSX only)

    This routine is used by the "ABS" function to make a negative integer contained in DAC positive. The operand is taken from DAC, negated and then placed back in DAC (3221H). If the operand's value is 8000H it is converted to single precision (2FCCH) as there is no integer of value +32768. Address ... 323AH

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to "MOD" two integer operands. The first operand is contained in register pair DE and the second in register pair HL, the result is returned in DAC. The sign of the first operand is saved and the two operands divided (31E6H). As the remainder is returned doubled by the division process register pair DE is shifted one place right to restore it. The sign of the first operand is then restored and, if it is negative, the remainder is negated before being placed in DAC as an integer (321DH). Address ... 324EH

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to add two single precision operands. The first operand is contained in registers C, B, E, D and the second in DAC, the result is returned in DAC. The first operand is copied to ARG (3280H), the second operand is converted to double precision (3042H) and control transfers to the double precision adder (269AH). Address ... 3257H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to subtract two single precision operands. The first operand is contained in registers C, B, E, D and the second in DAC, the result is returned in DAC. The second operand is negated (2E8DH) and control transfers to the single precision adder (324EH). Address ... 325CH

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to multiply two single precision operands. The first operand is contained in registers C, B, E, D and the second in DAC, the result is returned in DAC. The first operand is copied to ARG (3280H), the second operand is converted to double precision (3042H) and control transfers to the double precision multiplier (27E6H). - 103 - 5. ROM BASIC INTERPRETER Address ... 3265H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to divide two single precision operands. The first operand is contained in registers C, B, E, D and the second in DAC, the result is returned in DAC. The first and second operands are exchanged so that the first is in DAC and the second in the registers. The second operand is then copied to ARG (3280H), the first operand is converted to double precision (3042H) and control transfers to the double precision divider (289FH). Address ... 3280H

Address

@ (MSX only)

    This routine copies the single precision operand contained in registers C, B, E and D to ARG and then zeroes the four least significant bytes. Address ... 3299H

Address

@ (MSX only)

    This routine converts a number in textual form to one of the standard internal numeric types, it is used during tokenization and by the "VAL", "INPUT" and "READ" Statement handlers. On entry register pair HL points to the first character of the text string to be converted. On exit register pair HL points to the character following the string, the numeric operand is in DAC and the type code in VALTYP. Examples of the three types are:

    +-----+-----+-----+-----+-----+-----+-----+-----+
    |     |     | FFH | 7FH |     |     |     |     |
    +-----+-----+-----+-----+-----+-----+-----+-----+
    Integer +32767

    +-----+-----+-----+-----+-----+-----+-----+-----+
    | 42H | 17H | 39H | 04H |     |     |     |     |
    +-----+-----+-----+-----+-----+-----+-----+-----+
    Single Precision .173904*10^2

    +-----+-----+-----+-----+-----+-----+-----+-----+
    | C2H | 17H | 39H | 04H | 62H | 70H | 93H | 13H |
    +-----+-----+-----+-----+-----+-----+-----+-----+
    Double Precision -.17390462709313*10^2

Figure 41: Numeric Types in DAC.

    An integer is a sixteen bit binary number in two's complement form, it is stored LSB first, MSB second at DAC+2. An integer can range from 8000H (-32768) to 7FFFH (+32767).

    A floating point number consists of an exponent byte and a three or seven byte mantissa. The exponent is kept in signed binary form and can range from 01H (-63) through 40H (0) up to 7FH (+63), the special value of 00H is used for the number zero. These exponent values are for a normalized mantissa. The Interpreter presents exponent-form numbers to the user with a leading digit, this results in an asymmetric exponent range of E-64 to E+62. Bit 7 of the exponent byte holds the mantissa sign, 0 for positive and 1 for negative, the mantissa itself is kept in packed BCD form with two digits per byte. It should be noted that the Interpreter uses the contents of VALTYP to determine a number's type, not the format of the number itself.

    Conversion starts by examining the first text character. If this is an "&" control transfers to the special radix conversion routine (4EB8H), if it is a leading sign character it is temporarily saved. Successive numeric characters are then taken and added to the integer product with appropriate multiplications by ten as each new digit is found. If the value of the product exceeds 32767, or a decimal point is found, the product is converted to single precision and any further characters placed directly in DAC. If a seventh digit is found the product is changed to double precision, if more than fourteen digits are found the excess digits are read but ignored.

    Conversion ceases when a non-numeric character is found. If this a type definition character ("%", "#" or "!") the appropriate conversion routine is called and control transfers to the exit point (331EH). If it is an exponent prefix ("E", "e", "D" or "d") one of the conversion routines will also be used and then the following digits converted to a binary exponent in register E. At the exit point (331EH) the product's type is checked via the GETYPR standard routine. If it is single precision or double precision the exponent is calculated by first subtracting the fractional digit count, in register B, from the total digit count, in register D, to produce the leading digit count. This is then added to any explicitly stated exponent, in register E, and placed at DAC+0 as the exponent.

    The leading sign character is restored and the product negated if required (2E86H), if the product is integer the routine then terminates. If the product is single precision control terminates by checking for the special value of -32768 (2FA2H). If the product is double precision control terminates by rounding up from the fifteenth digit (273CH). Address ... 340AH

Address

@ (MSX only)

    This routine is used by the error handler to display the message " in " (6678H) followed by the line number supplied in register pair HL (3412H). Address ... 3412H

Address

@ (MSX only)

    This routine displays the unsigned binary integer supplied in register pair HL. The operand is placed in DAC as an integer (2F99H), converted to text (3441H) and then displayed (6677H). - 105 - 5. ROM BASIC INTERPRETER Address ... 3425H

Address

@ (MSX only)

    This routine converts the numeric operand contained in DAC to textual form in FBUFFR. The address of the first character of the resulting text is returned in register pair HL, the text is terminated by a zero byte. The operand is first converted to double precision (375FH). The BCD digits of the mantissa are then unpacked, converted to ASCII and placed in FBUFFR (36B3H). The position of the decimal point is determined by the exponent, for example:

        .999*10 ^ +2 = 99.9
        .999*10 ^ +1 = 9.99
        .999*10 ^ +0 = .999
        .999*10 ^ -1 = .0999

    If the exponent is outside the range 10^-1 to 10^14 the number is presented in exponential form. In this case the decimal point is placed after the first digit and the exponent is converted from binary and follows the mantissa.

    An alternative entry point to the routine exists at 3426H for the "PRINT USING" statement handler. With this entry point the number of characters to prefix the decimal point is supplied in register B, the number of characters to point fix it in register C and a format byte in register A:

       7     6     5     4     3     2     1     0
    +-----+-----+-----+-----+-----+-----+-----+-----+
    |  1  |  ,  |  *  |  $  |  +  |Sign |  0  |^^^^ |
    +-----+-----+-----+-----+-----+-----+-----+-----+

Figure 42: Format Byte.

    Operation in this mode is fairly similar to the normal mode but with the addition of extra facilities. Once the operand has been converted to double precision the exponential form will be assumed if bit 0 of the format byte is set. The mantissa is shifted to the right in DAC and rounded up to lose unwanted postfix digits (377BH). As the mantissa is converted to ASCII (36B3H) commas will be inserted at the appropriate points if bit 6 of the format byte is set. During post-conversion formatting (351CH) unused prefix positions will be filled with asterisks if bit 5 is set, a pound prefix may be added by setting bit 4. Bit 3 enables the "+" sign for positive numbers if set, otherwise a space is used. Bit 2 places any sign at the front if reset and at the back if set.

    The entry point to the routine at 3441H is used to convert unsigned integers, notably line numbers, to their textual form. For example 9000H, when treated as a normal integer, would be converted to -28672. By using this entry point 36864 would be produced instead. The operand is converted by successive division with the factors 10000, 1000, 100, 10 and 1 and the resulting digits placed in FBUFFR (36DBH). Address ... 3710H

Address

@ (MSX only)

    This table contains the five constants used by the numeric output routine: 10000, 1000, 100, 10, 1. Address ... 371AH

Address

@ (MSX only)

    This routine is used by the "BIN$" function to convert a numeric operand contained in DAC to textual form. Register B is loaded with the group size (1) and control transfers to the general conversion routine (3724H). Address ... 371EH

Address

@ (MSX only)

    This routine is used by the "OCT$" function to convert a numeric operand contained in DAC to textual form. Register B is loaded with the group size (3) and control transfers to the general conversion routine (3724H). Address ... 3722H

Address

@ (MSX only)

    This routine is used by the "HEX$" function to convert a numeric operand contained in DAC to textual form. Register B is loaded with the group size (4) and the operand converted to a binary integer in register pair HL (5439H). Successive groups of 1, 3 or 4 bits are shifted rightwards out of the operand, converted to ASCII digits and placed in FBUFFR. When the operand is all zeroes the routine terminates with the address of the first text character in register pair HL, the string is terminated with a zero byte. Address ... 3752H

Address

@ (MSX only)

    This routine is used during numeric output to return an operand's digit count in register B and the address of its least significant byte in register pair HL. For single precision B=6 and HL=DAC+3, for double precision B=14 and HL=DAC+7. Address ... 375FH

Address

@ (MSX only)

    This routine is used during numeric output to convert the numeric operand in DAC to double precision (303AH). Address ... 377BH

Address

@ (MSX only)

    This routine is used during numeric output to shift the mantissa in DAC rightwards (27DBH), the inverse of the digit count is supplied in register A. The result is then rounded up from the fifteenth digit (2741H). - 107 - 5. ROM BASIC INTERPRETER Address ... 37A2H

Address

@ (MSX only)

    This routine is used during numeric output to return the inverse of the fractional digit count in a floating point operand. This is computed by subtracting the exponent from the operand's digit count (6 or 14). Address ... 37B4H

Address

@ (MSX only)

    This routine is used during numeric output to locate the last non-zero digit of the mantissa contained in DAC. Its address is returned in register pair HL.

    Address ... 37C8H
    Name ...... SNGEXP (Single Precision Floating Point Exponent)
    Input ..... CBED=Base, DAC=Exponent
    Exit ...... DAC=CBED^DAC (Power)
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to exponentiate (^) two single precision operands. The first operand (base) is contained in registers C, B, E, D and the second (exponent) in DAC, the result (power) is returned in DAC. The first operand is copied to ARG (3280H), pushed onto the stack (2CC7H) and exchanged with DAC (2C6FH). The second operand is then popped into ARG and control drops into the double precision exponentiation routine.

    Address ... 37D7H
    Name ...... DBLEXP (Double Precision Floating Point Exponent)
    Input ..... ARG=Base, DAC=Exponent
    Exit ...... DAC=ARG^DAC (Power)
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to exponentiate (^) two double precision operands. The first operand (base) is contained in DAC and the second (exponent) in ARG, the result (power) is returned in DAC. The result is usually computed using:

X^P=EXP(P*LOG(X))

    An alternative, much faster, method is possible if the power operand is an integer. This is tested for by extracting the integer part of the operand and comparing for equality with the original value (391AH). A positive result to this test means that the faster method can be used, this is described below.

    Address ... 383FH
    Name ...... INTEXP (Integer Exponent Power)
    Input ..... DE=Base, HL=Exponent
    Exit ...... DAC=DE^HL (Power)
    Modifies .. AF, HL

    This routine is used by the Expression Evaluator to exponentiate (^) two integer operands. The first operand (base) is contained in register pair DE and the second (exponent) in register pair HL, the result (power) is returned in DAC. The routine operates by breaking the problem down into simple multiplications:

6^13=6^1101=(6^8)*(6^4)*(6^1)

    As the power operand is in binary form a simple right shift is sufficient to determine whether a particular intermediate product needs to be included in the result. The intermediate products themselves are obtained by cumulative multiplication of the operand each time the computation loop is traversed. If the product overflows at any time it is converted to single precision. Upon completion the power operand is checked, if it is negative the product is reciprocated as X^-P=1/X^P. Address ... 390DH

Address

@ (MSX only)

    This routine is used during exponentiation to multiply two integers (3193H), it returns Flag NZ if the result has overflowed to single precision. Address ... 391AH

Address

@ (MSX only)

    This routine is used during exponentiation to check whether a double precision power operand consists only of an integer part, if so it returns Flag NC. Address ... 392EH

Address

@ (MSX only)

    This table of addresses is used by the Interpreter Runloop to find the handler for a statement token. Although not part of the table the associated keywords are included below:

        TO     STMT        TO     STMT       TO     STMT
        ---------------------------------------------------
        63EAH  END         00C3H  CLS        5B11H  CIRCLE
        4524H  FOR         51C9H  WIDTH      7980H  COLOR
        6527H  NEXT        485DH  ELSE       5D6EH  DRAW
        485BH  DATA        6438H  TRON       59C5H  PAINT
        4B6CH  INPUT       6439H  TROFF      00C0H  BEEP
        5E9FH  DIM         643EH  SWAP       73E5H  PLAY
        4B9FH  READ        6477H  ERASE      57EAH  PSET
        4880H  LET         49AAH  ERROR      57E5H  PRESET
        47E8H  GOTO        495DH  RESUME     73CAH  SOUND
        479EH  RUN         53E2H  DELETE     79CCH  SCREEN
        49E5H  IF          49B5H  AUTO       7BE2H  VPOKE
        63C9H  RESTORE     5468H  RENUM      7A48H  SPRITE
        47B2H  GOSUB       4718H  DEFSTR     7B37H  VDP
        4821H  RETURN      471BH  DEFINT     7B5AH  BASE
        485DH  REM         471EH  DEFSNG     55A8H  CALL
        63E3H  STOP        4721H  DEFDBL     7911H  TIME
        4A24H  PRINT       4B0EH  LINE       786CH  KEY
        64AFH  CLEAR       6AB7H  OPEN       7E4BH  MAX
        522EH  LIST        7C52H  FIELD      73B7H  MOTOR
        6286H  NEW         775BH  GET        6EC6H  BLOAD
        48E4H  ON          7758H  PUT        6E92H  BSAVE
        401CH  WAIT        6C14H  CLOSE      7C16H  DSKO$
        501DH  DEF         6B5DH  LOAD       7C1BH  SET
        5423H  POKE        6B5EH  MERGE      7C20H  NAME
        6424H  CONT        6C2FH  FILES      7C25H  KILL
        6FB7H  CSAVE       7C48H  LSET       7C2AH  IPL
        703FH  CLOAD       7C4DH  RSET       7C2FH  COPY
        4016H  OUT         6BA3H  SAVE       7C34H  CMD
        4A1DH  LPRINT      6C2AH  LFILES     7766H  LOCATE
        5229H  LLIST

Address

@ (MSX only)

    This table of addresses is used by the Factor Evaluator to find the handler for a function token. Although not part of the table the associated keywords are included with the addresses shown below:

        TO     FUNCTION     TO     FUNCTION     TO     FUNCTION
        -------------------------------------------------------
        6861H  LEFT$        4FCCH  POS          30BEH  FIX
        6891H  RIGHT$       67FFH  LEN          7940H  STICK
        689AH  MID$         6604H  TR$          794CH  TRIG
        2E97H  SGN          68BBH  VAL          795AH  PDL
        30CFH  INT          680BH  ASC          7969H  PAD
        2E82H  ABS          681BH  CHR$         7C39H  DSKF
        2AFFH  SQR          541CH  PEEK         6D39H  FPOS
        2BDFH  RND          7BF5H  VPEEK        7C66H  CVI
        29ACH  SIN          6848H  SPACE$       7C6BH  CVS
        2A72H  LOG          7C70H  OCT$         7C70H  CVD
        2B4AH  EXP          65FAH  HEX$         6D25H  EOF
        2993H  COS          4FC7H  LPOS         6D03H  LOC
        29FBH  TAN          6FFFH  BIN$         6D14H  LOF
        2A14H  ATN          2F8AH  CINT         7C57H  MKI$
        69F2H  FRE          2FB2H  CSNG         7C5CH  MKS$
        4001H  INP          303AH  CDBL         7C61H  MKD$

Address

@ (MSX only)

    This table of addresses is used during program tokenization as an index into the BASIC keyword table (3A72H). Each of the twenty six entries defines the starting address of one of the keyword sub-blocks. The first entry points to the keywords beginning with the letter "A", the second to those beginning with the letter "B" and so on.

        3A72H ... A   3B9FH ... J    3C8EH ... S
        3A88H ... B   3BA0H ... K    3CDBH ... T
        3A9FH ... C   3BA8H ... L    3CF6H ... U
        3AF3H ... D   3BE8H ... M    3CFFH ... V
        3B2EH ... E   3C09H ... N    3D16H ... W
        3B4FH ... F   3C18H ... O    3D20H ... X
        3B69H ... G   3C2BH ... P    3D24H ... Y
        3B7BH ... H   3C5DH ... Q    3D25H ... Z
        3B80H ... I   3C5EH ... R

Address

@ (MSX only)

    This table contains the BASIC keywords and tokens. Each of the twenty-six blocks within the table contains all the keywords beginning with a particular letter, it is terminated with a zero byte. Each keyword is stored in plain text with bit 7 set to mark the last character, this is followed immediately by the associated token. The first character of the keyword need not be stored as this is implied by its position in the table' The keywords and tokens are listed below in full, note that the "J", "Q", "Y" and "Z" blocks are empty:

        AUTO   A9H  DSKF   26H  LIST    93H  REM     8FH
        AND    F6H  DRAW   BEH  LFILES  BBH  RESUME  A7H
        ABS    06H  ELSE   A1H  LOG     0AH  RSET    B9H
        ATN    0EH  END    81H  LOC     2CH  RIGHT$  02H
        ASC    15H  ERASE  A5H  LEN     12H  RND     08H
        ATTR$  E9H  ERROR  A6H  LEFT$   01H  RENUM   AAH
        BASE   C9H  ERL    E1H  LOF     2DH  SCREEN  C5H
        BSAVE  D0H  ERR    E2H  MOTOR   CEH  SPRITE  C7H
        BLOAD  CFH  EXP    0BH  MERGE   B6H  STOP    90H
        BEEP   C0H  EOF    2BH  MOD     FBH  SWAP    A4H
        BIN$   1DH  EQV    F9H  MKI$    2EH  SET     D2H
        CALL   CAH  FOR    82H  MKS$    2FH  SAVE    BAH
        CLOSE  B4H  FIELD  B1H  MKD$    30H  SPC(    DFH
        COPY   D6H  FILES  B7H  MID$    03H  STEP    DCH
        CONT   99H  FN     DEH  MAX     CDH  SGN     04H
        CLEAR  92H  FRE    0FH  NEXT    83H  SQR     07H
        CLOAD  9BH  FIX    21H  NAME    D3H  SIN     09H
        CSAVE  9AH  FPOS   27H  NEW     94H  STR$    13H
        CSRLIN E8H  GOTO   89H  NOT     E0H  STRING$ E3H
        CINT   1EH  GO TO  89H  OPEN    B0H  SPACE$  19H
        CSNG   1FH  GOSUB  8DH  OUT     9CH  SOUND   C4H
        CDBL   20H  GET    B2H  ON      95H  STICK   22H
        CVI    28H  HEX$   1BH  OR      F7H  STRIG   23H
        CVS    29H  INPUT  85H  OCT$    1AH  THEN    DAH
        CVD    2AH  IF     8BH  OFF     EBH  TRON    A2H
        COS    0CH  INSTR  E5H  PRINT   91H  TROFF   A3H
        CHR$   16H  INT    05H  PUT     B3H  TAB(    DBH
        CIRCLE BCH  INP    10H  POKE    98H  TO      D9H
        COLOR  BDH  IMP    FAH  POS     11H  TIME    CBH
        CLS    9FH  INKEY$ ECH  PEEK    17H  TAN     0DH
        CMD    D7H  IPL    D5H  PSET    C2H  USING   E4H
        DELETE A8H  KILL   D4H  PRESET  C3H  USR     DDH
        DATA   84H  KEY    CCH  POINT   EDH  VAL     14H
        DIM    86H  LPRINT 9DH  PAINT   BFH  VARPTR  E7H
        DEFSTR ABH  LLIST  9EH  PDL     24H  VDP     C8H
        DEFINT ACH  LPOS   1CH  PAD     25H  VPOKE   C6H
        DEFSNG ADH  LET    88H  PLAY    C1H  VPEEK   18H
        DEFDBL AEH  LOCATE D8H  RETURN  8EH  WIDTH   A0H
        DSKO$  D1H  LINE   AFH  READ    87H  WAIT    96H
        DEF    97H  LOAD   B5H  RUN     8AH  XOR     F8H
        DSKI$  EAH  LSET   B8H  RESTORE 8CH

Address

@ (MSX only)

    This twenty-one byte table is used by the Interpreter during program tokenization. It contains the ten single character keywords and their tokens:

  + ... F1H    * ... F3H   ^ ... F5H    ' ...  E6H = ... EFH
  - ... F2H    / ... F4H   \ ... FCH    > ...  EEH < ... F0H

Address

@ (MSX only)

    This table is used by the Expression Evaluator to determine the precedence level for a given infix operator, the higher the table value the greater the operator's precedence. Not included are the precedences for the relational operators (64H), the "NOT" operator (5AH) and the negation operator (7DH), these are defined directly by the Expression and Factor Evaluators.

        79H ... +       46H ... OR
        79H ... -       3CH ... XOR
        7CH ... *       32H ... EQV
        7CH ... /       28H ... IMP
        7FH ... ^       7AH ... MOD
        50H ... AND     7BH     \

Address

@ (MSX only)

    This table is used to convert the result of a user defined function to the same type as the Variable used in the function definition. It contains the addresses of the type conversion routines:

        303AH ... CDBL
        0000H ... Not used
        2F8AH ... CINT
        3058H ... Check string type
        2FB2H ... CSNG

Address

@ (MSX only)

    This table of addresses is used by the Expression Evaluator to find the handler for a particular infix math operator when both operands are double precision:

        269AH ... +
        268CH ... -
        27E6H ... *
        289FH ... /
        37D7H ... ^
        2F83H ... Relation

Address

@ (MSX only)

    This table of addresses is used by the Expression Evaluator to find the handler for a particular infix math operator when both operands are single precision:

        324EH ... +
        3257H ... -
        325CH ... *
        3267H ... /
        37C8H ... ^
        2F21H ... Relation

Address

@ (MSX only)

    This table of addresses is used by the Expression Evaluator to find the handler for a particular infix math operator when both operands are integer:

        3172H ... +
        3167H ... -
        3193H ... *
        4DB8H ... /
        383FH ... ^
        2F4DH ... Relation

Address

@ (MSX only)

    This table contains the Interpreter error messages, each one is stored in plain text with a zero byte terminator. The associated error codes are shown below for reference only, they do not form part of the table:

        01 NEXT without FOR             19 Device I/O error
        02 Syntax error                 20 Verify error
        03 RETURN without GOSUB         21 No RESUME
        04 Out of DATA                  22 RESUME without error
        05 Illegal function call        23 Unprintable error
        06 Overflow                     24 Missing operand
        07 Out of memory                25 Line buffer overflow
        08 Undefined line number        50 FIELD overflow
        09 Subscript out of range       51 Internal error
        10 Redimensioned array          52 Bad file number
        11 Division by zero             53 File not found
        12 Illegal direct               54 File already open
        13 Type mismatch                55 Input past end
        14 Out of string space          56 Bad file name
        15 String too long              57 Direct statement in file
        16 String formula too complex   58 Sequential I/O only
        17 Can't CONTINUE               59 File not OPEN
        18 Undefined user function

Address

@ (MSX only)

    This is the plain text message " in " terminated by a zero byte. Address ... 3FD7H

Address

@ (MSX only)

    This is the plain text message "Ok", CR, LF terminated by a zero byte. Address ... 3FDCH

Address

@ (MSX only)

    This is the plain text message "Break" terminated by a zero byte. - 113 - 5. ROM BASIC INTERPRETER Address ... 3FE2H

Address

@ (MSX only)

    This routine searches the Z80 stack for the "FOR" loop parameter block whose loop Variable address is supplied in register pair DE. The search is started four bytes above the current Z80 SP to allow for the caller's return address and the Runloop return address. If no "FOR" token (82H) exists the routine terminates Flag NZ, if one is found the loop Variable address is taken from the parameter block and checked. The routine terminates Flag Z upon a successful match with register pair HL pointing to the type byte of the parameter block. Otherwise the search moves up twenty-two bytes to the start of the next parameter block. Address ... 4001H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "INP" function to an operand contained in DAC. The port number is checked (5439H), the port read and the result placed in DAC as an integer (4FCFH). Address ... 400BH

Address

@ (MSX only)

    This routine first evaluates an operand in the range -32768 to +65535 (FRMQNT, 542FH) and places it in register pair BC. After checking for a comma, via the SYNCHR standard routine, it evaluates a second operand in the range 0 to 255 (GETBYT, 521CH) and places this in register A. Address ... 4016H

Address

@ (MSX only)

    This is the "OUT" statement handler. The port number and data byte are evaluated (400BH) and the data byte written to the relevant Z80 port. Address ... 401CH

Address

@ (MSX only)

    This is the "WAIT" statement handler. The port number and "AND" operands are first evaluated (400BH) followed by the optional "XOR" operand (GETBYT, 521CH). The port is then repeatedly read and the operands applied, XOR then AND, until a non-zero result is obtained. Contrary to the information given in some MSX manuals the loop can be broken by the CTRL-STOP key as the CKCNTC standard routine is called from inside it. Address ... 4039H

Address

@ (MSX only)

    This routine is used by the Runloop when it encounters the end of the program text while in program mode. ONEFLAG is checked to see whether it still contains an active error code. If so a "No RESUME" error is generated, otherwise program termination continues normally (6401H). The idea behind this routine is to catch any "ON ERROR" handlers without a "RESUME" statement at the end. - 114 - 5. ROM BASIC INTERPRETER Address ... 404FH

Address

@ (MSX only)

    This routine is used by the "READ" statement handler when an error is found in a "DATA" statement. The line number contained in DATLIN is copied to CURLIN so the error handler will flag the "DATA" line as the illegal statement rather than the program line. Control then drops into the "Syntax error" generator. Address ... 4055H

Address

@ (MSX only)

    This is a group of nine error generators, register E is loaded with the relevant error code and control drops into the error handler:

        ADDR. ERROR
        -------------------------------
        4055H Syntax error
        4058H Division by zero
        405BH NEXT without FOR
        405EH Redimensioned array
        4061H Undefined user function
        4064H RESUME without error
        4067H Overflow error
        406AH Missing operand
        406DH Type mismatch

Address

@ (MSX only)

    This is the Interpreter error handler, all error generators transfer to here with an error code in register E. VLZADR is first checked to see if the "VAL" statement handler has changed the program text, if so the original character is restored from VLZDAT. The current line number is then copied from CURLIN to ERRLIN and DOT and the Z80 stack is restored from SAVSTK (62F0H). The error code is placed in ERRFLG, for use by the "ERR" function, and the current program text position copied from SAVTXT to ERRTXT for use by the "RESUME" statement handler. The error line number and program text position are also copied to OLDLIN and OLDTXT for use by the "CONT" statement handler. ONELIN is then checked to see if a previous "ON ERROR" statement has been executed. If so, and providing no error code is already active, control transfers to the Runloop (4620H) to execute the BASIC error recovery statements.

    Otherwise the error code is used to count through the error message table at 3D75H until the required one is reached. A CR,LF is issued (7323H) and the screen forced back to text mode via the TOTEXT standard routine. A BELL code is then issued and the error message displayed (6678H). Assuming the Interpreter is in program mode, rather than direct mode, this is followed by the line number (340AH) and control drops into the "OK" point. - 115 - 5. ROM BASIC INTERPRETER Address ... 411FH

Address

@ (MSX only)

    This is the re-entry point to the Interpreter Mainloop for a terminating program. The screen is forced to text mode via the TOTEXT standard routine, the printer is cleared (7304H) and I/O buffer 0 closed (6D7BH). A CR,LF is then issued to the screen (7323H), the message "OK" is displayed (6678H) and control drops into the Mainloop. Address ... 4134H

Address

@ (MSX only)

    This is the Interpreter Mainloop. CURLIN is first set to FFFFH to indicate direct mode and AUTFLG checked to see if "AUTO" mode is on. If so the next line number is taken from AUTLIN and displayed (3412H). The Program Text Area is then searched to see if this line already exists (4295H) and either an asterisk or space displayed accordingly.

    The ISFLIO standard routine is then used to determine whether a "LOAD" statement is active. If so the program line is collected from the cassette (7374H), otherwise it is taken from the console via the PINLIN standard routine. If the line is empty or the CTRL-STOP key has been pressed control transfers back to the start of the Mainloop (4134H) with no further action. If the line commences with a line number this is converted to an unsigned integer in register pair DE (4769H). The line is then converted to tokenized form and placed in KBUF (CRUNCH, 42B2H). If no line number was found at the start of the line control then transfers to the Runloop (6D48H) to execute the statement.

    Assuming the line commences with a line number it is tested to see if it is otherwise empty and the result temporarily saved. The line number is copied to DOT and AUTLIN increased by the contents of AUTINC, if AUTLIN now exceeds 65530 the "AUTO" mode is turned off. The Program Text Area is then searched (4295H) to find a matching line number or, failing this, the position of the next highest line number. If no matching line number is found and the line is empty and "AUTO" mode is off an "Undefined line number" error is generated (481CH). If a matching line number is found and the line is empty and "AUTO" mode is on the Mainloop simply skips to the next statement (4237H).

    Otherwise any pointers in the Program Text Area are converted back to line numbers (54EAH) and any existing program line deleted (5405H). Assuming the new program line is non- empty the Program Text Area is opened up by the required amount (6250H) and the tokenized program line copied from KBUF.

    The Program Text Area links are then recalculated (4257H), the Variable Storage Areas are cleared (629AH) and control transfers back to the start of the Mainloop. - 116 - 5. ROM BASIC INTERPRETER Address ... 4253H

Address

@ (MSX only)

    This routine recalculates the Program Text Area links after a program modification. The first two bytes of each program line contain the starting address of the following line, this is called the link. Although the link increases the amount of storage required per program line it greatly reduces the time required by the Interpreter to locate a given line.

    An example of a typical program line is shown below, in this case the line "10 PRINT 9" situated at the start of the Program Text Area (8001H):

    +---------+---------+-----+-----+-----+-----+
    | 09H 80H | 0AH 00H | 91H | 20H | 1AH | 00H |
    +---------+---------+-----+-----+-----+-----+

Figure 43: Program Line.

    In the above example the link is stored in Z80 word order (LSB,MSB) and is immediately followed by the binary line number, also in word order. The statement itself is composed of a "PRINT" token (91H), a single space, the number nine and the end of line character (00H). Further details of the storage format can be found in the tokenizing routine (CRUNCH, 42B2H).

    Each link is recalculated simply by scanning through the line until the end of line character is found. The process is complete when the end of the Program Storage Area, marked by the special link of 0000H, is reached. Address ... 4279H

Address

@ (MSX only)

    This routine is used by the "LIST" statement handler to collect up to two line number operands from the program text. If the first line number is present it is converted to an unsigned integer in register pair DE (475FH), if not a default value of 0000H is returned. If the second line number is present it must be preceded by a "-" token (F2H) and is returned on the Z80 stack, if not a default value of 65530 is returned. Control then drops into the program text search routine to find the first referenced program line. Address ... 4295H

Address

@ (MSX only)

    This routine searches the Program Text Area for the program line whose line number is supplied in register pair DE. Starting at the address contained in TXTTAB each program line is examined for a match. If an equal line number is found the routine terminates with Flag Z and C and register pair BC pointing to the start of the program line. If a higher line number is found the routine terminates Flag NZ and NC and if the end link is reached the routine terminates Flag Z and NC. - 117 - 5. ROM BASIC INTERPRETER

    Address ... 42B2H
    Name ...... CRUNCH (Crunch)
    Input ..... HL=ASCII text address (ended by 00H)
    Exit ...... KBUF=tokenized line
    Modifies .. All registers

    This routine is used by the Interpreter Mainloop to tokenize a line of text. On entry register pair HL points to the first text character in BUF. On exit the tokenized line is in KBUF, register pair BC holds its length and register pair HL points to its start.

    Except after opening quotes or after the "REM", "CALL" or "DATA" keywords any string of characters matching a keyword is replaced by that keyword's token. Lower case alphabetics are changed to upper case for keyword comparison. The character "?" is replaced by the "PRINT" token (91H) and the character "'" by ":" (3AH), "REM" token (8FH), "'" token (E6H). The "ELSE" token (A1H) is preceded by a statement separator (3AH). Any other miscellaneous characters in the text are copied without alteration except that lower case alphabetics are converted to upper case. Those tokens smaller than 80H, the function tokens, cannot be stored directly in KBUF as they will conflict with ordinary text. Instead the sequence FFH, token+80H is used.

    Numeric constants are first converted into one of the standard types in DAC (3299H). They are then stored in one of several ways depending upon their type and magnitude, the general idea being to minimize memory usage:

        0BH LSB MSB ................... Octal number
        0CH LSB MSB ................... Hex number
        11H to 1AH .................... Integer 0 to 9
        0FH LSB ....................... Integer 10 to 255
        1CH LSB MSB ................... Integer 256 to 32767
        1DH EE DD DD DD ............... Single Precision
        1FH EE DD DD DD DD DD DD DD ... Double Precision

    There is no specific token for binary numbers, these are left as character strings. This would appear to be a legacy from earlier versions of Microsoft BASIC. Any sign prefixing a number is regarded as an operator and is stored as a separate token, negative numbers are not produced during tokenization. As double precision numbers occupy so much space a line containing too many, for example PRINT 1#,1#,1# etc. may cause KBUF to fill up. If this happens a "Line buffer overflow" error is generated.

    Any number following one of the keyword tokens in the table at 43B5H is considered to be a line number operand and is stored with a different token:

        0DH LSB MSB ................... Pointer
        0EH LSB MSB ................... Line number

    During tokenization only the normal type (0EH) is generated, when a program actually runs these line number operands are converted to the address pointer type (0DH). - 118 - 5. ROM BASIC INTERPRETER Address ... 43B5H

Address

@ (MSX only)

    This table of tokens is used during tokenization to check for the keywords which take line number operands. The keywords themselves are listed below:

        RESTORE    RUN
        AUTO       LIST
        RENUM      LLIST
        DELETE     GOTO
        RESUME     RETURN
        ERL        THEN
        ELSE       GOSUB

Address

@ (MSX only)

    This is the "FOR" statement handler. The loop Variable is first located and assigned its initial value by the "LET" handler (4880H), the address of the loop Variable is returned in register pair DE. The end of the statement is found (485BH) and its address placed in ENDFOR. The Z80 stack is then searched (3FE6H) for any parameter blocks using the same loop Variable. For each one found the current ENDFOR address is compared with that of the parameter block, if there is a match that section of the stack is discarded. This is done in case there are any incomplete loops as a result of a "GOTO" back to the "FOR" statement from inside the loop.

    The termination operand and optional "STEP" operand are then evaluated and converted to the same type as the loop Variable. After checking that stack space is available (625EH) a twenty- five byte parameter block is pushed onto the Z80 stack. This is made up of the following:

        2 bytes ... ENDFOR address
        2 bytes ... Current line number
        8 bytes ... Loop termination value
        8 bytes ... STEP value
        1 byte  ... Loop type
        1 byte  ... STEP direction
        2 bytes ... Address of loop Variable
        1 byte  ... FOR token (82H)

    The parameter block remains on the stack for use by the "NEXT" statement handler until termination is reached, it is then discarded. The size of the block remains constant even though, for integer and single precision loop Variables, the full eight bytes are not required for the termination and STEP values. In these cases the least significant bytes are packed out with garbage.

    It should be noted that the type of arithmetic operation performed by the "NEXT" statement handler, and hence the loop execution speed, depends entirely upon the loop Variable type and not the operand types. For the fastest program execution integer type Variables, N% for example, should be used.

    Address ... 4601H
    Name ...... NEWSTT (New Statement)
    Input ..... HL=pointer to program address
    Exit ...... None
    Modifies .. All registers

    This is the Runloop, each statement handler returns here upon completion so the Interpreter can proceed to the next statement. The current Z80 SP is copied to SAVSTK for error recovery purposes and the CTRL-STOP key checked via the ISCNTC standard routine. Any pending interrupts are processed (6389H) and the current program text position, held in register pair HL throughout the Interpreter, is copied to SAVTXT.

    The current program character is then examined, if this is a statement separator (3AH) control transfers immediately to the execution point (4640H). If it is anything else but an end of line character (00H) a "Syntax error" is generated (4055H) as there is spurious text at the end of the statement. Register pair HL is advanced to the first character of the new program line and the link examined, if this is zero the program is terminated (4039H). Otherwise the line number is taken from the new line and placed in CURLIN. If TRCFLG is non-zero the line number is displayed (3412H) enclosed by square brackets, control then drops into the execution point. Address ... 4640H

Address

@ (MSX only)

    This is the Runloop execution point. A return to the start of the Runloop (NEWSTT, 4601H) is pushed onto the Z80 stack and the first character taken from the new statement via the CHRGTR standard routine. If it is an underline character (5FH) control transfers to the "CALL" statement handler (55A7H). If it is smaller than 81H, the smallest statement token, control transfers to the "LET" handler (4880H). If it is larger than D8H, the largest statement token, it is checked to see if it is one of the function tokens allowed as a statement (51ADH). Otherwise the handler address is taken from the table at 392EH and pushed onto the stack. Control then drops into the CHRGTR standard routine to fetch the next program character before control transfers to the statement handler.

    Address ... 4666H (from 0010H)
    Name ...... CHRGTR (Character Get into Register A)
    Input ..... HL points to current program character
    Exit ...... A=Next program character
    Modifies .. AF, HL

    Standard routine to fetch the next character from the program text. Register pair HL is incremented and the character placed in register A. If it is a space, TAB code (09H) or LF code (0AH) it is skipped over. If it is a statement separator (3AH) or end of line character (00H) the routine terminates with Flag Z and NC. If it is a digit from "0" to "9" the routine terminates with Flag NZ and C. If it is any other character apart from the numeric prefix tokens the routine terminates Flag NZ and NC. If the character is one of the numeric prefix tokens then it is placed in CONSAV and the operand copied to CONLO. The type code is placed in CONTYP and the address of the trailing program character in CONTXT. Address ... 46E8H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator and during detokenization to recover a numeric operand when one of the prefix tokens is returned by the CHRGTR standard routine. The prefix token is first taken from CONSAV, if it is anything but a line number or pointer token the operand is copied from CONLO to DAC and the type code copied from CONTYP to VALTYP. If it is a line number it is converted to single precision and placed in DAC (3236H). If it is a pointer the original line number is recovered from the referenced program line, converted to single precision and placed in DAC (3236H). Address ... 4718H

Address

@ (MSX only)

    This is the "DEFSTR" statement handler. Register E is loaded with the string type code (03H) and control drops into the general type definition routine. Address ... 471BH

Address

@ (MSX only)

    This is the "DEFINT" statement handler. Register E is loaded with the integer type code (02H) and control drops into the general type definition routine. Address ... 471EH

Address

@ (MSX only)

    This is the "DEFSNG" statement handler. Register E is loaded with the single precision type code (04H) and control drops into the general type definition routine. Address ... 4721H

Address

@ (MSX only)

    This is the "DEFDBL" statement handler. Register E is loaded with the double precision type code (08H) and the first range definition character checked (64A7H). If this is not upper case alphabetic a "Syntax error" is generated (4055H). If a "-" token (F2H) follows the second range definition character is taken and checked (64A7H), the difference between the two determines the number of entries in DEFTBL that are filled with the type code. Address ... 4755H

Address

@ (MSX only)

    This routine evaluates an operand and converts it to an integer in register pair DE (520FH). If the operand is negative an "Illegal function call" error is generated. - 121 - 5. ROM BASIC INTERPRETER Address ... 475FH

Address

@ (MSX only)

    This routine is used by the statement handlers shown in the table at 43B5H to collect a single line number operand from the program text and convert it to an unsigned integer in register pair DE. If the first character in the text is a "." (2EH) the routine terminates with the contents of DOT. If it is one of the line number tokens (0DH or 0EH) the routine terminates with the contents of CONLO. Otherwise successive digits are taken and added to the product, with appropriate multiplications by ten, until a non-numeric character is found. Address ... 479EH

Address

@ (MSX only)

    This is the "RUN" statement handler. If no line number operand is present in the program text the system is cleared (629AH) and control returns to the Runloop with register pair HL pointing to the start of the Program Storage Area. If a line number operand is present the system is cleared (62A1H) and control transfers to the "GOTO" statement handler (47E7H). Otherwise a following filename is assumed, for example RUN "CAS:FILE", and control transfers to the "LOAD" statement handler (6B5BH). Address ... 47B2H

Address

@ (MSX only)

    This is the "GOSUB" statement handler. After checking that stack space is available (625EH) the line number operand is collected and placed in register pair DE (4769H). The seven byte parameter block is then pushed onto the stack and control transfers to the "GOTO" handler (47EBH). The parameter block is made up of the following:

        2 bytes ... End of statement address
        2 bytes ... Current line number
        2 bytes ... 0000H
        1 byte  ... GOSUB token (8DH)

    The parameter block remains on the stack until a "RETURN" statement is executed. It is then used to determine the original program text position after which it is discarded. Address ... 47CFH

Address

@ (MSX only)

    This routine is used by the Runloop interrupt processor (6389H) to create a "GOSUB" type parameter block on the Z80 stack. An interrupt block is identical to a normal block except that the two zero bytes shown above are replaced by the address of the device's entry in TRPTBL. This address will be used by the "RETURN" statement handler to update the device's interrupt status once a subroutine has terminated. After pushing the parameter block control transfers to the Runloop to execute the program line whose address is supplied in register pair DE. - 122 - 5. ROM BASIC INTERPRETER Address ... 47E8H

Address

@ (MSX only)

    This is the "GOTO" statement handler. The line number operand is collected (4769H) and placed in register pair HL. If it is a pointer control transfers immediately to the Runloop to begin execution at the new program text position. Otherwise the line number is compared with the current line number to determine the starting position for the program text search. If it is greater the search starts from the end of this line (4298H), if it is smaller it starts from the beginning of the Program Text Area (4295H). If the referenced line cannot be found an "Undefined line number" error is generated (481CH). Otherwise the line number operand is replaced by the referenced program line's address and its token changed to the pointer type (5583H). Control then transfers to the Runloop to execute the referenced program line. Address ... 481CH

Address

@ (MSX only)

    This is the "Undefined line number" error generator. Address ... 4821H

Address

@ (MSX only)

    This is the "RETURN" statement handler. A dummy loop Variable address is placed in register pair DE and the Z80 stack searched (3FE2H) to find the first parameter block not belonging to a "FOR" loop, this section of stack is then discarded. If no "GOSUB" token (8DH) is found at this point a "RETURN without GOSUB" error is generated.

    The next two bytes are then taken from the block, if they are non-zero the block was generated by an interrupt and the temporary "STOP" condition is removed (633EH). The program text is then examined, if anything follows the "RETURN" token itself it is assumed to be a line number operand and control transfers to the "GOTO" handler (47E8H). Otherwise the old line number and program text address are taken from the block and control returns to the Runloop. Address ... 485BH

Address

@ (MSX only)

    This is the "DATA" statement handler. The program text is skipped over until a statement separator (3AH) or end of line character (00H) is found. This routine is also the "REM" and "ELSE" statement handler via the entry point at 485DH, in this case only the end of line character acts as a terminator. Address ... 4880H

Address

@ (MSX only)

    This is the "LET" statement handler. The Variable is first located (PTRGET, 5EA4H), its address saved in TEMP and the operand evaluated (FRMEVL, 4C64H). If necessary the operand's type is then changed to match that of the Variable (517AH). Assuming the operand is one of the three numeric types it is simply copied - 123 - 5. ROM BASIC INTERPRETER from DAC to the Variable in the Variable Storage Area (2EF3H). If the operand is a string type the address of the string body is taken from the descriptor and checked. If it is in KBUF, as would be the case for an explicit string in a direct statement, the body is first copied to the String Storage Area and a new descriptor created (6611H). The descriptor is then freed from TEMPST (67EEH) and copied to the Variable in the Variable Storage Area (2EF3H). Address ... 48E4H

Address

@ (MSX only)

    This is the "ON ERROR", "ON DEVICE GOSUB" and "ON EXPRESSION" statement handler. If the next program text character is not an "ERROR" token (A6H) control transfers to the "ON DEVICE GOSUB" and "ON EXPRESSION" handler (490DH). The program text is checked to ensure that a "GOTO" token (89H) follows and then the line number operand collected (4769H). The program text is searched to obtain the address of the referenced line (4293H) and this is placed in ONELIN. If the line number operand is non-zero the routine then terminates. If the line number operand is zero ONEFLG is checked to see if an error situation already exists (implying that the statement is inside a BASIC error recovery routine). If so control transfers to the error handler (4096H) to force an immediate error, otherwise the routine terminates normally. Address ... 490DH

Address

@ (MSX only)

    This is the "ON DEVICE GOSUB" and "ON EXPRESSION" statement handler. If the next program text character is not a device token (7810H) control transfers to the "ON EXPRESSION" handler (4943H). After checking the program text for a "GOSUB" token (8DH) each of the line number operands required for a particular device is collected in turn (4769H). Assuming a given line number operand is non-zero the program text is searched to find the address of the referenced line (4293H) and this is placed in the device's entry in TRPTBL (785CH). The routine terminates when no more line number operands are found. Address ... 4943H

Address

@ (MSX only)

    This is the "ON EXPRESSION" statement handler. The operand is evaluated (GETBYT, 521CH) and the following "GOSUB" token (8DH) or "GOTO" token (89H) placed in register A. The operand is then used to count along the program text until register pair HL points to the required line number operand. Control then transfers back to the Runloop execution point (4646H) to decode the "GOSUB" or "GOTO" token. Address ... 495DH

Address

@ (MSX only)

    This is the "RESUME" statement handler. ONEFLG is first checked to make sure that an error condition already exists, if not a "RESUME without error" is generated (4064H). If a non-zero line number operand follows control transfers to the "GOTO" handler (47EBH). If a "NEXT" token (83H) follows the position of the error is restored from ERRTXT and ERRLIN, the start of the next statement is found (485BH) and the routine terminates. If there is no line number operand or if it is zero the position of the error is found from ERRTXT and ERRLIN and the routine terminates. Address ... 49AAH

Address

@ (MSX only)

    This is the "ERROR" statement handler. The operand is evaluated and placed in register E (GETBYT, 521CH). If it is zero an "Illegal function call" error is generated (475AH), otherwise control transfers to the error handler (406FH). Address ... 49B5H

Address

@ (MSX only)

    This is the "AUTO" Statement handler. The optional start and increment line number operands, both with a default value of ten, are collected (475FH) and placed in AUTLIN and AUTINC. After making AUTFLG non-zero the Runloop return is destroyed and control transfers directly to the Mainloop (4134H). Address ... 49E5H

Address

@ (MSX only)

    This is the "IF" statement handler. The operand is evaluated (FRMEVL, 4C64H) and, after checking for a "GOTO" token (89H) or "THEN" token (DAH), its sign is tested (2EA1H). If the operand is non-zero (true) the following text is executed either by an immediate transfer to the Runloop (4646H) or, for a line number operand, the "GOTO" handler (47E8H). If the operand is zero (false) the statement text is scanned (485BH) until an "ELSE" token (A1H) is found not balanced by an "IF" token (8BH) and execution re-commences. Address ... 4A1DH

Address

@ (MSX only)

    This is the "LPRINT" statement handler. PRTFLG is set to 01H, to direct output to the printer, and control transfers to the "PRINT" handler (4A29H). Address ... 4A24H

Address

@ (MSX only)

    This is the "PRINT" statement handler. The program text is first checked for a trailing buffer number and, if necessary, PTRFIL set to direct output to the required I/O buffer (6D57H). If no more program text exists a CR,LF is issued (7328H) and the routine terminates (4AFFH). Otherwise successive characters are taken from the program text and analyzed. If a "USING" token (E4H) is found control transfers to the "PRINT USING" handler (60B1H). If a ";" character is found control just transfers back to the start to fetch the next item (4A2EH). If a comma is found sufficient spaces are issued to bring the current print position, from TTYPOS, LPTPOS or an I/O buffer FCB, to an integral multiple of fourteen. If output is directed to the screen and the print position is equal to or greater than the contents of CLMLST or if output is directed to the printer and it is equal to or greater than 238 then a CR,LF is issued instead (7328H). If a "SPC(" token (DFH) is found the operand is evaluated (521BH) and the required number of spaces are output. If a "TAB(" token (DBH) is found the operand is evaluated (521BH) and sufficient spaces issued to bring the current print position, from TTYPOS, LPTPOS or an I/O buffer FCB, to the required point.

    If none of these characters is found the program text contains a data item which is then evaluated (FRMEVL, 4C64H). If the operand is a string it is simply displayed (667BH). If it is numeric it is first converted to text in FBUFFR (3425H) and a string descriptor created (6635H). If output is directed to an I/O buffer the resulting string is then displayed (667BH). If output is directed to the screen or printer the current print position, from TTYPOS or LPTPOS, is compared with the line length and a CR,LF issued (7328H) if the output will not fit on the line. The maximum line length is 255 for the printer and is taken from LINLEN for the screen. Once the string has been displayed control transfers back to the start of the handler. Address ... 4AFFH

Address

@ (MSX only)

    This routine zeroes PRTFLG and PTRFIL to return the Interpreter's output to the screen. Address ... 4B0EH

Address

@ (MSX only)

    This is the "LINE INPUT", "LINE INPUT#" and "LINE" statement handler. If the following program text character is anything other than an "INPUT" token (85H) control transfers to the "LINE" statement handler (58A7H). If the following program text character is a "#" (23H) control transfers to the "LINE INPUT#" statement handler (6D8FH).

    Any following prompt string is evaluated and displayed (4B7BH) and the Variable located (PTRGET, 5EA4H) and checked to ensure that it is a string type (3058H). The line of text is collected from the console via the INLIN standard routine, if Flag C (CTRL-STOP) is returned control transfers to the "STOP" statement handler (63FEH). Otherwise the input string is analyzed and a descriptor created (6638H), control then transfers to the "LET" statement handler for assignment (4892H). It should be noted that the screen is not forced to text mode before the input is collected. Address ... 4B3AH

Address

@ (MSX only)

    This is the plain text message "?Redo from start", CR, LF terminated by a zero byte. - 126 - 5. ROM BASIC INTERPRETER Address ... 4B4DH

Address

@ (MSX only)

    This routine is used by the "READ/INPUT" statement handler if it has failed to convert a data item to numeric form. If in "READ" mode (FLGINP is non-zero) a "Syntax error" is generated (404FH). Otherwise the message "?Redo from start" is displayed (6678H) and control returns to the statement handler. Address ... 4B62H

Address

@ (MSX only)

    This is the "INPUT#" Statement handler. The buffer number is evaluated and PTRFIL set to direct input from the required I/O buffer (6D55H), control then transfers to the combined "READ/INPUT" statement handler (4B9BH). Address ... 4B6CH

Address

@ (MSX only)

    This is the "INPUT" statement handler. If the next program text character is a "#" control transfers to the "INPUT#" statement handler (4B62H). Otherwise the screen is forced to text mode, via the TOTXT standard routine, and any prompt string analyzed (6636H) and displayed (667BH). A question mark is then displayed and a line of text collected from the console via the QINLIN standard routine. If this returns Flag C (CTRL- STOP) control transfers to the "STOP" handler (63FEH). If the first character in BUF is zero (null input) the handler terminates by skipping to the end of the statement (485AH), otherwise control drops into the combined "READ/INPUT" handler. Address ... 4B9FH

Address

@ (MSX only)

    This is the "READ" statement handler, a large section is also used by the "INPUT" and "INPUT#" statements so the structure is rather awkward. Each Variable found in the program text is located in turn (PTRGET, 5EA4H), for each one the corresponding data item is obtained and assigned to the Variable by the "LET" handler (4893H). When in "READ" mode the data items are taken from the program text using the initial contents of DATPTR (4C40H). When in "INPUT" or "INPUT#" mode the data items are taken from the text buffer BUF.

    If the data items are exhausted in "READ" mode an "Out of DATA" error is generated. If they are exhausted in "INPUT" mode two question marks are displayed and another line fetched from the console via the QINLIN standard routine. If they are exhausted in "INPUT#" mode another line of text is copied to BUF from the relevant I/O buffer (6D83H). If the Variable list is exhausted while in "INPUT" mode the message "Extra ignored" is displayed (6678H) and the handler terminates (4AFFH). In "INPUT#" mode no message is displayed while in "READ" mode control terminates by updating DATPTR (63DEH). If a data item cannot be converted to numeric form (3299H) to match a numeric Variable control transfers to the "?Redo from start" routine (4B4DH). - 127 - 5. ROM BASIC INTERPRETER Address ... 4C2FH

Address

@ (MSX only)

    The is the plain text message "?Extra ignored", CR, LF terminated by a zero byte. Address ... 4C40H

Address

@ (MSX only)

    This routine is used by the "READ" handler to locate the next "DATA" statement in the program text, the address to start from is supplied in register pair HL. Each program statement is examined until a "DATA" token (84H) is found whereupon the routine terminates (4BD1H). If the end link is reached an "Out of DATA" error is generated. As the search proceeds the line number of each program line is placed in DATLIN for use by the error handler. Address ... 4C5FH

Address

@ (MSX only)

    This routine checks that the next character in the program text is the "=" token (EFH) and then drops into the Expression Evaluator. When entered at 4C62H it checks for "(".

    Address ... 4C64H
    Name ...... FRMEVL (Formula Evaluator)
    Input ..... HL points to current program expression
    Exit ...... HL=Next Address, DAC=Result, VALTYP=Result Type
    Modifies .. All registers

    This is the Expression Evaluator. On entry register pair HL points to the first character of the expression to be evaluated. On exit register pair HL points to the character following the expression, the result is in DAC and the type code in VALTYP. For a string result the address of the string descriptor is returned at DAC+2. The descriptor itself comprising a single byte for the string length and two bytes for its address, will be in TEMPST or inside a string Variable.

    An expression is a list of factors (4DC7H) linked together by operators with differing precedence levels. To process such an expression correctly the Expression Evaluator must be able to temporarily stack an intermediate result, if the next operator has a higher precedence than the current operator, and start afresh on a new calculation. It therefore has two basic operations, STACK and APPLY. For example:

        3+250\2^2*3^3+1,

        STACK:    3+   (\ follows)
        STACK:    250\ ( follows)
        APPLY:    2^2=4     (* follows)
        STACK:    4*   ( follows)
        APPLY:    3^3=27    (+ follows)
        APPLY:    4*27=108  (+ follows)
        APPLY:    250\108=2 (+ follows)
        APPLY:    3+2=5     (+ follows)
        APPLY:    5+1=6     (, follows)

    Evaluation terminates when the next operator has a precedence equal to or lower than the initial precedence and the stack is empty. The expression delimiter, shown as a comma in the example, is regarded as having a precedence of zero and so will always halt evaluation. Normally the Expression Evaluator starts off with an initial precedence of zero but the entry point at 4C67H may be used to supply an alternative value in register D. This facility is used by the Factor Evaluator to restrict the range of evaluation when applying the monadic negation and "NOT" operators. Address ... 4D22H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to apply an infix math operator (+-*/ ) to a pair of numeric operands. There are separate routines for the relational operators (4F57H) and the logical operators (4F78H). The first operand, its type code, and the operator token are supplied on the Z80 stack, the second operand and its type code are supplied in DAC and VALTYP. The types of both operands are first compared, if they differ the lowest precision operand is converted to match the higher. The operands are then moved to the positions required by the math routines. For integers the first operand is placed in register pair DE and the second in register pair HL. For single precision the first operand is placed in registers C, B, E, D and the second in DAC. For double precision the first operand is placed in DAC and the second in ARG. The operator token is then used to obtain the required address from the table at 3D51H, 3D5DH or 3D69H, depending upon the operand type, and control transfers to the relevant math routine. Address ... 4DB8H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to divide two integer operands. The first operand is contained in register pair DE and the second in register pair HL, the result is returned in DAC. Both operands are converted to single precision (2FCBH) and control transfers to the single precision division routine (3265H). Address ... 4DC7H

Address

@ (MSX only)

    This is the Factor Evaluator. On entry register pair HL points to the character before the factor to be evaluated. On exit register pair HL points to the character following the factor, the result is in DAC and the type code in VALTYP. A factor may be one of the following:

1. A numeric or string constant
2. A numeric or string Variable
3. A function
4. A monadic operator (+-NOT)
5. A parenthesized expression

    The first character is taken from the program text via the CHRGTR standard routine and examined. If it is an end of Statement character a "Missing operand" error is generated (406AH). If it is an ASCII digit it is converted from textual form to one of the standard numeric types in DAC (3299H).

    If it is upper case alphabetic (64A8H) it is a Variable and its current value is returned (4E9BH). If it is a numeric token the number is copied from CONLO to DAC (46B8H). If it is one of the FFH prefixed function tokens shown in the table at 39DEH it is decoded to transfer control to the relevant function handler (4ECFH). If it is the monadic "+" operator it is simply skipped over, only the monadic "-" operator (4E8DH) and monadic "NOT" operator (4F63H) require any action.

    If it is an opening quote the following explicit string is analyzed and a descriptor created (6636H). If it is an "&" it is a non-decimal numeric constant and it is converted to one of the standard numeric types in DAC (4EB8H). If it is not one of the functions shown below then it must be a parenthesized expression (4E87H), otherwise a "Syntax error" is generated. The following function tokens are tested for directly and control transferred to the address shown:

        ERR ....  4DFDH   ATTR$ .... 7C43H
        ERL ....  4E0BH   VARPTR ... 4E41H
        POINT ..  5803H   USR....... 4FD5H
        TIME ...  7900H   INSTR .... 68EBH
        SPRITE .  7A84H   INKEY$ ... 7347H
        VDP ....  7B47H   STRING$ .. 6829H
        BASE ...  7BCBH   INPUT$ ... 6C87H
        PLAY ...  791BH   CSRLIN ... 790AH
        DSKI$ ..  7C3EH   FN ....... 5040H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "ERR" function. The contents of ERRFLG are placed in DAC as an integer (4FCFH). Address ... 4E0BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "ERL" function. The contents of ERRLIN are copied to DAC as a single precision number (3236H). Address ... 4E41H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "VARPTR" function. If the function token is followed by a "#" the buffer number is evaluated (521BH), the I/O buffer FCB located (6A6DH) and its address placed in DAC as an integer (2F99H). Otherwise the Variable is located (5F5DH) and its address placed in DAC as an integer (2F99H). - 130 - 5. ROM BASIC INTERPRETER Address ... 4E8DH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the monadic "-" operator. Register D is set to a precedence value of 7DH, the factor evaluated (4C67H) and then negated (2E86H). Address ... 4E9BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to return the current value of a Variable. The Variable is first located (PTRGET, 5EA4H). If it is a string Variable its address is placed in DAC to point to the descriptor. Otherwise the contents of the Variable are copied to DAC (2F08). Address ... 4EA9H

Address

@ (MSX only)

    This routine returns the single character pointed to by register pair HL in register A, if it is a lower case alphabetic it converts it to upper case. Address ... 4EB8H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator and the numeric input routine (3299H) to convert an ampersand ("&") Prefixed number from textual form to an integer in DAC. As each legal character is found the product is multiplied by 2, 8 or 16, depending upon the character which initially followed the ampersand, and the new digit added to it. If the product overflows an "Overflow" error is generated (4067H). The routine terminates when an unacceptable character is found. Address ... 4EFCH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to process the FFH prefixed function tokens. If the token is either "LEFT$", "RIGHT$" or "MID$" the string operand is evaluated (4C62H), the address of its descriptor pushed onto the Z80 stack and the following numeric operand also evaluated (GETBYT, 521CH) and stacked. Otherwise the function's parenthesized operand is evaluated (4E87H) and, for "SQR", "RND", "SIN", "LOG", "EXP", "COS", "TAN" or "ATN" only, converted to double precision (303AH). The function token is then used to obtain the required address from the table at 39DEH and control transfers to the function handler. Address ... 4F47H

Address

@ (MSX only)

    This routine is used by the numeric input conversion routine (3299H) to test for a "+" or "-" character or token. It returns register D=0 for positive and register D=FFH for negative. Address ... 4F57H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to apply a relational operator (<>= or combinations) to a pair of operands. If the operands are numeric the Expression Evaluator first uses the math operator routine (4D22H) to apply the general relation operation to the operands. If the operands are strings the string comparison routine (65C8H) is used first. When control arrives here the relation result is in register A and the Z80 Flags:

Operand 1	=	Operand 2		A=00H (0), Flag Z, NC
Operand 1	<	Operand 2		A=01H (1), Flag NZ, NC
Operand 1	>	Operand 2		A=FFH (-1), Flag NZ, C

    The Expression Evaluator also supplies a bit mask defining the original operators on the Z80 stack. This has a 1 in each position if the associated operation is required: 00000<=>. The mask is applied to the relation result producing zero if none of the conditions is satisfied. This is then placed in DAC as a true (-1) or false (0) integer (2E9AH). Address ... 4F63H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the monadic "NOT" operator. Register D is set to an initial precedence level of 5AH and the expression evaluated (4C67H) and converted to an integer (2F8AH). It is then inverted and restored to DAC. Address ... 4F78H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to apply a logical operator ("OR", "AND", "XOR", "EQV" and "IMP") or the "MOD" and "\" operators to a pair of numeric operands. The first operand, which has already been converted to an integer, is supplied on the Z80 stack and the second is supplied in DAC. The operator token (actually its precedence level) is supplied in register B. After converting the second operand to an integer (2F8AH) the operator is examined. There are separate routines for "MOD" (323AH) and "\" (31E6H) but the logical operators are processed locally using the corresponding Z80 logical instructions on register pairs DE and HL. The result is stored in DAC as an integer (2F99H). Address ... 4FC7H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "LPOS" function to an operand contained in DAC. The contents of LPTPOS are placed in DAC as an integer (4FCFH). Address ... 4FCCH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "POS" function to an operand contained in DAC. The contents of TTYPOS are placed in DAC as an integer (2F99). - 132 - 5. ROM BASIC INTERPRETER Address ... 4FD5H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "USR" function. The user number is collected directly from the program text, it cannot be an expression, and the associated address taken from USRTAB (4FF4H). The following parenthesized operand is then evaluated (4E87H) and left in DAC as the passed parameter. If it is a string type its storage is freed (67D3H). The current program text position is pushed onto the Z80 stack followed by a return to 3297H, the routine at this address will restore the program text position after the user function has terminated. Control then transfers to the user address with register pair HL pointing to the first byte of DAC and the type code, from VALTYP, in register A. Additionally, for a string parameter, the descriptor address is taken from DAC and placed in register pair DE.

    The user routine may modify any register except the Z80 SP and should terminate with a RET instruction, interrupts may be left disabled if necessary as the Runloop will re-enable them. Any numeric parameter to be returned to the Interpreter should be placed in DAC. Strictly speaking this should be the same numeric type as the passed parameter, however if VALTYP is modified the Interpreter will always accept it.

    Returning a string type is more difficult. Using the same method as the Factor Evaluator string functions, which involves copying the string to the String Storage Area and pushing a new descriptor onto TEMPST, is complicated and vulnerable to changes in the MSX system. A simpler and more reliable method is to use the passed parameter to create the space for the result. This should not be an explicitly stated string as the program text will have to be modified, instead an implicit parameter should be used. This must be done with care however, it is very easy to gain the impression that the Interpreter has accepted the string when in fact it has not. Take the following example which does nothing but return the passed parameter: 10 POKE &H9000,&HC9 20 DEFUSR=&H9000 30 A$=USR(STRING$(12,"!")) 40 PRINT A$ 50 B$=STRING$(9,"X") 60 PRINT A$

    At first it seems that the passed string has been correctly assigned to A$. When line 60 is reached however it becomes apparent that A$ has been corrupted by the subsequent assignment of a string to B$. What has happened is that the temporary storage allocated to the passed parameter was reclaimed from the String Storage Area before control transferred to the user routine. This region was then used to store the string belonging to B$ thus modifying A$. - 133 - 5. ROM BASIC INTERPRETER

    This situation can be avoided by assigning the parameter to a Variable beforehand and then passing the Variable, for example:

        10 A$=STRING$(12,"!")
        20 A$=USR(A$)

    Line 10 results in twelve bytes of the String Storage Area being permanently allocated to A$. When the user function is entered the descriptor, which is pointed to by register pair DE, will contain the starting address of the twelve byte region where the result should be placed. If the returned string is shorter than the passed one the length byte of the descriptor may be changed without any side effects. For further details on string storage see the garbage collector (66B6H).

    A point worth noting is that a "CLEAR" operation is not strictly necessary before a machine language program is loaded. The region between the top of the Array Storage Area and the base of the Z80 stack is never used by the Interpreter. A program can exist in this region provided that the two enclosing areas do not overlap it. Address ... 500EH

Address

@ (MSX only)

    This is the "DEFUSR" statement handler. The user number is collected directly from the program text, it cannot be an expression, and the associated entry in USRTAB located (4FF4H). The address operand is then evaluated (FRMQNT, 542FH) and placed in USRTAB. Address ... 501DH

Address

@ (MSX only)

    This is the "DEF FN" and "DEFUSR" statement handler. If the following character is a "USR" token (DDH) control transfers to the "DEFUSR" statement handler (500EH), otherwise the program text is checked for a trailing "FN" token (DEH). The function name Variable is located (51A1H) and, after checking that the Interpreter is in program mode (5193H), the current program text position is placed there. Each of the Variables in the formal parameter list is then located in succession (PTRGET, 5EA4H), this is simply to ensure that they are created. The routine terminates by skipping over the remainder of the statement (485BH) as the function body is not required at this time. Address ... 5040H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "FN" function. The function name Variable is first located (51A1H) to obtain the address of the function definition in the program text. Each formal Variable from the function definition is located in turn (PTRGET, 5EA4H) and its address pushed onto the Z80 stack. As each one is found the corresponding actual parameter is evaluated (4C64H) and pushed onto the stack with it. If necessary the type of the actual parameter is converted to match that of the formal parameter (517AH).

    When both lists are exhausted each formal Variable address and actual parameter are popped from the stack in turn. Each Variable is then copied from the Variable Storage Area to PARM2 with its value replaced by the actual parameter. It should be noted that, because PARM2 is only a hundred bytes long, a maximum of nine double precision parameters is allowed. When all the actual parameters have been copied to PARM2 the entire contents of PARM1 (the current parameter area) are pushed onto the Z80 stack and PARM2 is copied to PARM1 (518EH). Register pair HL is then set to the start of the function body in the program text and the expression is evaluated (4C5FH). The old contents of PARM1 are popped from the stack and restored. Finally the result of the evaluation is type converted if necessary to match the function name type (517AH).

    A user defined function differs from a normal expression in only one respect, it has its own set of local Variables. These Variables are created in PARM1 when the function is invoked and disappear when it terminates. When a normal Variable search is initiated by the Expression Evaluator the region examined is the Variable Storage Area. However, if NOFUNS is non-zero, indicating at least one active user function, PARM1 will be searched instead, only if this fails will the search move on to the global Variables in the Variable Storage Area. Using a local Variable area specific to each invocation of a function means that the same Variable names can be used throughout without the Variables overwriting each other or the global Variables.

    It is worth noting that a user defined function is slower than an inline expression or even a subroutine. The search carried out to find the function name Variable, plus the large amount of stacking and destacking, are significant overheads. Address ... 5189H

Address

@ (MSX only)

    This routine moves a block of memory from the address pointed to by register pair DE to that pointed to by register pair HL, register pair BC defines the length. Address ... 5193H

Address

@ (MSX only)

    This routine generate an "Illegal direct" error if CURLIN shows the Interpreter to be in direct mode. Address ... 51A1H

Address

@ (MSX only)

    This routine checks the program text for an "FN" token (DEH) and then creates the function name Variable (5EA9H). These are distinguished from ordinary Variables by having bit 7 set in the first character of the Variable's name. - 135 - 5. ROM BASIC INTERPRETER Address ... 51ADH

Address

@ (MSX only)

    Control transfers to this routine from the Runloop execution point (4640H) if a token greater than D8H is found at the start of a statement. If the token is not an FFH prefixed function token a "Syntax error" is generated (4055H). If the function token is one of those which double as statements control transfers to the relevant handler, otherwise a "Syntax error" is generated. The statements in question are "MID$" (696EH), "STRIG" (77BFH) and "INTERVAL" (77B1H). There is actually no separate token for "INTERVAL", the "INT" token (85H) suffices with the remaining characters being checked by the statement handler. Address ... 51C9H

Address

@ (MSX only)

    This is the "WIDTH" statement handler. The operand is evaluated (GETBYT, 521CH) and its magnitude checked. If it is zero or greater than thirty-two or forty, depending upon the screen mode held in OLDSCR an "Illegal function call" error is generated (475AH). If it is the same as the current contents of LINLEN the routine terminates with no further action. Otherwise the current screen is cleared with a FORMFEED control code (0CH) via the OUTDO standard routine in case the screen is to be made smaller. The operand is then placed in LINLEN and either LINL32 or LINL40, depending upon the screen mode held in OLDSCR, and the screen cleared again in case it has been made larger. Because the line length variable to be changed is selected by OLDSCR, rather than SCRMOD, the width can still be changed even if the screen is currently in Graphics Mode or Multicolour Mode. In this case the change is effective when a return is made to the Interpreter Mainloop or an "INPUT" statement is executed. Address ... 520EH

Address

@ (MSX only)

    This routine evaluates the next expression in the program text (4C64H), converts it to an integer (2F8AH) and places the result in register pair DE. The magnitude and sign of the MSB are then tested and the routine terminates. Address ... 521BH

Address

@ (MSX only)

    This routine evaluates the next operand in the program text (4C64H) and converts it to an integer (5212H). If the operand is greater than 255 an "Illegal function call" error is generated (475AH).

    Address ... 521CH
    Name ...... GETBYT (Get Byte Parameter)
    Input ..... HL points to current program expression
    Exit ...... HL=Next Address, A=E=Byte Read
    Modifies .. All registers

    This routine evaluates the next operand in the program text to a byte datum. - 136 - 5. ROM BASIC INTERPRETER Address ... 5229H

Address

@ (MSX only)

    This is the "LLIST" statement handler. PRTFLG is set to 01H, to direct output to the printer, and control drops into the "LIST" statement handler. Address ... 522EH

Address

@ (MSX only)

    This is the "LIST" statement handler. The optional start and termination line number operands are collected and the starting position found in the program text (4279H). Successive program lines are listed until the end link is found, the CTRL-STOP key is pressed or the termination line number is reached, control then transfers directly to the Mainloop "OK" point (411FH). Each program line is listed by displaying its line number (3412H), detokenizing (5284H) and displaying (527BH) the line itself and then issuing a CR,LF (7328H). Address ... 5284H

Address

@ (MSX only)

    This routine is used by the "LIST" statement handler to convert a tokenized program line to textual form. On entry register pair HL points to the first character of the tokenized line. On exit the line of text is in BUF and is terminated by a zero byte.

    Any normal or FFH prefixed token is converted to the corresponding keyword by a simple linear search of the tokens in the table at 3A72H. Exceptions are made if either an opening quote character, a "REM" token, or a "DATA" token has previously been found. Normally these tokens will be followed by plain text anyway, the check is made to stop graphics characters being interpreted as tokens. The three byte sequence ":" (3AH), "REM" token (8FH), " " token (E6H) is converted to the single " " character (27H) and the statement separator (3AH) preceding an "ELSE" token (A1H) is scrubbed out.

    If one of the numeric tokens is found its value and type are first copied from CONLO and CONTYP to DAC and VALTYP (46E8H). It is then converted to textual form in FBUFFR by the decimal (3425H), octal (371EH) or hex (3722H) conversion routines. For octal and hex types the number is prefixed by an ampersand and an "O" or "H" letter. A type suffix, "'" or "#", is added to single precision or double precision numbers only if there is no decimal part and no exponent part ("E" or "D"). Address ... 53E2H

Address

@ (MSX only)

    This is the "DELETE" statement handler. The optional start and termination line number operands are collected and the starting position found in the program text (4279H). If any pointers exist in the program text they are converted back to line numbers (54EAH). The terminating program line is found by a search of the program text (4295H), if this address is smaller than that of the starting program line an "Illegal function call" error is generated (475AH), otherwise the message "OK" is displayed (6678H). The block of memory from the end of the terminating line to the start of the Variable Storage Area is copied down to the beginning of the starting line and VARTAB, ARYTAB and STREND are reset to the new (lower) end of the program text. Control then transfers directly to the end of the Mainloop (4237H) to reset the remaining pointers and to relink the Program Text Area. Note that, because control does not return to the normal "OK" point, the screen will not be returned to text mode. If the screen is in Graphics Mode or Multicolour mode when a "DELETE" is executed, which is admittedly rather unlikely, the system will crash. Address ... 541CH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "PEEK" function to an operand contained in DAC. The address operand is checked (5439H) then the byte read from memory and placed in DAC as an integer (4FCFH). Address ... 5423H

Address

@ (MSX only)

    This is the "POKE" statement handler. The address operand is evaluated (FRMQNT, 542FH) then the data operand evaluated (GETBYT, 521CH) and written to memory.

    Address ... 542FH
    Name ...... FRMQNT (Formula Quantificator)
    Input ..... HL points to current program expression
    Exit ...... HL=Next Address, DE=Integer Datum
    Modifies .. All registers

    This routine evaluates the next operand in the program text (4C64H) and places it in register pair DE as an integer (5439H). Address ... 5439H

Address

@ (MSX only)

    This routine converts the numeric operand contained in DAC into an integer in register pair HL. The operand must be in the range -32768 to +65535 and is normally an address as required by "POKE", "PEEK", "BLOAD", etc. The operand's type is first checked via the GETYPR standard routine, if it is already an integer it is simply placed in register pair HL (2F8AH). Assuming the operand is single precision or double precision its sign is checked, if it is negative it is converted to integer (2F8AH). Otherwise it is converted to single precision (2FB2H) and its magnitude checked (2F21H). If it is greater than 32767 and smaller than 65536 then -65536 is added (324EH) before it is converted to integer (2F8AH). Address ... 5468H

Address

@ (MSX only)

    This is the "RENUM" statement handler. If a new initial line number operand exists it is collected (475FH), otherwise a default value of ten of taken. If an old initial line number operand exists it is collected (475FH), otherwise a default value of zero is taken. If an increment line number operand exists it is collected (4769H), otherwise a default value of ten is taken.

    The program text is then searched for existing line numbers equal to or greater than the new initial line number (4295H) and the old initial line number (4295H), an "Illegal function call" error is generated (475AH) if the new address is smaller than the old address. This is to catch any attempt to renumber high program lines down to existing low ones.

    A dummy renumbering run of the program text is first carried out to check than no new line number will be generated with a value greater than 65529. This must be done as an error midway through the conversion would leave the program text in a confused state. Assuming all is well any line number operands in the program text are converted to pointers (54F6H). This neatly solves the problem of line number references, GOTO 50 for example, as the program text is not moved during renumbering. Starting at the old initial program text position each existing program line number is replaced with its new value. When the end link is reached any program text pointers are converted back to line number operands (54F1H) and control transfers directly to the Mainloop "OK" point (411EH). Address ... 54F6H

Address

@ (MSX only)

    When entered at 54F6H this routine converts every line number operand in the program text to a pointer. When entered at 54F7H it performs the reverse operation and converts every pointer in the program text back to a line number operand. Starting at the beginning of the Program Text Area each line is examined for a pointer token (0DH) or a line number operand token (0EH) depending upon the mode. In pointer to line number operand mode the pointer is replaced by the line number from the referenced program line and the token changed to 0EH. In line number operand to pointer mode the program text is searched (4295H) to find the relevant line, its address replaces the line number operand and the token is changed to 0DH. If the search is unsuccessful a message of the form "Undefined line NNNN in NNNN" is displayed (6678H) and the conversion process continues. A special check is made for the "ON ERROR GOTO 0" statement, to prevent the generation of a spurious error message, but no check is made for the similar "RESUME 0" statement. In this case an error message will be displayed, this should be ignored. Address ... 555AH

Address

@ (MSX only)

    This is the plain text message "Undefined line " terminated by a zero byte.

    Address ... 558CH (from 0008H)
    Name ...... SYNCHR (Syntax check Character)
    Input ..... HL points to character to check
    Exit ...... A=Next program character
    Modifies .. AF, HL

    Standard routine to check the current program text character, whose address is supplied in register pair HL, against a reference character. The reference character is supplied as a single byte immediately following the CALL or RST instruction, for example:

        RST 08H
        DEFB ","

    If the characters do not match a "Syntax error" is generated (4055H), otherwise control transfers to the CHRGTR standard routine to fetch the next program character (4666H).

    Address ... 5597H (from 0028H)
    Name ...... GETYPR (Get Type into Register A)
    Input ..... None
    Exit ...... AF=Type
    Modifies .. AF

    Standard routine to return the type of the current operand, determined by VALTYP, as follows:

Integer ............ A=FFH, Flag M, NZ, C
String ............. A=00H, Flag P, Z, C
Single Precision ... A=01H, Flag P, NZ, C
Double Precision ... A=05H, Flag P, NZ, NC

Address

@ (MSX only)

    This is the "CALL" statement handler. The extended statement name, which is an unquoted string up to fifteen characters long terminated by a "(", ":" or end of line character (00H), is first copied from the program text to PROCNM, any unused bytes are zero filled. Bit 5 of each entry in SLTATR is then examined for an extension ROM with a statement handler. If a suitable ROM is found its position in SLTATR is converted to a Slot ID in register A and a ROM base address in register H (7E2AH). The statement handler address is read from ROM locations four and five (7E1AH) and placed in register pair IX. The Slot ID is placed in the high byte of register pair IY and the ROM statement handler called via the CALSLT standard routine.

    The ROM will examine the statement name and return Flag C if it does not recognize it, otherwise it performs the required operation. If the ROM call fails the search of SLTATR continues until the table is exhausted whereupon a "Syntax error" is generated (4055H). If the ROM call is successful the handler terminates. Address ... 55F8H

Address

@ (MSX only)

    This routine is used by the device name parser (6F15H) when it cannot recognize a device name found in the program text. Upon entry register pair HL points to the first character of the name and register B holds its length. The name is first copied to PROCNM and terminated by a zero byte. Bit 6 of each entry in SLTATR is then examined for an extension ROM with a device handler. If a suitable ROM is found its position in SLTATR is converted to a Slot ID in register A and a ROM base address in register H (7E2AH). The device handler address is read from ROM locations six and seven (7E1AH) and placed in register pair IX. The Slot ID is placed in the high byte of register pair IY, the unknown device code (FFH) in register A and the ROM device handler called via the CALSLT standard routine.

    The ROM will examine the device name and return Flag C if it does not recognize it, otherwise it returns its own internal code from zero to three. If the ROM call fails the search of SLTATR continues until the table is exhausted whereupon a "Bad file name" error is generated (6E6BH). If the ROM call is successful the ROM's internal code is added to its SLTATR position, multiplied by a factor of four, to produce a global device code. The base code for each entry in SLTATR is shown below in hexadecimal. The "SS" and "PS" markers show the corresponding Secondary and Primary Slot numbers, each slot is composed of four pages:

          SS0           SS1           SS2           SS3
    +-------------+-------------+-------------+-------------+
    | 00 04 08 0C | 10 14 18 1C | 20 24 28 2C | 30 34 38 3C | PS0
    +-------------+-------------+-------------+-------------+
    | 40 44 48 4C | 50 54 58 5C | 60 64 68 6C | 70 74 78 7C | PS1
    +-------------+-------------+-------------+-------------+
    | 80 84 88 8C | 90 94 98 9C | A0 A4 A8 AC | B0 B4 B8 BC | PS2
    +-------------+-------------+-------------+-------------+
    | C0 C4 C8 CC | D0 D4 D8 DC | E0 E4 E8 EC | F0 F4 F8 FC | PS3
    +-------------+-------------+-------------+-------------+

Figure 44: Device Codes.

    The global device code is used by the Interpreter until the time comes for the ROM to perform an actual device operation. It is then converted back into the ROM's Slot ID, base address and internal device code to perform the ROM access. Note that the codes from 0 to 8 are reserved for disk drive identifiers and those from FCH to FFH for the standard devices GRP, CRT, LPT and CAS. With the current MSX hardware structure these codes correspond to physically improbable ROM configurations and are therefore safe to be used for specific purposes by the Interpreter. Address ... 564AH

Address

@ (MSX only)

    This routine is used by the function dispatcher (6F8FH) when it encounters a device code not belonging to one of the standard devices. The device code is first converted to a SLTATR position and then to a Slot ID in register A and ROM base address in register H (7E2DH). The ROM device handler address is read from ROM locations six and seven (7E1AH) and placed in register pair IX. The Slot ID is placed in the high byte of register pair IY, the ROM's internal device code in DEVICE and the ROM device handler called via the CALSLT standard routine. Address ... 566CH

Address

@ (MSX only)

    This entry point to the macro language parser is used by the "DRAW" statement handler, a later entry point (56A2H) is used by the "PLAY" statement handler. The command string is evaluated (4C64H) and its storage freed (FRESTR, 67D0H). After pushing a zero termination block onto the Z80 stack the length and address of the string body are placed in MCLLEN and MCLPTR and control drops into the parser mainloop. Address ... 56A2H

Address

@ (MSX only)

    This is the macro language parser mainloop, it is used to process the command string associated with a "DRAW" or "PLAY" statement. On entry the string length is in MCLLEN, the string address is in MCLPTR and the address of the relevant command table is in MCLTAB. The command tables contain the legal command letters, together with the associated command handler addresses, for each statement. The "DRAW" table is at 5D83H and the "PLAY" table at 752EH.

    The parser mainloop first fetches the next character from the command string (56EEH). If there are no more characters left the next string descriptor is popped from the stack (568CH). If this is zero the parser terminates (5709H) if MCLFLG shows a "DRAW" statement to be active, otherwise control transfers back to the "PLAY" statement handler (7494H).

    Assuming a command character exists the current command table is searched to check its legality, if no match is found an "Illegal function call" error is generated (475AH). The command table entry is then examined, if bit 7 is set the command takes an optional numeric parameter. If this is present it is collected and placed in register pair DE (571CH), otherwise a default value of one is taken. After pushing a return to the start of the parser mainloop onto the Z80 stack control transfers to the command handler at the address taken from the command table. Address ... 56EEH

Address

@ (MSX only)

    This routine is used by the macro language parser to fetch the next character from the command string. If MCLLEN is zero the routine terminates with Flag Z, there are no characters left. Otherwise the next character is taken from the address contained in MCLPTR and returned in register A, if the character is lower case it is converted to upper case. MCLPTR is then incremented and MCLLEN decrement Ed. - 142 - 5. ROM BASIC INTERPRETER Address ... 570BH

Address

@ (MSX only)

    This routine is used by the macro language parser to return an unwanted character to the command string. MCLLEN is incremented and MCLPTR decremented. Address ... 5719H

Address

@ (MSX only)

    This routine is used by the macro language parser to collect a numeric parameter from the command string. The result is a signed integer and is returned in register pair DE, it cannot be an expression. The first character is taken and examined, if it is a "+" it is ignored and the next character taken (5719H). If it is a "-" a return is set up to the negation routine (5795H) and the next character taken (5719H). If it is an "=" the value of the following Variable is returned (577AH). Otherwise successive characters are taken and a binary product accumulated until a non-numeric character is found. Address ... 575AH

Address

@ (MSX only)

    This routine is used by the macro language parser "=" and "X" handlers. The Variable name is copied to BUF until the ";" delimiter is found, if this takes more than thirty-nine characters to find an "Illegal function call" error is generated (475AH). Otherwise control transfers to the Factor Evaluator Variable handler (4E9BH) and the Variable contents are returned in DAC. Address ... 577AH

Address

@ (MSX only)

    This routine is used by the macro language parser to process the "=" character in a command parameter. The Variable's value is obtained (575AH), converted to an integer (2F8AH) and placed in register pair DE. Address ... 5782H

Address

@ (MSX only)

    This routine is used by the macro language parser to process the "X" command. The Variable is processed (575AH) and, after checking that stack space is available (625EH), the current contents of MCLLEN and MCLPTR are stacked. Control then transfers to the parser entry point (5679H) to obtain the Variable's descriptor and process the new command string. Address ... 579CH

Address

@ (MSX only)

    This routine is used by various graphics statements to evaluate a coordinate pair in the program text. The coordinates must be parenthesized with a comma separating the component operands. If the coordinate pair is preceded by a "STEP" token (DCH) each component value is added to the corresponding component of the current graphics coordinates in GRPACX and GRPACY, otherwise the absolute values are returned. The X coordinate is returned in GRPACX, GXPOS and register pair BC. The Y coordinate is returned in GRPACY, GYPOS and register pair DE.

    There are two entry points to the routine, the one which is used depends on whether the caller is expecting more than one coordinate pair. The "LINE" statement, for example, expects two coordinate pairs the first of which is the more flexible. The entry point at 579CH is used to collect the first coordinate pair and will accept the characters "-" or "@-" as representing the current graphics coordinates. The entry point at 57ABH is used for the second coordinate pair and requires an explicit operand. Address ... 57E5H

Address

@ (MSX only)

    This is the "PRESET" statement handler. The current background colour is taken from BAKCLR and control drops into the "PSET" handler. Address ... 57EAH

Address

@ (MSX only)

    This is the "PSET" statement handler. After the coordinate pair has been evaluated (57ABH) the current foreground colour is taken from FORCLR and used as the default when setting the ink colour (5850H). The current graphics coordinates are converted to a physical address, via the SCALXY and MAPXYC standard routines, and the colour of the current pixel set via the SETC standard routine. Address ... 5803H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "POINT" function. The current contents of CLOC, CMASK, GYPOS, GXPOS, GRPACY and GRPACX are stacked and the coordinate pair operand evaluated (57ABH). The colour of the new pixel is read via the SCALXY, MAPXYC and READC standard routines and placed in DAC as an integer (2F99H), the old coordinate values are then popped and restored. Note that a value of -1 is returned if the point coordinates are outside the screen. Address ... 5850H

Address

@ (MSX only)

    This graphics routine is used to evaluate an optional colour operand in the program text and to make it the current ink colour. After checking the screen mode (59BCH) the colour operand is evaluated (GETBYT, 521CH) and placed in ATRBYT. If no operand exists the colour code supplied in register A is placed in ATRBYT instead. Address ... 5871H

Address

@ (MSX only)

    This graphics routine returns the difference between the contents of GXPOS and register pair BC in register pair HL. If the result is negative (GXPOS<BC) it is negated to produce the absolute magnitude and Flag C is returned. Address ... 5883H

Address

@ (MSX only)

    This graphics routine returns the difference between the contents of GYPOS and register pair DE in register pair HL. If the result is negative (GYPOS<DE) it is negated to produce the absolute magnitude and Flag C is returned. Address ... 588EH

Address

@ (MSX only)

    This graphics routine swaps the contents of GYPOS and register pair DE. Address ... 5898H

Address

@ (MSX only)

    This graphics routine first swaps the contents of GYPOS and register pair DE (588EH) then swaps the contents of GXPOS and register pair BC. When entered at 589BH only the second operation is performed. Address ... 58A7H

Address

@ (MSX only)

    This is the "LINE" statement handler. The first coordinate pair (X1,Y1) is evaluated (579CH) and placed in register pairs BC,DE. After checking for the "-" token (F2H) the second coordinate pair (X2,Y2) is evaluated (57ABH) and left in GRPACX, GRPACY and GXPOS, GYPOS. After setting the ink colour (584DH) the program text is checked for a following "B" or "BF" option and either the box (5912H), boxfill (58BFH) or linedraw (58FCH) operation performed. None of these operations affects the current graphics coordinates in GRPACX and GRPACY, these are left at X2,Y2. Address ... 58BFH

Address

@ (MSX only)

    This routine performs the boxfill operation. Given that the supplied coordinate pairs define diagonally opposed points of the box two quantities must be derived from them. The horizontal size of the box is obtained from the difference between X1 and X2, this gives the number of pixels to set per row. The vertical size is obtained from the difference between Y1 and Y2 giving the number of rows required. Starting at the physical address of X1,Y1, and moving successively lower via the DOWNC standard routine, the required number of pixel rows are filled in by repeated use of the NSETCX standard routine. Address ... 58FCH

Address

@ (MSX only)

    This routine performs the linedraw operation. After drawing the line (593CH) GXPOS and GYPOS are reset to X2,Y2 from GRPACX and GRPACY. - 145 - 5. ROM BASIC INTERPRETER Address ... 5912H

Address

@ (MSX only)

    This routine performs the box operation. The box is produced by drawing a line (58FCH) between each of the four corner points. The coordinates of each corner are derived from the initial operands by interchanging the relevant component of the pair. The drawing sequence is:

1. X1,Y2 to X2,Y2
2. X1,Y1 to X2,Y1
3. X2,Y1 to X2,Y2
4. X1,Y1 to X1,Y2

Address

@ (MSX only)

    This routine draws a line between the points X1,Y1, supplied in register pairs BC and DE and X2,Y2, supplied in GXPOS and GYPOS. The operation of the drawing mainloop (5993H) is best illustrated by an example, say LINE(0,0)-(10,4). To reach the end point of the line from its start ten horizontal steps (X2- X1) and four downward steps (Y2-Y1) must be taken altogether. The best approximation to a straight line therefore requires two and a half horizontal steps for every downward step (X2- X1/Y2-Y1). While this is impossible in practice, as only integral steps can be taken, the correct ratio can be achieved on average.

    The method employed is to add the Y difference to a counter each time a rightward step is taken. When the counter exceeds the value of the X difference it is reset and one downward step is taken, this is in effect an integer division of the two difference values. Sometimes downward steps will be produced every two rightward steps and sometimes every three rightward steps. The average, however, will be one downward step every two and a half rightward steps. An equivalent BASIC program is shown below with a slightly offset BASIC line for comparison:

        10 SCREEN 0
        20 INPUT"START X,Y";X1,Y1
        30 INPUT"END X,Y",X2,Y2
        40 SCREEN 2
        50 X=X1:Y=Y1:L=X2-X1:S=Y2-Y1:CTR=L/2
        60 PSET(X,Y)
        70 CTR=CTR+S:IF CTR<L THEN 90
        80 CTR=CTR-L:Y=Y+1
        90 X=X+1:IF X<=X2 THEN 60
        100 LINE(X1,Y1+5)-(X2,Y2+5)
        110 GOTO 110

    The above example suffers from three limitations. The line must slope downwards, it must slope to the right and the slope cannot exceed forty-five degrees from the horizontal (one downward step for one rightward step). - 146 - 5. ROM BASIC INTERPRETER

    The routine overcomes the first limitation by examining the Y1 and Y2 coordinates before drawing commences. If Y2 is greater than or equal to Y1, showing the line to slope upwards or to be horizontal, both coordinate pairs are exchanged. The line is now sloping downwards and will be drawn from the end point to the start.

    The second limitation is overcome by examining X1 and X2 beforehand to determine which way the line is sloping. If X2 is greater than or equal to X1 the line slopes to the right and a Z80 JP to the RIGHTC standard routine is placed in MINUPD/MAXUPD (see below) for use by the drawing mainloop, otherwise a JP to the LEFTC standard routine is placed there.

    The third limitation is overcome by comparing the X coordinate difference to the Y coordinate difference before drawing to determine the slope steepness. If X2-X1 is smaller than Y2-Y1 the slope of the line is less than forty-five degrees from the horizontal. The simple method shown above for LINE(0,0)-(10,4) will not work for slopes greater than forty- five degrees as the maximum rate of descent is achieved when one downward step is taken for every horizontal step. It will work however if the step directions are exchanged. Thus LINE(0,0)-(4,10) requires one rightward step for every two and a half downward steps. MINUPD holds a Z80 JP to the "normal" step direction standard routine for the drawing mainloop and MAXUPD holds a JP to the "slope" step direction standard routine. For shallow angles MINUPD will vector to DOWNC and MAXUPD to LEFTC or RIGHTC. For steep angles MINUPD will vector to LEFTC or RIGHTC and MAXUPD to DOWNC. For steep angles the counter values must also be exchanged, the X difference must now be added to the counter and the Y difference used as the counter limit. The variables MINDEL and MAXDEL are used by the drawing mainloop to hold these counter values, MINDEL holds the smaller end point difference and MAXDEL the larger.

    An interesting point is that the reference counter, held in CTR in the above program and in register pair DE in the ROM, is preloaded with half the largest end point difference rather than being set to zero. This has the effect of splitting the first "stair" in the line into two sections, one at the start of the line and one at its end, and improving the line's appearance. Address ... 59B4H

Address

@ (MSX only)

    This graphics routine shifts the contents of register pair DE one bit to the right. Address ... 59BCH

Address

@ (MSX only)

    This routine generates an "Illegal function call" error (475AH) if the screen is not in Graphics Mode or Multicolour Mode. - 147 - 5. ROM BASIC INTERPRETER Address ... 59C5H

Address

@ (MSX only)

    This is the "PAINT" statement handler. The starting coordinate pair is evaluated (579CH), the ink colour set (584DH) and the optional boundary colour operand evaluated (GETBYT, 521CH) and placed in BDRATR. The starting coordinate pair is checked to ensure that it is within the screen (5E91H) and is made the current pixel physical address by the MAPXYC standard routine. The distance to the right hand boundary is then measured (5ADCH) and, if it is zero, the handler terminates. Otherwise the distance to the left hand boundary is measured (5AEDH) and the sum of the two placed in register pair DE as the zone width. The current position is then stacked twice (5ACEH), first with a termination flag (00H) and then with a down direction flag (40H). Control then transfers to the paint mainloop (5A26H) with an up direction flag (C0H) in register B. Address ... 5A26H

Address

@ (MSX only)

    This is the paint mainloop. The zone width is held in register pair DE, the paint direction, up or down, in register B and the current pixel physical address is that of the pixel adjacent to the left hand boundary. A vertical step is taken to the next line, via the TUPC or TDOWNC standard routines, and the distance to the right hand boundary measured (5ADCH). The distance to the left hand boundary is then measured and the line between the boundaries filled in (5AEDH). If no change is found in the position of either boundary control transfers to the start of the mainloop to continue painting in the same direction. If a change is found an inflection has occurred and the appropriate action must be taken.

    There are four types of inflection, LH or RH incursive, where the relevant boundary moves inward, and LH or RH excursive, where it moves outward. An example of each type is shown below with numbered zones indicating the order of painting during upward movement. A secondary zone is shown within each inflective region for completeness:

             +-+   +-+            +----------+   +----------+
             | |   | |            |     3    |   |    3     |
        +-+  |2|   |3|  +-+       |          |   |          |
        |3|  | |   | |  |2|       +---+ +--+ |   | +--+ +---+
    +---+ +--+ |   | +--+ +---+       |2|  | |   | |  |2|
    |          |   |          |       +-+  |1|   |1|  +-+
    |     1    |   |    1     |            | |   | |
    +----------+   +----------+            +-+   +-+
    LH Incursion   RH Incursion   LH Excursion   RH Excursion

Figure 45: Boundary Inflections.

    A LH excursion has occurred when the distance to the left hand boundary is non-zero, a RH excursion has occurred when the current zone width is greater than that of the previous line. Unless the excursion is less than two pixels, in which case it will be ignored, the current position (the bottom left of zone 3 in Figure 45) is stacked (5AC2H), the paint direction reversed and painting restarts at the top left of the excursive region .

    A RH incursion has occurred when the current zone width is smaller than that of the previous line. If the incursion is total, that is the current zone width is zero, a dead end has been reached and the last position and direction are popped (5AIFH) and painting restarts at that point. Otherwise the current position and direction are stacked (5AC2H) and painting restarts at the bottom left of the incursive region.

    A LH incursion is dealt with automatically during the search for the right hand boundary and requires no explicit action by the paint mainloop. Address ... 5AC2H

Address

@ (MSX only)

    This routine is used by the "PAINT" statement handler to save the current paint position and direction on the Z80 stack. The six byte parameter block is made up of the following:

        2 bytes ... Current contents of CLOC
        1 byte  ... Current direction
        1 byte  ... Current contents of CMASK
        2 bytes ... Current zone width

    After the parameters have been stacked a check is made that sufficient stack space still exists (625EH). Address ... 5ADCH

Address

@ (MSX only)

    This routine is used by the "PAINT" statement handler to locate the right hand boundary. The zone width of the previous line is passed to the SCANR standard routine in register pair DE, this determines the maximum number of boundary colour pixels that may initially be skipped over. The returned skip count remainder is placed in SKPCNT and the number of non- boundary colour pixels traversed in MOVCNT. Address ... 5AEDH

Address

@ (MSX only)

    This routine is used by the "PAINT" statement handler to locate the left hand boundary. The end point of the right hand boundary search is temporarily saved and the starting point taken from CSAVEA and CSAVEM and made the current pixel physical address. The left hand boundary is then located via the SCANL standard routine, which also fills in the entire zone, and the right hand end point recovered and placed in CSAVEA and CSAVEM. - 149 - 5. ROM BASIC INTERPRETER Address ... 5B0BH

Address

@ (MSX only)

    This routine is used by the "CIRCLE" statement handler to negate the contents of register pair DE. Address ... 5B11H

Address

@ (MSX only)

    This is the "CIRCLE" statement handler. After evaluating the centre coordinate pair (579CH) the radius is evaluated (520FH), multiplied (325CH) by SIN(PI/4) and placed in CNPNTS. The ink colour is set (584DH), the start angle evaluated (5D17H) and placed in CSTCNT and the end angle evaluated (5D17H) and placed in CENCNT. If the end angle is smaller than the start angle the two values are swapped and CPLOTF is made non-zero. The aspect ratio is evaluated (4C64H) and, if it is greater than one, its reciprocal is taken (3267H) and CSCLXY is made non-zero to indicate an X axis squash. The aspect ratio is multiplied (325CH) by 256, converted to an integer (2F8AH) and placed in ASPECT as a single byte binary fraction. Register pairs HL and DE are set to the starting position on the circle perimeter (X=RADIUS,Y=0) and control drops into the circle mainloop. Address ... 5BBDH

Address

@ (MSX only)

    This is the circle mainloop. Because of the high degree of symmetry in a circle it is only necessary to compute the coordinates of the arc from zero to forty-five degrees. The other seven segments are produced by rotation and reflection of these points. The parametric equation for a unit circle, with T the angle from zero to PI/4, is:

        X=COS(T)
        Y=SIN(T)

    Direct computation using this equation, or the corresponding functional form X=SQR(1-Y^2), is too slow, instead the first derivative is used:

         dx
        ---- = -Y/X
         dy

    Given that the starting position is known (X=RADIUS,Y=0), the X coordinate change for each unit Y coordinate change may be computed using the derivative. Furthermore, because graphics resolution is limited to one pixel, it is only necessary to know when the sum of the X coordinate changes reaches unity and then to decrement the X coordinate. Therefore:

        Decrement X when (Y1/X)+(Y2/X)+(Y3/X)+... ⇒ 1
        Therefore decrement when (Y1+Y2+Y3+...)/X ⇒ 1
        Therefore decrement when     Y1+Y2+Y3+... ⇒ X

    All that is required to identify an X coordinate change is to totalize the Y coordinate values from each step until the X coordinate value is exceeded. The circle mainloop holds the X coordinate in register pair HL, the Y coordinate in register pair DE and the running total in CRCSUM. An equivalent BASIC program for a circle of arbitrary radius 160 pixels is:

        10 SCREEN 2
        20 X=160:Y=0:CRCSUM=0
        30 PSET(X,191-Y)
        40 CRCSUM=CRCSUM+Y :Y=Y+1
        50 IF CRCSUM<X THEN 30
        60 CRCSUM=CRCSUM-X:X=X-1
        70 IF X>Y THEN 30
        80 CIRCLE(0,191),155
        90 GOTO 90

    The coordinate pairs generated by the mainloop are those of a "virtual" circle, such tasks as axial reflection, elliptic squash and centre translation are handled at a lower level (5C06H). Address ... 5C06H

Address

@ (MSX only)

    This routine is used to by the circle mainloop to convert a coordinate pair, in register pairs HL and DE, into eight symmetric points on the screen. The Y coordinate is initially negated (5B0BH), reflecting it about the X axis, and the first four points produced by successive clockwise rotations through ninety degrees (5C48H). The Y coordinate is then negated again (5B0BH) and a further four points produced (5C48H).

    Clockwise rotation is performed by exchanging the X and Y coordinates and negating the new Y coordinate, thus a point (40,10) would become (10,-40). Assuming an aspect ratio of 0.5, for example, the complete sequence of eight points would therefore be:

1. X,-Y*0.5
2. -Y,-X*0.5
3. -X, Y*0.5
4. Y, X*0.5
5. Y,-X*0.5
6. -X,-Y*0.5
7. -Y, X*0.5
8. X, Y*0.5

    It can be seen from the above that, ignoring the sign of the coordinates for the moment, there are only four terms involved. Therefore, rather than performing the relatively slow aspect ratio multiplication (5CEBH) for each point, the terms X*0.5 and Y*0.5 can be prepared in advance and the complete sequence generated by interchanging and negating the four terms. With the aspect ratio shown above the initial conditions are set up so that register pair HL=X, register pair DE= -Y*0.5, CXOFF=Y and CYOFF=X*0.5 and successive points are produced by the operations:

1. Exchange HL and CXOFF, negate HL.
2. Exchange DE and CYOFF, negate DE.

    In parallel with the computation of each circle coordinate the number of points required to reach the start of the segment containing the point is kept in CPCNT8. This will initially be zero and will increase by 2*RADIUS*SIN(PI/4) as each ninety degree rotation is made. As each of the eight points is produced its Y coordinate value is added to the contents of CPCNT8 and compared to the start and end angles to determine the appropriate course of action. If the point is between the two angles and CPLOTF is zero, or if it is outside the angles and CPLOTF is non-zero, the coordinates are added to the circle centre coordinates (5CDCH) and the point set via the SCALXY, MAPXYC and SETC standard routines. If the point is equal to either of the two angles, and the associated bit is set in CLINEF, the coordinates are added to the circle centre coordinates (5CDCH) and a line drawn to the centre (593CH). If none of these conditions is applicable no action is taken other than to proceed to the next point. Address ... 5CEBH

Address

@ (MSX only)

    This routine multiplies the coordinate value supplied in register pair DE by the aspect ratio contained in ASPECT, the result is returned in register pair DE. The standard binary shift and add method is used but the operation is performed as two single byte multiplications to avoid overflow problems. Address ... 5D17H

Address

@ (MSX only)

    This routine is used by the "CIRCLE" statement handler to convert an angle operand to the form required by the circle mainloop, the result is returned in register pair DE. While the method used is basically sound, and eliminates one trigonometric computation per angle, the results produced are inaccurate. This is demonstrated by the following example which draws a line to the true thirty degree point on a circle's perimeter:

        10 SCREEN 2
        20 PI = 4 * ATN(1)
        30 CIRCLE(100,100),80,,PI/6
        40 LINE(100,100)-(100+80*COS(PI/6),100-80*SIN(PI/6))
        50 GOTO 50

    The result that the routine should produce is the number of points that must be produced by the circle mainloop before the required angle is reached. This can be computed by first noting that there will be INT(ANGLE/(PI/4)) forty-five degree segments prior to the segment containing the required angle. Furthermore each forty-five segment will contain RADIUS*SIN(PI/4) points as this is the value of the terminating Y coordinate. Therefore the number of points required to reach the start of the segment containing the angle is the product of these two numbers. The total count is produced by adding this figure to the number of points required to cover any remaining angle within the final segment, that is RADIUS*SIN(REMAINING ANGLE) points.

    Unfortunately the routine computes the number of points within a segment by linear approximation from the total segment size on the mistaken assumption that successive points subtend equal angles. Thus in the above example the point count computed for the angle is 30/45*(80*0.707107)=37 instead of the correct value of forty. The error produced by the routine is therefore at a maximum at the centre of each forty-five degree segment and reduces to zero at the end points. Address ... 5D6EH

Address

@ (MSX only)

    This is the "DRAW" statement handler. Register pair DE is set to point to the command table at 5D83H and control transfers to the macro language parser (566CH). Address ... 5D83H

Address

@ (MSX only)

    This table contains the valid command letters and associated addresses for the "DRAW" statement commands. Those commands which takes a parameter, and consequently have bit 7 set in the table, are shown with an asterisk:

        CMD TO
        -----------
        U*  5DB1H
        D*  5DB4H
        L*  5DB9H
        R*  5DBCH
        M   5DD8H
        E*  5DCAH
        F*  5DC6H
        G*  5DD1H
        H*  5DC3H
        A*  5E4EH
        B   5E46H
        N   5E42H
        X   5782H
        C*  5E87H
        S*  5E59H

Address

@ (MSX only)

    This is the "DRAW" statement "U" command handler. The operation of the "D", "L", "R", "E", "F", "G" and "H" commands is very similar so no separate description of their handlers is given. The optional numeric parameter is supplied by the macro language parser in register pair DE. This initial parameter is modified by a given handler into a horizontal offset in register pair BC and a vertical offset in register pair DE. For example if leftward or upward movement is required the parameter is negated (5B0BH), if diagonal movement is required the parameter is duplicated so that equal horizontal and vertical offsets are produced. Once the offsets have been prepared control transfers to the line drawing routine (5DFFH). Address ... 5DD8H

Address

@ (MSX only)

    This is the "DRAW" statement "M" command handler. The character following the command letter is examined then the two parameters collected from the command string (5719H). If the initial character is "+" or "-" the parameters are regarded as offsets and are scaled (5E66H), rotated through successive ninety degree steps as determined by DRWANG and then added to the current graphics coordinates (5CDCH) to determine the termination point. If DRWFLG shows the "B" mode to be inactive a line is then drawn (5CCDH) from the current graphics coordinates to the termination point. If DRWFLG shows the "N" mode to be inactive the termination coordinates are placed in GRPACX and GRPACY to become the new current graphics coordinates. Finally DRWFLG is zeroed, turning the "B" and "N" modes off, and the handler terminates. Address ... 5E42H

Address

@ (MSX only)

    This is the "DRAW" statement "N" command handler, DRWFLG is simply set to 40H. Address ... 5E46H

Address

@ (MSX only)

    This is the "DRAW" statement "B" command handler, DRWFLG is simply set to 80H. Address ... 5E4EH

Address

@ (MSX only)

    This is the "DRAW" statement "A" command handler. The parameter is checked for magnitude and placed in DRWANG. Address ... 5E59H

Address

@ (MSX only)

    This is the "DRAW" statement "S" command handler. The parameter is checked for magnitude and placed in DRWSCL. Address ... 5E66H

Address

@ (MSX only)

    This routine is used by the "DRAW" statement "U", "D", "L", "R", "E", "F", "G", "H" and "M" (in offset mode) command handlers to scale the offset supplied in register pair DE by the contents of DRWSCL. Unless DRWSCL is zero, in which case the routine simply terminates, the offset is multiplied using repeated addition and then divided by four (59B4H). To eliminate scaling an "S0" or "S4" command should be used. - 154 - 5. ROM BASIC INTERPRETER Address ... 5E87H

Address

@ (MSX only)

    This is the "DRAW" statement "C" command handler. The parameter is placed in ATRBYT via the SETATR standard routine. There is no check on the MSB of the parameter so illegal values such as "C265" will be accepted without an error message. Address ... 5E91H

Address

@ (MSX only)

    This routine is used by the "PAINT" statement handler to check, via the SCALXY standard routine, that the coordinates in register pairs BC and DE are within the screen. If not an "Illegal function call" error is generated (475AH). Address ... 5E9FH

Address

@ (MSX only)

    This is the "DIM" statement handler. A return is set up to 5E9AH, so that multiple Arrays can be processed, DIMFLG is made non-zero and control drops into the Variable search routine.

    Address ... 5EA4H
    Name ...... PTRGET (Pointer Get)
    Input ..... HL=program address, SUBFLG=0 single variable else array
    Exit ...... HL=next address, DE=pointer to the first variable
    Modifies .. All registers

    This is the Variable search routine. On entry register pair HL points to the first character of the Variable name in the program text. On exit register pair HL points to the character following the name and register pair DE to the first byte of the Variable contents in the Variable Storage Area. The first character of the name is taken from the program text, checked to ensure that it is upper case alphabetic (64A7H) and placed in register C. The optional second character, with a default value of zero, is placed in register B, this character may be alphabetic or numeric. Any further alphanumeric characters are then simply skipped over. If a type suffix character ("%", "$", "!" or "#") follows the name this is converted to the corresponding type code (2, 3, 4 or 8) and placed in VALTYP. Otherwise the Variable's default type is taken from DEFTBL using the first letter of the name to locate the appropriate entry.

    SUBFLG is then checked to determine how any parenthesized subscript following the name should be treated. This flag is normally zero but is modified by the "ERASE" (01H), "FOR" (64H), "FN" (80H) or "DEF FN" (80H) statement handlers to force a particular course of action. In the "ERASE" case control transfers straight to the Array search routine (5FE8H), no parenthesized subscript need be present. In the "FOR", "FN" and "DEF FN" cases control transfers straight to the simple Variable search routine (5F08H), no check is made for a parenthesized subscript. Assuming that the situation is normal the program text is checked for the characters "(" or "[". If either is present control transfers to the Array search routine (5FBAH), otherwise control drops into the simple Variable search routine. - 155 - 5. ROM BASIC INTERPRETER Address ... 5F08H

Address

@ (MSX only)

    This is the simple Variable search routine. There are four types of simple Variable each composed of a header followed by the Variable contents. The first byte of the header contains the type code and the next two bytes the Variable name. The contents of the Variable will be one of the three standard numeric forms or, for the string type, the length and address of the string. Each of the four types is shown below:

    +-----+-----+-----+-----+-----+
    | 02H | "A" | "B" | LSB | MSB |
    +-----+-----+-----+-----+-----+
    Integer

    +-----+-----+-----+-----+-----+-----+
    | 03H | "A" | "B" | LEN | LSB | MSB |
    +-----+-----+-----+-----+-----+-----+
    String

    +-----+-----+-----+----+----+----+----+
    | 04H | "A" | "B" | EE | DD | DD | DD |
    +-----+-----+-----+----+----+----+----+
    Single Precision

    +-----+-----+-----+----+----+----+----+----+----+----+----+
    | 08H | "A" | "B" | EE | DD | DD | DD | DD | DD | DD | DD |
    +-----+-----+-----+----+----+----+----+----+----+----+----+
    Double Precision

Figure 46: Simple Variables.

    NOFUNS is first checked to determine whether a user defined function is currently being evaluated. If so the search is carried out on the contents of PARM1 first of all, only if this fails will it move onto the main Variable Storage Area. A linear search method is used, the two name characters and type byte of each Variable in the storage area are compared to the reference characters and type until a match is found or the end of the storage area is reached. If the search is successful the routine terminates with the address of the first byte of the Variable contents in register pair DE. If the search is unsuccessful the Array Storage Area is moved upwards and the new Variable is added to the end of the existing ones and initialized to zero.

    There are two exceptions to this automatic creation of a new Variable. If the search is being carried out by the "VARPTR" function, and this is determined by examining the return address, no Variable will be created. Instead the routine terminates with register pair DE set to zero (5F61H) causing a subsequent "Illegal function call" error. The second exception occurs when the search is being carried out by the Factor Evaluator, that is when the Variable is newly declared inside an expression. In this case DAC is zeroed for numeric types, and loaded with the address of a dummy zero length descriptor for a string type, thus returning a zero result (5FA7H). These actions are designed to prevent the Expression Evaluator creating a new Variable ("VARPTR") is the only function to take a Variable argument directly rather than via an expression and so requires separate protection). If this were not so then assignment to an Array, via the "LET" statement handler, would fail as any simple Variable created during expression evaluation would change the Array's address. Address ... 5FBAH

Address

@ (MSX only)

    This is the Array search routine. There are four types of Array each composed of a header plus a number of elements. The first byte of the header contains the type code, the next two bytes the Array name and the next two the offset to the start of the following Array. This is followed by a single byte containing the dimensionality of the Array and the element count list. Each two byte element count contains the maximum number of elements per dimension. These are stored in reverse order with the first one corresponding to the last subscript. The contents of each Array element are identical to the contents of the corresponding simple Variable. The integer Array AB%(3,4) is shown below with each element identified by its subscripts, high memory is towards the top of the page:

    +-----------------------+
    |(0,4) (1,4) (2,4) (3,4)|
    |(0,3) (1,3) (2,3) (3,3)|
    |(0,2) (1,2) (2,2) (3,2)|
    |(0,1) (1,1) (2,1) (3,1)|
    |(0,0) (1,0) (2,0) (3,0)|
    +-----------------------+

    +-----+-----+-----+-----------+-----+-----------+-----------+
    |     |     |     |  Offset   | Dim |   Count   |   Count   |
    | 02H | "A" | "B" | 2DH   00H | 02H | 05H   00H | 04H   00H |
    +-----+-----+-----+-----------+-----+-----------+-----------+

Figure 47: Integer Array.

    Each subscript is evaluated, converted to an integer (4755H) and pushed onto the Z80 stack until a closing parenthesis is found, it need not match the opening one. A linear search is then carried out on the Array Storage Area for a match with the two name characters and the type. If the search is successful DIMFLG is checked and a "Redimensioned array" error generated (405EH) if it shows a "DIM" statement to be active. Unless an "ERASE" statement is active, in which case the routine terminates with register pair BC pointing to the start of the Array (3297H), the dimensionality of the Array is then checked against the subscript count and a "Subscript out of range" error generated if they fail to match. Assuming these tests are passed control transfers to the element address computation point (607DH).

    If the search is unsuccessful and an "ERASE" statement is active an "Illegal function call" error is generated (475AH), otherwise the new Array is added to the end of the existing Array Storage Area. Initialization of the new Array proceeds by storing the two name characters, the type code and the dimensionality (the subscript count) followed by the element count for each dimension. If DIMFLG shows a "DIM" statement to be active the element counts are determined by the subscripts. If the Array is being created by default, with a statement such as "A(1,2,3)=5" for example, a default value of eleven is used. As each element count is stored the total size of the Array is accumulated in register pair DE by successive multiplications (314AH) of the element counts and the element size (the Array type). After a check that this amount of memory is available (6267H) STREND is increased the new area is zeroed and the Array size is stored, in slightly modified form, immediately after the two name characters. Unless the Array is being created by default, in which case the element address must be computed, the routine then terminates.

    This is the element address computation point of the Array search routine. The location of a particular element within an Array involves the multiplication (314AH) of subscripts, element counts and element sizes. As there are a variety of ways this could be done the actual method used is best illustrated with an example. The location of element (1,2,3) in a 4*5*6 Array would initially be computed as (((3*5)+2)*4)+1. This is then multiplied by the element size (type) and added to the Array base address to obtain the address of the required element. The computation method is an optimized form which minimizes the number of steps needed, it is equivalent to evaluating (3*(4*5))+(2*4)+(1). The element address is returned in register pair DE. Address ... 60B1H

Address

@ (MSX only)

    This is the "PRINT USING" statement handler. Control transfers here from the general "PRINT" statement handler after the applicable output device has been set up. Upon termination control passes back to the general "PRINT" statement exit point (4AFFH) to restore the normal video output. The format string is evaluated (4C65H) and the address and length of the string body obtained from the descriptor. The program text pointer is then temporarily saved. Each character of the format string is examined until one of the possible template characters is found. If the character does not belong in a template it is simply output via the OUTDO standard routine. Once the start of a template is found this is scanned along until a non-template character is found. Control then passes to the numeric output routine (6192H) or the string output routine (6211H).

    In either case the program text pointer is restored to register pair HL and the next operand evaluated (4C64H). For numeric output the information gained from the template scan is passed to the numeric conversion routine (3426H) in registers A, B and C and the resulting string displayed (6678H). For string output the required character count is passed to the "LEFT$" statement handler (6868H) in register C and the resulting string displayed (667BH). For either type of output the program text and format string are then examined to determine whether there are any further characters. If no operands exist the handler terminates. If the format string has been exhausted then it is restarted from the beginning (60BFH), otherwise scanning continues from the current position for the next operand (60f6H). Address ... 6250H

Address

@ (MSX only)

    This routine is used by the Interpreter Mainloop and the Variable search routine to move a block of memory upwards. A check is first made to ensure that sufficient memory exists (6267H) and then the block of memory is moved. The top source address is supplied in register pair BC and the top destination address in register pair HL. Copying stops when the contents of register pair BC equal those of register pair DE. Address ... 625EH

Address

@ (MSX only)

    This routine is used to check that sufficient memory is available between the top of the Array Storage Area and the base of the Z80 stack. On entry register C contains the number of words the caller requires. If this would narrow the gap to less than two hundred bytes an "Out of memory" error is generated. Address ... 6286H

Address

@ (MSX only)

    This is the "NEW" statement handler. TRCFLG, AUTFLG and PTRFLG are zeroed and the zero end link is placed at the start of the Program Text Area. VARTAB is set to point to the byte following the end link and control drops into the run-clear routine. Address ... 629AH

Address

@ (MSX only)

    This routine is used by the "NEW", "RUN" and "CLEAR" statement handlers to initialize the Interpreter variables. All interrupts are cleared (636EH) and the default Variable types in DEFTBL set to double precision. RNDX is reset (2C24H) and ONEFLG, ONELIN and OLDTXT are zeroed. MEMSIZ is copied to FRETOP to clear the String Storage Area and DATPTR set to the start of the Program Text Area (63C9H). The contents of VARTAB are copied into ARYTAB and STREND, to clear any Variables, all the I/O buffers are closed (6C1CH) and NLONLY is reset. SAVSTK and the Z80 SP are reset from STKTOP and TEMPPT is reset to the start of TEMPST to clear any string descriptors. The printer is shut down (7304H) and output restored to the screen (4AFFH). Finally PRMLEN, NOFUNS, PRMLN2, FUNACT, PRMSTK and SUBFLG are zeroed and the routine terminates. - 159 - 5. ROM BASIC INTERPRETER Address ... 631BH

Address

@ (MSX only)

    This routine is used by the "DEVICE ON" statement handlers to enable an interrupt source, the address of the relevant device's TRPTBI. status byte is supplied in register pair HL. Interrupts are enabled by setting bit 0 of the status byte. Bits 1 and 2 are then examined and, if the device has been stopped and an interrupt has occurred, ONGSBF is incremented (634FH) so that the Runloop will process it at the end of the statement. Finally bit 1 of the status byte is reset to release any existing stop condition. Address ... 632EH

Address

@ (MSX only)

    This routine is used by the "DEVICE OFF" statement handlers to disable an interrupt source, the address of the relevant device's TRPTBL status byte is supplied in register pair HL. Bits 0 and 2 are examined to determine whether an interrupt has occurred since the end of the last statement, if so ONGSBF is decremented (6362H) to prevent the Runloop from picking it up. The status byte is then zeroed. Address ... 6331H

Address

@ (MSX only)

    This routine is used by the "DEVICE STOP" statement handlers to suspend processing of interrupts from an interrupt source, the address of the relevant device's TRPTBL status byte is supplied in register pair HL. Bits 0 and 2 are examined to determine whether an interrupt has occurred since the end of the last statement, if so ONGSBF is decremented (6362H) to prevent the Runloop from picking it up. Bit 1 of the status byte is then set. Address ... 633EH

Address

@ (MSX only)

    This routine is used by the "RETURN" statement handler to release the temporary stop condition imposed during interrupt driven BASIC subroutines, the address of the relevant device's TRPTBL status byte is supplied in register pair HL. Bits 0, and 2 are examined to determine whether a stopped interrupt has occurred since the subroutine was first activated. If so ONGSBF is incremented (634FH) so that the Runloop will pick it up at the end of the statement. Bit 1 of the status byte is then reset. It should be noted that any "DEVICE STOP" Statement within an interrupt driven subroutine will therefore be ineffective. Address ... 6358H

Address

@ (MSX only)

    This routine is used by the Runloop interrupt processor (6389H) to clear an interrupt prior to activating the BASIC subroutine, the address of the relevant device's TRPTBL status byte is supplied in register pair HL. ONGSBF is decremented and bit 2 of the status byte is reset. - 160 - 5. ROM BASIC INTERPRETER Address ... 636EH

Address

@ (MSX only)

    This routine is used by the run-clear routine (629AH) to clear all interrupts. The seventy-eight bytes of TRPTBL and the ten bytes of FNKFLG are zeroed. Address ... 6389H

Address

@ (MSX only)

    This is the Runloop interrupt processor. ONEFLG is first examined to determine whether an error condition currently exists. If so the routine terminates, no interrupts will be processed until the error clears. CURLIN is then examined and, if the Interpreter is in direct mode, the routine terminates. Assuming all is well a search is made of the twenty-six status bytes in TRPTBL to find the first active interrupt. Note that devices near the start of the table will consequently have a higher priority than those lower down. When the first active status byte is found, that is one with bits 0 and 2 set, the associated address is taken from TRPTBL and placed in register pair DE. The interrupt is then cleared (6358H) and the device stopped (6331H) before control transfers to the "GOSUB" handler (47CFH). Address ... 63C9H

Address

@ (MSX only)

    This is the "RESTORE" statement handler. If no line number operand exists DATPTR is set to the start of the Program Storage Area. Otherwise the operand is collected (4769H), the program text searched to find the relevant line (4295H) and its address placed in DATPTR. Address ... 63E3H

Address

@ (MSX only)

    This is the "STOP" statement handler. If further text exists in the statement control transfers to the "STOP ON/OFF/STOP" statement handler (77A5H). Otherwise register A is set to 01H and control drops into the "END" statement handler. Address ... 63EAH

Address

@ (MSX only)

    This is the "END" statement handler. It is also used, with differing entry points, by the "STOP" statement and for CTRL- STOP and end of text program termination. ONEFLG is first zeroed and then, for the "END" statement only, all I/O buffers are closed (6C1CH). The current program text position is placed in SAVTXT and OLDTXT and the current line number in OLDLIN for use by any subsequent "CONT" statement. The printer is shut down (7304H), a CR LF issued to the screen (7323H) and register pair HL set to point to the "Break" message at 3FDCH. For the "END" statement and end of text cases control then transfers to the Mainloop "OK" point (411EH). For the CTRL-STOP case control transfers to the end of the error handler (40FDH) to display the "Break" message. - 161 - 5. ROM BASIC INTERPRETER Address ... 6424H

Address

@ (MSX only)

    This is the "CONT" statement handler. Unless they are zero, in which case a "Can't CONTINUE" error is generated, the contents of OLDTXT are placed in register pair HL and those of OLDLIN in CURLIN. Control then returns to the Runloop to execute at the old program text position. A program cannot be continued after CTRL-STOP has been used to break from WITHIN a statement, via the CKCNTC standard routine, rather than from between statements. Address ... 6438H

Address

@ (MSX only)

    This is the "TRON" statement handler, TRCFLG is simply made non-zero. Address ... 6439H

Address

@ (MSX only)

    This is the "TROFF" statement handler, TRCFLG is simply made zero. Address ... 643EH

Address

@ (MSX only)

    This is the "SWAP" statement handler. The first Variable is located (PTRGET, 5EA4H) and its contents copied to SWPTMP. The location of this Variable and of the end of the Variable Storage Area are temporarily saved. The second Variable is then located (PTRGET, 5EA4H) and its type compared with that of the first. If the types fail to match a "Type mismatch" error is generated (406DH). The current end of the Variable Storage Area is then compared with the old end and an "Illegal function call" error generated (475AH) if they differ. Finally the contents of the second Variable are copied to the location of the first Variable (2EF3H) and the contents of SWPTMP to the location of the second Variable (2EF3H).

    The checks performed by the handler mean that the second Variable, if it is simple and not an Array, must always be in existence before a "SWAP" Statement is encountered or an error will be generated. The reason for this is that, supposing the first Variable was an Array, then the creation of a second (simple) Variable would move the Array Storage Area upwards invalidating its saved location. Note that the perfectly legal case of a simple first Variable and a newly created simple second Variable is also rejected. Address ... 6477H

Address

@ (MSX only)

    This is the "ERASE" statement handler. SUBFLG is first set to 01H, to control the Variable search routine, and the Array located (PTRGET, 5EA4H). All the following Arrays are moved downward and STREND set to its new, lower value. The program text is then checked and, if a comma follows, control transfers back to the start of the handler. - 162 - 5. ROM BASIC INTERPRETER Address ... 64A7H

Address

@ (MSX only)

    This routine checks whether the character whose address is supplied in register pair HL is upper case alphabetic, if so it returns Flag NC. Address ... 64AFH

Address

@ (MSX only)

    This is the "CLEAR" statement handler. If no operands are present control transfers to the run-clear routine (62A1H) to remove all current Variables. Otherwise the string space operand is evaluated (4756H) followed by the optional top of memory operand (FRMQNT, 542FH). The top of memory value is checked and an "Illegal function call" error generated (475AH) if it is less than 8000H or greater than F380H. The space required by the I/O buffers (267 bytes each) and the String Storage Area is subtracted from the top of memory value and an "Out of memory" error generated (6275H) if there is less than 160 bytes remaining to the base of the Variable Storage Area. Assuming all is well HIMEM, MEMSIZ and STKTOP are set to their new values and the remaining storage pointers reset via the run- clear routine (62A1H). The I/O buffer storage is re-allocated (7E6BH) and the handler terminates.

    Unfortunately the computation of MEMSIZ and STKTOP, when a new top of memory is specified, is incorrect resulting in the top of the String Storage Area being set one byte too high. This can be seen with the following where an illegal string is accepted:

        10 CLEAR 200,&HF380
        20 A$=STRING$(201,"A")
        30 PRINT FRE("")

    Because there should be an extra DEC HL instruction at 64EBH the new values of MEMSIZ and STKTOP are initially set one byte too high. When the run-clear routine is called MEMSIZ is copied into FRETOP, the top of the String Storage Area, which results in this being one byte too high as well. Although MEMSIZ and STKTOP are correctly recomputed when the file pointers are reset, FRETOP is left with its incorrect value. When the "FRE" statement is executed in line thirty, and string garbage collection initiated, FRETOP is restored to its correct value but, because the string overflows the String Storage Area by one byte, the amount of free space displayed is -1 byte. To correctly set all the system pointers any alteration of the top of memory should be followed immediately by another "CLEAR" statement with no operands. Address ... 6520H

Address

@ (MSX only)

    This routine computes the difference between the contents of register pairs HL and DE. It is a duplicate of the short section of code from 64ECH to 64F1H and is completely unused. - 163 - 5. ROM BASIC INTERPRETER Address ... 6527H

Address

@ (MSX only)

    This is the "NEXT" statement handler. Assuming further text is present in the statement the loop Variable is located (PTRGET, 5EA4H), otherwise a default address of zero is taken. The stack is then searched for the corresponding "FOR" parameter block (3FE2H). If no parameter block is found, or if a "GOSUB" parameter block is found first, a "NEXT without FOR" error is generated (405BH). Assuming the parameter block is found the intervening section of stack, together with any "FOR" blocks it may contain, is discarded. The loop Variable type is then taken from the parameter block and examined to determine the precision required during subsequent operations.

    The STEP value is taken from the parameter block and added (3172H, 324EH or 2697H) to the current contents of the loop Variable which is then updated. The new value is compared (2F4DH, 2F21H or 2F5CH) with the termination value from the parameter block to determine whether the loop has terminated (65B6H). The loop will terminate for a positive STEP if the new loop value is GREATER than the termination value. The loop will terminate for a negative step if the new loop value is LESS than the termination value. If the loop has not terminated the original program text position and line number are taken from the parameter block and control transfers to the Runloop (45FDH). If the loop has terminated the parameter block is discarded from the stack and, unless further program text is present in which control transfers back to the start of the handler, control transfers to the Runloop to execute the next statement (NEWSTT, 4601H). Address ... 65C8H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to find the relation (<>=) between two string operands. The address of the first string descriptor is supplied on the Z80 stack and the address of the second in DAC. The result is returned in register A and the flags as for the numeric relation routines:

        String 1=String 2 ... A=00H, Flag Z, NC
        String 1<String 2 ... A=01H, Flag NZ, NC
        String 1>String 2 ... A=FFH, Flag NZ, C

    Comparison commences at the first character of each string and continues until the two characters differ or one of the strings is exhausted. Control then returns to the Expression Evaluator (4F57H) to place the true or false numeric result in DAC. Address ... 65F5H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "OCT$" function to an operand contained in DAC. The number is first converted to textual form in FBUFFR (371EH) and then the result string is created (6607H). - 164 - 5. ROM BASIC INTERPRETER Address ... 65FAH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "HEX$" function to an operand contained in DAC. The number is first converted to textual form in FBUFFR (3722H) and then the result string is created (6607H). Address ... 65FFH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "BIN$" function to an operand contained in DAC. The number is first converted to textual form in FBUFFR (371AH) and then the result string is created (6607H). Address ... 6604H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "STR$" function to an operand contained in DAC. The number is first converted to textual form in FBUFFR (3425H) then analyzed to determine its length and address (6635H). After checking that sufficient space is available (668EH) the string is copied to the String Storage Area (67C7H) and the result descriptor created (6654H). Address ... 6627H

Address

@ (MSX only)

    This routine first checks that there is sufficient space in the String Storage Area for the string whose length is supplied in register A (668EH). The string length and the address where the string will be placed in the String Storage Area are then copied to DSCTMP. Address ... 6636H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to analyze the character string whose address is supplied in register pair HL. The character string is scanned until a terminating character (00H or ") is found. The length and starting address are then placed in DSCTMP (662AH) and control drops into the descriptor creation routine. Address ... 6654H

Address

@ (MSX only)

    This routine is used by the string functions to create a result descriptor. The descriptor is copied from DSCTMP to the next available position in TEMPST and its address placed in DAC. Unless TEMPST is full, in which case a "String formula too complex" error is generated, TEMPPT is increased by three bytes and the routine terminates. Address ... 6678H

Address

@ (MSX only)

    This routine displays the message, or string, whose address is supplied in register pair HL. The string is analyzed (6635H) and its storage freed (67D3H). Successive characters are then taken from the string and displayed, via the OUTDO standard routine, until the string is exhausted. Address ... 668EH

Address

@ (MSX only)

    This routine checks that there is room in the String Storage Area to add the string whose length is supplied in register A. On exit register pair DE points to the starting address in the String Storage Area where the string should be placed. The length of the string is first subtracted from the current free location contained in FRETOP. This is then compared with STKTOP, the lowest allowable location for string storage, to determine whether there is space for the string. If so FRETOP is updated with the new position and the routine terminates. If there is insufficient space for the string then garbage collection is initiated (66B6H) to try and eliminate any dead strings. If, after garbage collection, there is still not enough space an "Out of string space" error is generated. Address ... 66B6H

Address

@ (MSX only)

    This is the string garbage collector, its function is to eliminate any dead strings from the String Storage Area. The basic problem with string Variables, as opposed to numeric ones, is that their lengths vary. If string bodies were stored with their Variables in the Variable Storage Area even such apparently simple statements as A$=A$+"X" would require the movement of thousands of bytes of memory and slow execution speeds dramatically. The method used by the Interpreter to overcome this problem is to keep the string bodies separate from the Variables. Thus strings are kept in the String Storage Area and each Variable holds a three byte descriptor containing the length and address of the associated string. Whenever a string is assigned to a Variable it is simply added to the heap of existing strings in the String Storage Area and the Variable's descriptor changed. No attempt is made to eliminate any previous string belonging to the Variable, by restructuring the heap, as this would wipe out any throughput gains.

    If sufficient Variable assignments are made it is inevitable that the String Storage Area will fill up. In a typical program many of these strings will be unused, that is the result of previous assignments. Garbage collection is the process whereby these dead strings are removed. Every string Variable in memory, including Arrays and the local Variables present during evaluation of user defined functions, is examined until the one is found whose string is stored highest in the heap. This string is then moved to the top of the String Storage Area and the Variable contents modified to point to the new location. The owner of the next highest string is then found and the process repeated until every string belonging to a Variable has been compacted. - 166 - 5. ROM BASIC INTERPRETER

    If a large number of Variables are present garbage collection may take an appreciable time. The process can be seen at work with the following program which repeatedly assigns the string "AAAA" to each element of the Array A$. The program will run at full speed for the first two hundred and fifty assignments and then pause to eliminate the fifty dead strings. A further fifty assignments can then be made before a further garbage collection is required: 10 CLEAR 1000 20 DIM A$(200) 30 FOR N=0 TO 200 40 A$(N)=STRING$(4,"A") 50 PRINT"."; 60 NEXT N 70 GOTO 30

    The String Storage Area is also used to hold the intermediate strings produced during expression evaluation. Because so many string functions take multiple arguments, "MID$" takes three for example, the management of intermediate results is a major problem. To deal with it a standardized approach to string results is taken throughout the Interpreter. A producer of a string simply adds the string body to the heap in the String Storage Area, adds the descriptor to the descriptor heap in TEMPST and places the address of the descriptor in DAC. It is up to the user of the result to free this storage (FRESTR, 67D0H) once it has processed the string. This rule applies to all parts of the system, from the individual function handlers back through the Expression Evaluator to the statement handlers, with only two exceptions.

    The first exception occurs when the Factor Evaluator finds an explicitly stated string, such as "SOMETHING" in the program text. In this case it is not necessary to copy the string to the String Storage Area as the original will suffice.

    The second exception occurs when the Factor Evaluator finds a reference to a Variable. In this case it is not necessary to place a copy of the descriptor in TEMPST as one already exists inside the Variable. Address ... 6787H

Address

@ (MSX only)

    This routine is used by the Expression Evaluator to concatenate two string operands. Control transfers here when a "+" token is found following a string operand so the first action taken is to fetch the second string operand via the Factor Evaluator (4DC7H). The lengths are then taken from both string descriptors and added together to check the length of the combined string. If this is greater than two hundred and fifty-five characters a "String too long" error is generated. After checking that space is available in the String Storage Area (6627H) the storage of both operands is freed (67D6H). The first string is then copied to the String Storage Area (67BFH) and followed by the second one (67BFH). The result descriptor is created (6654H) and control transfers back to the Expression Evaluator (4C73H).

    Address ... 67D0H
    Name ...... FRESTR (Free String)
    Input ..... DAC=String Descriptor, VALTYP=3 (String)
    Exit ...... HL=pointer to the string descriptor
    Modifies .. All registers

    This routine frees any storage occupied by the string whose descriptor address is contained in DAC. The address of the descriptor is taken from DAC and examined to determine whether it is that of the last descriptor in TEMPST (67EEH), if not the routine terminates. Otherwise TEMPPT is reduced by three bytes clearing this descriptor from TEMPST. The address of the string body is then taken from the descriptor and compared with FRETOP to see if this is the lowest string in the String Storage Area, if not the routine terminates. Otherwise the length of the string is added to FRETOP, which is then updated with this new value, freeing the storage occupied by the string body. Address ... 67FFH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "LEN" function to an operand contained in DAC. The operand's storage is freed (FRESTR, 67D0H) and the string length taken from the descriptor and placed in DAC as an integer (4FCFH). Address ... 680BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "ASC" function to an operand contained in DAC. The operand's storage is freed and the string length examined (6803H), if it is zero an "Illegal function call" error is generated (475AH). Otherwise the first character is. taken from the string and placed in DAC as an integer (4FCFH). Address ... 681BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CHR$" function to an operand contained in DAC. After checking that sufficient space is available (6625H) the operand is converted to a single byte integer (521FH). This character is then placed in the String Storage Area and the result descriptor created (6654H). Address ... 6829H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "STRING$" function. After checking for the open parenthesis character the length operand is evaluated and placed in register E (GETBYT, 521CH). The second operand is then evaluated (4C64H). If it is numeric it is converted to a single byte integer (521FH) and placed in register A. If it is a string the first character is taken from it and placed in register A (680FH). Control then drops into the "SPACE$" function to create the result string. - 168 - 5. ROM BASIC INTERPRETER Address ... 6848H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "SPACE$" function to an operand contained in DAC. The operand is first converted to a single byte integer in register E (521FH). After checking that sufficient space is available (6627H) the required number of spaces are copied to the String Storage Area and the result descriptor created (6654H). Address ... 6861H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "LEFT$" function. The first operand's descriptor address and the integer second operand are supplied on the Z80 stack. The slice size is taken from the stack (68E3H) and compared to the source string length. If the source string length is less than the slice size it replaces it as the length to extract. After checking that sufficient space is available (668EH) the required number of characters are copied from the start of the source string to the String Storage Area (67C7H). The source string's storage is then freed (67D7H) and the result descriptor created (6654H). Address ... 6891H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "RIGHT$" function. The first operand's descriptor address and the integer second operand are supplied on the Z80 stack. The slice size is taken from the stack (68E3H) and subtracted from the source string length to determine the slice starting position. Control then transfers to the "LEFT$" routine to extract the slice (6865H). Address ... 689AH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "MID$" function. The first operand's descriptor address and the integer second operand are supplied on the Z80 stack. The starting position is taken from the stack (68E6H) and checked, if it is zero an "Illegal function call" error is generated (475AH). The optional slice size is then evaluated (69E4H) and control transfers to the "LEFT$" routine to extract the slice (6869H). Address ... 68BBH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "VAL" function to an operand contained in DAC. The string length is taken from the descriptor (6803H) and checked, if it is zero it is placed in DAC as an integer (4FCFH). The length is then added to the starting address of the string body to give the location of the character immediately following it. This is temporarily replaced with a zero byte and the string is converted to numeric form in DAC (3299H). The original character is then restored and the routine terminates. The temporary zero byte delimiter is necessary because strings are packed together in the String Storage Area, without it the numeric converter would run on into succeeding strings. Address ... 68E3H

Address

@ (MSX only)

    This routine is used by the "LEFT$", "MID$" and "RIGHT$" function handlers to check that the next program text character is ")" and then to pop an operand from the Z80 stack into register pair DE. Address ... 68EBH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "INSTR" function. The first operand, which may be the starting position or the source string, is evaluated (4C62H) and its type tested. If it is the source string a default starting position of one is taken. If it is the starting position operand its value is checked and the source string operand evaluated (4C64H). The pattern string is then evaluated (4C64H) and the storage of both operands freed (FRESTR, 67D0H). The length of the pattern string is checked and, if zero, the starting position is placed in DAC (4FCFH). The pattern string is then checked against successive characters from the source string, commencing at the starting position, until a match is found or the source string is exhausted. With a successful search the character position of the substring is placed in DAC as an integer (4FCFH), otherwise a zero result is returned. Address ... 696EH

Address

@ (MSX only)

    This is the "MID$" statement handler. After checking for the open parenthesis character the destination Variable is located (PTRGET, 5EA4H) and checked to ensure that it is a string type (3058H). The address of the string body is then taken from the Variable and examined to determine whether it is inside the Program Text Area, as would be the case for an explicitly stated string. If this is the case the string body is copied to the String Storage Area (6611H) and a new descriptor copied to the Variable (2EF3H). This is done to avoid modifying the program text. The starting position is then evaluated (GETBYT, 521CH) and checked, if it is zero an "Illegal function call" error is generated (475AH). The optional slice length operand is evaluated (69E4H) followed by the replacement string (4C5FH) whose storage is then freed (FRESTR, 67D0H). Characters are then copied from the replacement string to the destination string until either the slice length is completed or the replacement string is exhausted. Address ... 69E4H

Address

@ (MSX only)

    This routine is used by various string functions to evaluate an optional operand (GETBYT, 521CH) and return the result in register E. If no operand is present a default value of 255 is returned. Address ... 69F2H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "FRE" function to an operand contained in DAC. If the operand is numeric the single precision difference between the Z80 Stack Pointer and the contents of STREND is placed in DAC (4FC1H). If the operand is a string type its storage is freed (67D3H) and garbage collection initiated (66B6H). The single precision difference between the contents of FRETOP and those of STKTOP is then placed in DAC (4FC1H). Address ... 6A0EH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to analyze a filespec such as "A:FILENAME.BAS". The filespec consists of three parts, the device, the filename and the type extension. On entry register pair HL points to the start of the filespec in the program text. On exit register D holds the device code, the filename is in positions zero to seven of FILNAM and the type extension in positions eight to ten. Any unused positions are filled with spaces.

    The filespec string is evaluated (4C64H) and its storage freed (FRESTR, 67D0H), if the string is of zero length a "Bad file name" error is generated (6E6BH). The device name is parsed (6F15H) and successive characters taken from the filespec and placed in FILNAM until the string is exhausted, a "." character is found or FILNAM is full. A "Bad file name" error is generated (6E6BH) if the filespec contains any control characters, that is those whose value is smaller than 20H. If the filespec contains a type extension a "Bad file name" error is generated (6E6BH) if it is longer than three characters or if the filename is longer than eight characters. If no type extension is present the filename may be any length, extra characters are simply ignored. Address ... 6A6DH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to locate the I/O buffer FCB whose number is supplied in register A. The buffer number is first checked against MAXFIL and a "Bad file number" error generated (6E7DH) if it is too large. Otherwise the required address is taken from the file pointer block and placed in register pair HL and the buffer's mode taken from byte 0 of the FCB and placed in register A. Address ... 6A9EH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to evaluate an I/O buffer number and to locate its FCB. Any "#" character is skipped (4666H) and the buffer number evaluated (GETBYT, 521CH). The FCB is located (6A6DH) and a "File not open" error generated (6E77H) if the buffer mode byte is zero. Otherwise the FCB address is placed in PTRFIL to redirect the Interpreter's output. Address ... 6AB7H

Address

@ (MSX only)

    This is the "OPEN" statement handler. The filespec is analyzed (6A0EH) and any following mode converted to the corresponding mode byte, these are: "FOR INPUT" (01H), "FOR OUTPUT" (02H) and "FOR APPEND" (08H). If no mode is explicitly stated random mode (04H) is assumed. The "AS" characters are checked and the buffer number evaluated (GETBYT, 521CH), if this is zero a "Bad file number" error is generated (6E7DH). The FCB is then located (6A6DH) and a "File already open" error generated (6E6EH) if the buffer's mode byte is anything other than zero. The device code is placed in byte 4 of the FCB, the open function dispatched (6F8FH) and the Interpreter's output reset to the screen (4AFFH). Address ... 6B24H

Address

@ (MSX only)

    This routine is used by the file I/O handlers to close the I/O buffer whose number is supplied in register A. The FCB is located (6A6DH) and, provided the buffer is in use, the close function dispatched (6F8FH) and the buffer filled with zeroes (6CEAH). PTRFIL and the FCB mode byte are then zeroed to reset the Interpreter's output to the screen. Address ... 6B5BH

Address

@ (MSX only)

    This is the "LOAD", "MERGE" and "RUN filespec" statement handler. The filespec is analyzed (6A0EH) and then, for "LOAD" and "RUN" only, the program text examined to determine whether the auto-run "R" option is specified. I/O buffer 0 is opened for input (6AFAH) and the first byte of FILNAM set to FFH if auto-run is required. For "LOAD" and "RUN" only any program text is then cleared via the "NEW" statement handler (6287H). As this will reset the Interpreter's output to the screen the buffer FCB is again located and placed in PTRFIL (6AAAH). Control then transfers directly to the Interpreter Mainloop (4134H) for the program text to be loaded as if typed from the keyboard. Note that no error checking of any sort is carried out on the data read. Address ... 6BA3H

Address

@ (MSX only)

    This is the "SAVE" statement handler. The filespec is analyzed (6A0EH) and the program text examined to determine whether the ASCII "A" suffix is present. This is only relevant under Disk BASIC, it makes no difference on a standard MSX machine. I/O buffer 0 is opened for output (6AFAH) and control transfers to the "LIST" statement handler (522EH) to output the program text. Note that no error checking information of any sort accompanies the text. - 172 - 5. ROM BASIC INTERPRETER Address ... 6BDAH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to return the device code for the currently active I/O buffer. The FCB address is taken from PTRFIL then the device code taken from byte 4 of the FCB and placed in register A. Address ... 6BE7H

Address

@ (MSX only)

    This routine is used by the file I/O handlers to perform an operation on a number of I/O buffers. The address of the relevant routine is supplied in register pair BC and the buffer count in register A. For example if register pair BC contained 6B24H and register A contained 03H buffers 3, 2, 1 and 0 would be closed. The routine has a slightly different function if it is entered with FLAG NZ. In this case the I/O buffer numbers are taken sequentially from the program text and evaluated (GETBYT, 521CH) before the operation is performed, a typical case might be "#1,#2". Address ... 6C14H

Address

@ (MSX only)

    This is the "CLOSE" statement handler. Register pair BC is set to 6B24H, register A is loaded with the contents of MAXFIL and the required number of buffers closed (6BE7H). Address ... 6C1CH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to close every I/O buffer. Register pair BC is set to 6B24H, register A is loaded with the contents of MAXFIL and all buffers closed (6BE7H). Address ... 6C2AH

Address

@ (MSX only)

    This is the "LFILES" statement handler. PRTFLG is made non- zero, to direct output to the printer, and control drops into the "FILES" statement handler. Address ... 6C2FH

Address

@ (MSX only)

    This is the "FILES" statement handler, an "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 6C35H

Address

@ (MSX only)

    Control transfers here from the general "PUT" and "GET" handlers (7758H) when the program text contains anything other than a "SPRITE" token. A "Sequential I/O only" error is generated (6E86H) on a standard MSX machine. Address ... 6C48H

Address

@ (MSX only)

    This routine is used by the file I/O handlers to sequentially output the character supplied in register A. The character is placed in register C and the sequential output function dispatched (6F8FH). Address ... 6C71H

Address

@ (MSX only)

    This routine is used by the file I/O handlers to sequentially input a single character. The sequential input function is dispatched (6F8FH) and the character returned in register A, FLAG C indicates an EOF (End Of File) condition. Address ... 6C87H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "INPUT$" function. The program text is checked for the "$" and "(" characters and the length operand evaluated (GETBYT, 521CH). If an I/O buffer number is present it is evaluated, the FCB located (6A9EH) and the mode byte examined. An "Input past end" error is generated (6E83H) if the buffer is not in input or random mode. After checking that sufficient space is available (6627H) the required number of characters are sequentially input (6C71H), or collected via the CHGET standard routine, and copied to the String Storage Area. Finally the result descriptor is created (6654H). Address ... 6CEAH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to fill the buffer whose FCB address is contained in PTRFIL with two hundred and fifty-six zeroes. Address ... 6CFBH

Address

@ (MSX only)

    This routine is used by the file I/O handlers to return, in register pair HL, the starting address of the buffer whose FCB address is contained in PTRFIL. This just involves adding nine to the FCB address. Address ... 6D03H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "LOC" function to the I/O buffer whose number is contained in DAC. The FCB is located (6A6AH) and the LOC function dispatched (6F8FH). An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 6D14H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "LOF" function to the I/O buffer whose number is contained in DAC. The FCB is located (6A6AH) and the LOF function dispatched (6F8FH). An "Illegal function call" error is generated (475AH) on a standard MSX machine. - 174 - 5. ROM BASIC INTERPRETER Address ... 6D25H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "EOF" function to the I/O buffer whose number is contained in DAC. The FCB is located (6A6AH) and the EOF function dispatched (6F8FH). Address ... 6D39H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "FPOS" function to the I/O buffer whose number is contained in DAC. The FCB is located (6A6AH) and the FPOS function dispatched (6F8FH). An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 6D48H

Address

@ (MSX only)

    Control transfers to this routine when the Interpreter Mainloop encounters a direct statement, that is one with no line number. The ISFLIO standard routine is first used to determine whether a "LOAD" statement is active. If input is coming from the keyboard control transfers to the Runloop execution point (4640H) to execute the statement. If input is coming from the cassette buffer 0 is closed (6B24H) and a "Direct statement in file" error generated (6E71H). This could happen on a standard MSX machine either through a cassette error or by attempting to load a text file with no line numbers. Address ... 6D57H

Address

@ (MSX only)

    This routine is used by the "INPUT", "LINE INPUT" and "PRINT" statement handlers to check for the presence of a "#" character in the program text. If one is found the I/O buffer number is evaluated (521BH), the FCB located and its address placed in PTRFIL (6AAAH). The mode byte of the FCB is then compared with the mode number supplied by the statement handler in register C, if they do not match a "Bad file number" error is generated (6E7DH). With "PRINT" the allowable modes are output, random and append. With "INPUT" and "LINE INPUT" the allowable modes are input and random. Note that on a standard MSX machine not all these modes are supported at lower levels. Some sort of error will consequently be generated at a later stage for illegal modes. Address ... 6D83H

Address

@ (MSX only)

    This routine is used by the "INPUT" statement handler to input a string from an I/O buffer. A return is first set up to the "READ/INPUT" statement handler (4BF1H). The characters which delimit the input string, comma and space for a numeric Variable and comma only for a string Variable, are placed in registers D and E and control transfers to the "LINE INPUT" routine (6DA3H). - 175 - 5. ROM BASIC INTERPRETER Address ... 6D8FH

Address

@ (MSX only)

    This is the "LINE INPUT" statement handler when input is from an I/O buffer. The buffer number is evaluated, the FCB located and the mode checked (6D55H). The Variable to assign to is then located (PTRGET, 5EA4H) and its type checked to ensure it is a string type (3058H). A return is set up to the "LET" statement handler (487BH) to perform the assignment and the input string collected.

    Characters are sequentially input (6C71H) and placed in BUF until the correct delimiter is found, EOF is reached or BUF fills up (6E41H). When the terminating condition is reached and assignment is to a numeric Variable the string is converted to numeric form in DAC (3299H). When assignment is to a string Variable the string is analyzed and the result descriptor created (6638H).

    For "LINE INPUT" all characters are accepted until a CR code is reached. Note that if this CR code is preceded by a LF code then it will not function as a delimiter but will merely be accepted as part of the string. For "INPUT" to a numeric Variable leading spaces are stripped then characters accepted until a CR code, a space or a comma is reached. Note that as for "LINE INPUT" a CR code will not function as a delimiter when preceded by a LF code. In this case however the CR code will not be placed in BUF but ignored. For "INPUT" to a string Variable leading spaces are stripped then characters accepted until a CR or comma is reached. Note that as for "LINE INPUT" a CR code will not function as a delimiter when preceded by a LF code. In this case however neither code will be placed in BUF both are ignored. An alternative mode is entered when the first character read, after any spaces, is a double quote character. In this case all characters will be accepted, and stored in BUF, until another double quote delimiter is read.

    Once the input string has been accepted the terminating delimiter is examined to see if any special action is required with respect to trailing characters. If the input string was delimited by a double quote character or a space then any succeeding spaces will be read in and ignored until a non-space character is found. If this character is a comma or CR code then it is accepted and ignored. Otherwise a putback function is dispatched (6F8FH) to return the character to the I/O buffer. If the input string was delimited by a CR code then the next character is read in and checked. If this is a LF code it will be accepted but ignored. If it is not a LF code then a putback function is dispatched (6F8FH) to return the character to the I/O buffer. Address ... 6E6BH

Address

@ (MSX only)

    This is a group of ten file I/O related error generators. Register E is loaded with the relevant error code and control transfers to the error handler (406FH):

        ADDR. ERROR
        -------------------------------
        6E6BH Bad file name
        6E6EH File already open
        6E71H Direct statement in file
        6E74H File not found
        6E77H File not open
        6E7AH Field overflow
        6E7DH Bad file number
        6E80H Internal error
        6E83H Input past end
        6E86H Sequential I/O only

Address

@ (MSX only)

    This is the "BSAVE" statement handler. The filespec is analyzed (6A0EH) and the start address evaluated (6F0BH). The stop address is then evaluated (6F0BH) and placed in SAVEND followed by the optional entry address (6F0BH) which is placed in SAVENT. If no entry address exists the start address is taken instead. The device code is checked to ensure that it is CAS, if not a "Bad file name" error is generated (6E6BH), and the data written to cassette (6FD7H). Note that no buffering is involved, data is written directly to the cassette, and no error checking information accompanies the data. Address ... 6EC6H

Address

@ (MSX only)

    This is the "BLOAD" statement handler. The filespec is analyzed (6A0EH) and RUNBNF made non-zero if the auto-run "R" option is present in the program text. The optional load offset, with a default value of zero, is then evaluated (6F0BH) and the device code checked to ensure that it is CAS, if not a "Bad file name" error is generated (6E6BH). Datum is then read directly from cassette (7014H), as with "BSAVE" no buffering or error checking is involved. Address ... 6EF4H

Address

@ (MSX only)

    Control transfers to this routine when the "BLOAD" statement handler has completed loading data into memory. If RUNBNF is zero buffer 0 is closed (6B24H) and control returns to the Runloop. Otherwise buffer 0 is closed (6B24H), a return address of 6CF3H is set up (this routine just pops the program text pointer back into register pair HL and returns to the Runloop) and control transfers to the address contained in SAVENT. Address ... 6F0BH

Address

@ (MSX only)

    This routine is used by the "BLOAD" and "BSAVE" handlers to evaluate an address operand, the result is returned in register pair DE. The operand is evaluated (4C64H) then converted to an integer (5439H). Address ... 6F15H

Address

@ (MSX only)

    This routine is used by the filespec analyzer to parse a device name such as "CAS:". On entry register pair HL points to the start of the filespec string and register E contains its length. If no device name is present the default device code (CAS=FFH) is returned in register A with FLAG Z. If a legal device name is present its code is returned in register A with FLAG NZ.

    The filespec is examined until a ":" character is found then the name compared with each of the legal device names in the device table at 6F76H. If a match is found the device code is taken from the table and returned in register A. If no match is found control transfers to the external ROM search routine (55F8H). Note that any lower case characters are turned to upper case for comparison purposes. Thus crt and CRT, for example, are the same device. Address ... 6F76H

Address

@ (MSX only)

    This table is used by the device name parser, it contains the four device names and codes available on a standard MSX machine:

        CAS ... FFH  LPT ... FEH  CRT ... FDH  GRP ... FCH

Address

@ (MSX only)

    This table is used by the function dispatcher (6F8FH), it contains the address of the function decoding table for each of the four standard MSX devices:

        CAS ... 71C7H  LPT ... 72A6H  CRT ... 71A2H  GRP ... 7182H

Address

@ (MSX only)

    This is the file I/O function dispatcher. In conjunction with the Interpreter's buffer structure it provides a consistent, device independent method of inputting or outputting data. The required function code is supplied in register A and the address of the buffer FCB in register pair HL.

    The device code is taken from byte 4 of the FCB and examined to determine whether it is one of the four standard devices, if not control transfers to the external ROM function dispatcher (564AH). Otherwise the address of the device's function decoding table is taken from the table at 6F87H, the required function's address taken from it and control transferred to the relevant function handler. - 178 - 5. ROM BASIC INTERPRETER Address ... 6FB7H

Address

@ (MSX only)

    This is the "CSAVE" statement handler. The filename is evaluated (7098H) followed by the optional baud rate operand (7A2DH). The identification block is then written to cassette (7125H) with a filetype byte of D3H. The contents of the Program Text Area are written directly to cassette as a single data block (713EH). Note that no error checking information accompanies the data. Address ... 6FD7H

Address

@ (MSX only)

    Control transfers to this routine from the "BSAVE" statement handler to write a block of memory to cassette. The identification block is first written to cassette (7125H) with a filetype byte of D0H. The motor is then turned on and a short header written to cassette (72F8H) The starting address is popped from the Z80 stack and written to cassette LSB first, MSB second (7003H). The stop address is taken from SAVEND and written to cassette LSB first, MSB second (7003H). The entry address is taken from SAVENT and written to cassette LSB first, MSB second (7003H). The required area of memory is then written to cassette one byte at a time (72DEH) and the cassette motor turned off via the TAPOOF standard routine. Note that no error checking information accompanies the data. Address ... 7003H

Address

@ (MSX only)

    This routine writes the contents of register pair HL to cassette with register L first (72DEH) and register H second (72DEH). Address ... 700BH

Address

@ (MSX only)

    This routine reads two bytes from cassette and places the first in register L (72D4H), the second in register H (72D4H). Address ... 7014H

Address

@ (MSX only)

    Control transfers to this routine from the "BLOAD" statement handler to load data from the cassette into memory. The cassette is read until an identification block with a file type of D0H and the correct filename is found (70B8H). The data block header is then located on the cassette (72E9H). The offset value is popped from the Z80 stack and added to the start address from the cassette (700BH). The stop address is read from cassette (700BH) and the offset added to this as well. The entry address is read from cassette (700BH) and placed in SAVENT in case auto-run is required. Successive data bytes are then read from cassette (72D4H) and placed in memory, at the start address initially, until the stop address is reached. Finally the motor is turned off via the TAPIOF standard routine and control transfers to the "BLOAD" termination point (6EF4H). - 179 - 5. ROM BASIC INTERPRETER Address ... 703FH

Address

@ (MSX only)

    This is the "CLOAD" and "CLOAD?" statement handler. The program text is first checked for a trailing "PRINT" token (91H) which is how the "?" character is tokenized. The filename is then evaluated (708CH) and the cassette read until an identification block with a filetype of D3H and the correct filename is found (70B8H). For "CLOAD" a "NEW" operation is then performed (6287H) to erase the current program text. For "CLOAD?" all pointers in the Program Text Area are converted to line numbers (54EAH) to match the cassette data.

    The data block header is located on the cassette and successive data bytes read from cassette and placed in memory or compared with the current memory contents (715DH). When the data block has been completely read the message "OK" is displayed (6678H) and control transfers directly to the end of the Interpreter Mainloop (4237H) to reset the Variable storage pointers. For "CLOAD?" reading of the data block will terminate if the cassette byte is not the same as the program text byte in memory. If the address where this occurred is above the end of the Program Text Area then the handler terminates with an "OK" message as before. Otherwise a "Verify error" is generated. Address ... 708CH

Address

@ (MSX only)

    This routine is used by the "CLOAD" and "CSAVE" statement handlers to evaluate a filename in the program text. The two handlers use different entry points so that a null filename is allowed for "CLOAD" but not for "CSAVE". The filename string is evaluated (4C64H), its storage freed (680FH) and the first six characters copied to FILNAM. If the filename is longer than six characters the excess is ignored. If the filename is shorter than six characters then FILNAM is padded with spaces. Address ... 70B8H

Address

@ (MSX only)

    This routine is used by the "CLOAD" and "BLOAD" statement handlers and for the dispatcher open function (when the device is CAS and the mode is input) to locate an identification block on the cassette. On entry the filename is in FILNAM and the file type in register C, D3H for a tokenized BASIC (CLOAD) file, D0H for a binary (BLOAD) file and EAH for an ASCII (LOAD or data) file.

    The cassette motor is turned on and the cassette read until a header is found (72E9H). Each identification block is prefixed by ten file type characters so successive characters are read from cassette (72D4H) and compared to the required file type. If the file type characters do not match control transfers back to the start of the routine to find the next header. Otherwise the next six characters are read in (72D4H) and placed in FILNAM. If FILNAM is full of spaces no filename match is attempted and the identification block has been found. Otherwise the contents of FILNAM and FILNM2 are compared to determine whether this is the required file. If the match is unsuccessful, and the Interpreter is in direct mode, the message "Skip:" is displayed (710DH) followed by the filename. Control then transfers back to the start of the routine to try the next header. If the match is successful, and the Interpreter is in direct mode, the message "Found:" is displayed (710DH) followed by the filename and the routine terminates. Address ... 70FFH

Address

@ (MSX only)

    This is the plain text message "Found:" terminated by a zero byte. Address ... 7106H

Address

@ (MSX only)

    This is the plain text message "Skip :" terminated by a zero byte. Address ... 710DH

Address

@ (MSX only)

    Unless CURLIN shows the Interpreter to be in program mode this routine first displays (6678H) the message whose address is supplied in register pair HL, followed by the six characters contained in FILENAM2. Address ... 7125H

Address

@ (MSX only)

    This routine is used by the "CSAVE" and "BSAVE" statement handlers and for the dispatcher open function (when the device is CAS and the mode is output) to write an identification block to cassette. On entry the filename is in FILNAM and the filetype in register A, D3H for a tokenized BASIC (CSAVE) file, D0H for a binary (BSAVE) file and EAH for an ASCII (SAVE or data) file. The cassette motor is turned on and a long header written to cassette (72F8H) The filetype byte is then written to cassette (72DEH) ten times followed by the first six characters from FILNAM (72DEH). The cassette motor is turned off via the TAPOOF standard routine and the routine terminates. Address ... 713EH

Address

@ (MSX only)

    This routine is used by the "CSAVE" statement handler to write the Program Text Area to cassette as a single data block. All pointers in the program text are converted back to line numbers (54EAH) to make the text address independent. The cassette motor is turned on and a short header written to cassette (72F8H) The entire Program Text Area is then written to cassette a byte at a time (72DEH) and followed with seven zero bytes (72DEH) as a terminator. The cassette motor is then turned off via the TAPOOF standard routine and the routine terminates. - 181 - 5. ROM BASIC INTERPRETER Address ... 715DH

Address

@ (MSX only)

    This routine is used by the "CLOAD" and "CLOAD?" statement handlers to read a single data block into the Program Text Area or to compare it with the current contents. On entry register A contains a flag to distinguish between the two statements, 00H for "CLOAD" and FFH for "CLOAD?". The cassette motor is turned on and the first header located (72E9H). Successive characters are read from cassette (72D4H) and placed in the Program Text Area or compared with the current contents. If the current statement is "CLOAD?" the routine will terminate with FLAG NZ if the cassette character is not the same as the memory character. Otherwise data will be read until ten successive zeroes are found. This sequence of zeroes is composed of the last program line end of line character, the end link and the seven terminator zeroes added by "CSAVE". Note that the routine will probably terminate during this sequence, when used by "CLOAD?", as memory comparison is still in progress. This accounts for the rather peculiar coding of the "CLOAD?" handler terminating conditions. Address ... 7182H

Address

@ (MSX only)

    This table is used by the dispatcher when decoding function codes for the GRP device. It contains the address of the handler for each of the function codes, most are in fact error generators:

        TO    FUNCTION
        -------------------------------
        71B6H  0, open
        71C2H  2, close
        6E86H  4, random
        7196H  6, sequential output
        475AH  8, sequential input
        475AH 10, loc
        475AH 12, lof
        475AH 14, eof
        475AH 16, fpos
        475AH 18, putback

Address

@ (MSX only)

    This is the dispatcher sequential output routine for the GRP device. SCRMOD is first checked and an "Illegal function call" error generated (475AH) if the screen is in either text mode. The character to output is taken from register C and control transfers to the GRPPRT standard routine. Address ... 71A2H

Address

@ (MSX only)

    This table is used by the DEVICE DISPATCHER when decoding function codes for the CRT device. It contains the address of the handler for each of the function codes, most are in fact error generators:

        TO    FUNCTION
        -------------------------------
        71B6H  0, open
        71C2H  2, close
        6E86H  4, random
        71C3H  6, sequential output
        475AH  8, sequential input
        475AH 10, loc
        475AH 12, lof
        475AH 14, eof
        475AH 16, fpos
        475AH 18, putback

Address

@ (MSX only)

    This is the dispatcher open routine for the CRT, LPT and GRP devices. The required mode, in register E, is checked and a "Bad file name" error generated (6E6BH) for input or append. The FCB address is then placed in PTRFIL, the mode in byte 0 of the FCB and the routine terminates. Note that the Z80 RET instruction at the end of this routine (71C2H) is the dispatcher close routine for the CRT, LPT and GRP devices. Address ... 71C3H

Address

@ (MSX only)

    This is the dispatcher sequential output routine for the CRT device. The character to output is taken from register C and control transfers to the CHPUT standard routine. Address ... 71C7H

Address

@ (MSX only)

    This table is used by the dispatcher when decoding function codes for the CAS device. It contains the address of the handler for each of the function codes, several are error generators:

        TO    FUNCTION
        -------------------------------
        71DBH  0, open
        7205H  2, close
        6E86H  4, random
        722AH  6, sequential output
        723FH  8, sequential input
        475AH 10, loc
        475AH 12, lof
        726DH 14, eof
        475AH 16, fpos
        727CH 18, putback

Address

@ (MSX only)

    This is the dispatcher open routine for the CAS device. The current I/O buffer position, held in byte 6 of the FCB, and CASPRV, which holds any putback character are both zeroed. The required mode, supplied in register E, is examined and a "Bad file name" error generated (6E6BH) for append or random modes. For output mode the identification block is then written to cassette (7125H) while for input mode the correct identification block is located on the cassette (70B8H). The FCB address is then placed in PTRFIL, the mode in byte 0 of the FCB and the routine terminates. Address ... 7205H

Address

@ (MSX only)

    This is the dispatcher close routine for the CAS device. Byte 0 of the FCB is examined and, if the mode is input, CASPRV is zeroed and the routine terminates. Otherwise the remainder of the I/O buffer is filled with end of file characters (1AH) and the I/O buffer contents written to cassette (722FH). CASPRV is then zeroed and the routine terminates. Address ... 722AH

Address

@ (MSX only)

    This is the dispatcher sequential output routine for the CAS device. The character to output is taken from register C and placed in the next free position in the I/O buffer (728BH). Byte 6 of the FCB, the I/O buffer position, is then incremented. If the I/O buffer position has wrapped round to zero this means that there are two hundred and fifty-six characters in the I/O buffer and it has to be written to cassette. The cassette motor is turned on, a short header is written to cassette (72F8H) followed by the I/O buffer contents (72DEH), and the motor is turned off via the TAPOOF standard routine. Address ... 723FH

Address

@ (MSX only)

    This is the dispatcher sequential input routine for the CAS device. CASPRV is first checked (72BEH) to determine whether it contains a character which has been putback, in which case its contents will be non-zero. If so the routine terminates with the character in register A. Otherwise the I/O buffer position is checked (729BH) to determine whether it contains any characters. If the I/O buffer is empty the cassette motor is turned on and the header located (72E9H). Two hundred and fifty-six characters are then read in (72D4H), the cassette motor turned off via the TAPION standard routine and the I/O buffer position reset to zero. The character is then taken from the current I/O buffer position and the position incremented. Finally the character is checked to see if it is the end of file character (1AH). If it is not the routine terminates with the character in register A and FLAG NC. Otherwise the end of file character is placed in CASPRV, so that succeeding sequential input requests will always return the end of file condition, and the routine terminates with FLAG C. - 184 - 5. ROM BASIC INTERPRETER Address ... 726DH

Address

@ (MSX only)

    This is the dispatcher eof routine for the CAS device. The next character is input (723FH) and placed in CASPRV. It is then tested for the end of file code (1AH) and the result placed in DAC as an integer, zero for false, FFFFH for true. Address ... 727CH

Address

@ (MSX only)

    This is the dispatcher putback routine for the CAS device. The character is simply placed in CASPRV to be picked up at the next sequential input request. Address ... 7281H

Address

@ (MSX only)

    This routine is used by the dispatcher close function to check if there are any characters in the I/O buffer and then zero the I/O buffer position byte in the FCB. Address ... 728BH

Address

@ (MSX only)

    This routine is used by the dispatcher sequential output function to place the character in register A in the I/O buffer at the current I/O buffer position, which is then incremented. Address ... 729BH

Address

@ (MSX only)

    This routine is used by the dispatcher sequential input function to collect the character at the current I/O buffer position, which is then incremented. Address ... 72A6H

Address

@ (MSX only)

    This table is used by the dispatcher when decoding function codes for the LPT device. It contains the address of the handler for each of the function codes, most are in fact error generators:

        TO    FUNCTION
        -------------------------------
        71B6H  0, open
        71C2H  2, close
        6E86H  4, random
        72BAH  6, sequential output
        475AH  8, sequential input
        475AH 10, loc
        475AH 12, lof
        475AH 14, eof
        475AH 16, fpos
        475AH 18, putback

Address

@ (MSX only)

    This is the dispatcher sequential output routine for the LPT device. The character to output is taken from register C and control transfers to the OUTDLP standard routine. Address ... 72BEH

Address

@ (MSX only)

    This routine is used by the dispatcher sequential input function to check if a putback character exists in CASPRV, and if not to return Flag Z. Otherwise CASPRV is zeroed and the character tested to see if it is the end of file character (1AH). If not it returns with the character in register A and FLAG NZ and NC. Otherwise the end of file character is placed back in CASPRV and the routine returns with FLAG Z and C. Address ... 72CDH

Address

@ (MSX only)

    This routine is used by various dispatcher functions to check if the mode in register E is append, if so a "Bad file name" error is generated (6E6BH). Address ... 72D4H

Address

@ (MSX only)

    This routine is used by various dispatcher functions to read a character from the cassette. The character is read via the TAPIN standard routine and a "Device I/O error" generated (73B2H) if FLAG C is returned. Address ... 72DEH

Address

@ (MSX only)

    This routine is used by various dispatcher functions to write a character to cassette. The character is written via the TAPOUT standard routine and a "Device I/O error" generated (73B2H) if FLAG C is returned. Address ... 72E9H

Address

@ (MSX only)

    This routine is used by various dispatcher functions to turn the cassette motor on for input. The motor is turned on via the TAPION standard routine and a "Device I/O error" generated (73B2H) if FLAG C is returned. Address ... 72F8H

Address

@ (MSX only)

    This routine is used by various dispatcher functions to turn the cassette motor on for output, control simply transfers to the TAPOON standard routine. Address ... 7304H

Address

@ (MSX only)

    This routine is used by the Interpreter Mainloop "OK" point, the "END" statement handler and the run-clear routine to shut down the printer. PRTFLG is first zeroed and then LPTPOS tested to see if any characters have been output but left hanging in the printer's line buffer. If so a CR,LF sequence is issued to flush the printer and LPTPOS zeroed. - 186 - 5. ROM BASIC INTERPRETER Address ... 7323H

Address

@ (MSX only)

    This routine issues a CR,LF sequence to the current output device via the OUTDO standard routine. LPTPOS or TTYPOS is then zeroed depending upon whether the printer or the screen is active. Address ... 7347H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "INKEY$" function. The state of the keyboard buffer is examined via the CHSNS standard routine. If the buffer is empty the address of a dummy null string descriptor is returned in DAC. Otherwise the next character is read from the keyboard buffer via the CHGET standard routine. After checking that sufficient space is available (6625H) the character is copied to the String Storage Area and the result descriptor created (6821H). Address ... 7367H

Address

@ (MSX only)

    This routine is used by the "LIST" statement handler to output a character to the current output device via the OUTDO standard routine. If the character is a LF code then a CR code is also issued. Address ... 7374H

Address

@ (MSX only)

    This routine is used by the Interpreter Mainloop to collect a line of text when input is from an I/O buffer rather than the keyboard, that is when a "LOAD" statement is active. Characters are sequentially input (6C71H) and placed in BUF until BUF fills up, a CR is detected or the end of file is reached. All characters are accepted apart from LF codes which are filtered out. If BUF fills up or a CR is detected the routine simply returns the line to the Mainloop. If the end of file is reached while some characters are in BUF the line is returned to the Mainloop. When end of file is reached with no characters in BUF then I/O buffer 0 is closed (6D7BH) and FILNAM checked to determine whether auto-run is required. If not control returns to the Interpreter "OK" point (411EH). Otherwise the system is cleared (629AH) and control transfers to the Runloop (NEWSTT, 4601H) to execute the program. Address ... 73B2H

Address

@ (MSX only)

    This is the "Device I/O error" generator. Address ... 73B7H

Address

@ (MSX only)

    This is the "MOTOR" statement handler. If no operand is present control transfers to the STMOTR standard routine with FFH in register A. If the "OFF" token (EBH) follows control transfers with 00H in register A. If the "ON" token (95H) follows control transfers with 01H in register A. - 187 - 5. ROM BASIC INTERPRETER Address ... 73CAH

Address

@ (MSX only)

    This is the "SOUND" statement handler. The register number operand, which must be less than fourteen, is evaluated (GETBYT, 521CH) and placed in register A. The data operand is evaluated (GETBYT, 521CH) and bit 7 set, bit 6 reset to avoid altering the PSG auxiliary I/O port modes. The data operand is placed in register E and control transfers to the WRTPSG standard routine. Address ... 73E4H

Address

@ (MSX only)

    This is a single ASCII space used by the "PLAY" statement handler to replace a null string operand with a one character blank string. Address ... 73E5H

Address

@ (MSX only)

    This is the "PLAY" statement handler. The address of the "PLAY" command table at 752EH is placed in MCLTAB for the macro language parser and PRSCNT zeroed. The first string operand, which is obligatory, is evaluated (4C64H), its storage freed (FRESTR, 67D0H) and its length and address placed in VCBA at bytes 2, 3 and 4. The channel's stack pointer is initialized to VCBA+33 and placed in VCBA at bytes 5 and 6. If further text is present in the statement this process is repeated for voices B and C until a maximum of three operands have been evaluated, after this a "Syntax error" is generated (4055H). If there are less than three string operands present an end of queue mark (FFH) is placed in the queue (7507H) of each unused voice. Register A is then zeroed, to select voice A, and control drops into the play mainloop. Address ... 744DH

Address

@ (MSX only)

    This is the play mainloop. The number of free bytes in the current queue is checked (7521H) and, if less than eight bytes remain, the next voice is selected (74D6H) to avoid waiting for the queue to empty. The remaining length of the operand string is then taken from the current voice buffer and, if zero bytes remain to be parsed, the loop again skips to the next voice (74D6H). Otherwise the current string length and address are taken from the voice buffer and placed in MCLLEN and MCLPTR for the macro language parser. The old stack contents are copied from the voice buffer to the Z80 stack (6253H), MCLFLG is made non-zero and control transfers to the macro language parser (56A2H).

    The macro language parser will normally scan along the string, using the "PLAY" statement command handlers, until the string is exhausted. However, if a music queue fills up during note generation an abnormal termination is forced back to the play mainloop (748EH) so that the next voice can be processed without waiting for the queue to empty. When control returns normally an end of queue mark is placed in the current queue (7507H) and PRSCNT is incremented to show the number of strings completed. If control returns abnormally then anything left on the Z80 stack is copied into the current voice buffer (6253H). Because of the recursive nature of the macro language parser where the "X" command is involved there may be a number of four byte string descriptors, marking the point where the original string was suspended, left on the Z80 stack at termination. Saving the stack contents in the voice buffer means they can be restored when the loop gets around to that voice again. Note that as there are only sixteen bytes available in each voice buffer an "Illegal function call" error is generated (475AH) if too much data remain on the stack. This will occur when a queue fills up and multiple, nested "X" commands exist, for example:

        10 A$="XB$;"
        20 B$="XC$;"
        30 C$="XD$;"
        40 D$=STRING$(150,"A")
        50 PLAY A$

    There seems to be a slight bug in this section as only fifteen bytes of stack data are allowed, instead of sixteen, before an error is generated.

    When control returns from the macro language parser register A is incremented to select the next voice for processing. When all three voices have been processed INTFLG is checked and, if CTRL-STOP has been detected by the interrupt handler, control transfers to the GICINI standard routine to halt all music and terminate. Assuming bit 7 of PRSCNT shows this to be the first pass through the mainloop, that is no voice has been temporarily suspended because of a full queue, PLYCNT is incremented and interrupt dequeueing started via the STRTMS standard routine. PRSCNT is then checked to determine the number of strings completed by the macro language parser. If all three operand strings have been completed the handler terminates, otherwise control transfers back to the start of the play mainloop to try each voice again. Address ... 7507H

Address

@ (MSX only)

    This routine is used by the "PLAY" statement handler to place an end of queue mark (FFH) in the current queue via the PUTQ standard routine. If the queue is full it waits until space becomes available. Address ... 7521H

Address

@ (MSX only)

    This routine is used by the "PLAY" statement handler to check how much space remains in the current queue via the LFTQ standard routine. If less than eight bytes remain (the largest possible music data packet is seven bytes long) FLAG C is returned. - 189 - 5. ROM BASIC INTERPRETER Address ... 752EH

Address

@ (MSX only)

    This table contains the valid command letters and associated addresses for the "PLAY" statement commands. Those commands which take a parameter, and consequently have bit 7 set in the table, are shown with an asterisk:

        CMD  TO
        -----------
        A    763EH
        B    763EH
        C    763EH
        D    763EH
        E    763EH
        F    763EH
        G    763EH
        M*   759EH
        V*   7586H
        S*   75BEH
        N*   7621H
        O*   75EFH
        R*   75FCH
        T*   75E2H
        L*   75C8H
        X    5782H

Address

@ (MSX only)

    This table is used by the "PLAY" statement "A" to "G" command handler to translate a note number from zero to fourteen to an offset into the tone divider table at 756EH. The note itself, rather than the note number, is shown below with each offset value:

        16 ... A-
        18 ... A
        20 ... A+ or B-
        22 ... B or C-
        00 ... B+
        00 ... C
        02 ... C+ or D-
        04 ... D
        06 ... D+ or E-
        08 ... E or F-
        10 ... E+
        10 ... F
        12 ... F+ or G-
        14 ... G
        16 ... G+

Address

@ (MSX only)

    This table contains the twelve PSG divider constants required to produce the tones of octave 1. For each constant the corresponding note and frequency are shown:

        0D5DH (3421) ... C  32.698 Hz
        0C9CH (3228) ... C+ 34.653 Hz
        0BE7H (3047) ... D  36.712 Hz
        0B3CH (2876) ... D+ 38.895 HZ
        0A9BH (2715) ... E  41.201 Hz
        0A02H (2562) ... F  43.662 Hz
        0973H (2419) ... F+ 46.243 Hz
        08EBH (2283) ... G  48.997 Hz
        086BH (2155) ... G+ 51.908 Hz
        07F2H (2034) ... A  54.996 Hz
        0780H (1920) ... A+ 58.261 Hz
        0714H (1812) ... B  61.733 Hz

Address

@ (MSX only)

    This is the "PLAY" statement "V" command handler. The parameter, with a default value of eight, is placed in byte 18 of the current voice buffer without altering bit 6 of the existing contents. No music datum is generated. Address ... 759EH

Address

@ (MSX only)

    This is the "PLAY" statement "M" command handler. The parameter, with a default value of two hundred and fifty-five, is compared with the existing modulation period contained in bytes 19 and 20 of the current voice buffer. If they are the same the routine terminates with no action. Otherwise the new modulation period is placed in the voice buffer and bit 6 set in byte 18 of the voice buffer to indicate that the new value must be incorporated into the next music data packet produced. No music datum is generated. Address ... 75BEH

Address

@ (MSX only)

    This is the "PLAY" statement "S" command handler. The parameter is placed in byte 18 of the current voice buffer and bit 4 of the same byte set to indicate that the new value must be incorporated into the next music data packet produced. No music data is generated. Because of the PSG characteristics the shape and volume parameters are mutually exclusive so the same byte of the voice buffers is used for both. Address ... 75C8H

Address

@ (MSX only)

    This is the "PLAY" statement "L" command handler. The parameter, with a default value of four, is placed in byte 16 of the current voice buffer where it is used in the computation of succeeding note durations. No music datum is generated. Address ... 75E2H

Address

@ (MSX only)

    This is the "PLAY" statement "T" command handler. The parameter, with a default value of one hundred and twenty, is placed in byte 17 of the current voice buffer where it will be used in the computation of succeeding note durations. no music datum is generated. Address ... 75EFH

Address

@ (MSX only)

    This is the "PLAY" statement "O" command handler. The parameter, with a default value of four, is placed in byte 15 of the current voice buffer where it is used in the computation of succeeding note frequencies. No music datum is generated. Address ... 75FCH

Address

@ (MSX only)

    This is the "PLAY" statement "R" command handler. The length parameter, with a default value of four, is left in register pair DE and a zero tone divider value placed in register pair HL. The existing volume value is taken from byte 18 of the current voice buffer, temporarily replaced with a zero value and control transferred to the note generator (769CH). Address ... 7621H

Address

@ (MSX only)

    This is the "PLAY" statement "N" command handler. The obligatory parameter is first examined, if it is zero a rest is generated (760BH). If it is greater than ninety-six an "Illegal function call" error is generated (475AH). Otherwise twelve is repeatedly subtracted from the note number until underflow to obtain an octave number from one to nine in register E and a note number from zero to eleven in register C. Control then transfers to the note generator (7673H). Address ... 763EH

Address

@ (MSX only)

    This is the "PLAY" statement "A" to "G" command handler. The note letter is first converted into a note number from zero to fourteen, this extended range being necessary because of the redundancy implicit in the notation. The table at 755FH is then used to obtain the offset into the tone divider table and the divider constant for the note placed in register pair DE. The octave value is taken from byte 15 of the current voice buffer and the divider constant halved until the correct octave is reached. The string operand is then examined directly (56EEH) to determine whether a trailing note length parameter exists. If so it is converted (572FH) and placed in register C. If no parameter exists the default length is taken from byte 16 of the current voice buffer. The duration of the note is then computed from:

        Duration (Interrupt ticks) = 14,400/(LENGTH*TEMPO)

        Duration (Interrupt ticks) = 12,000/(LENGTH*TEMPO) (UK)

    With the normal length value (4) and tempo value (120) this gives a note duration of 30 (25 in UK) interrupt ticks of 1/60 s each (20 ms in UK) or 0.5 seconds. The string operand is then examined (56EEH) for trailing "." characters and, for each one, the duration multiplied by one and a half. Finally the resulting duration is checked and, if it is less than five interrupt ticks, it is replaced with a value of five. Thus the shortest note that can be generated on a UK machine is 0.10 seconds whatever the tempo or note length.

    The music data packet, which will be three, five or seven bytes long, is then assembled in bytes 8 to 14 of the current voice buffer prior to placing it in the queue. The duration is placed in bytes 8 and 9 of the voice buffer. The volume and flag byte is taken from byte 18 and placed in byte 10 of the voice buffer with bit 7 set to indicate a volume change to the interrupt dequeuing routine. If bit 6 of the volume byte is set then the modulation period is taken from bytes 19 and 20 and added to the data packet at bytes 11 and 12. If the tone divider value is non-zero then it is added to the data packet at bytes 11 and 12 (without a modulation period) or bytes 13 and 14 (with a modulation period). Finally the byte count is mixed into the three highest bits of byte 8 of the voice buffer to complete the preparation of the music data packet.

    If the tone divider value is zero, indicating a rest, the contents of SAVVOL are restored to byte 18 of the static buffer. The music data packet is then placed in the current queue via the PUTQ standard routine and the number of free bytes remaining checked (7521H). If less than eight bytes remain control transfers directly to the "PLAY" statement handler (748EH), otherwise control returns normally to the macro language parser. Address ... 7754H

Address

@ (MSX only)

    This is the single precision constant 14,400 (12,000 in UK machines) used in the computation of note duration. The constant is inverted: first 2 bytes at 7756H and last 2 at 7754H (both MSB LSB). Address ... 7758H

Address

@ (MSX only)

    This is the "PUT" statement handler. Register B is set to 80H and control drops into the "GET" statement handler. Address ... 775BH

Address

@ (MSX only)

    This is the "GET" statement handler. Register B is zeroed, to distinguish "GET" from "PUT" and the next program token examined. Control then transfers to the "PUT SPRITE" statement handler (7AAFH) or the Disk BASIC "GET/PUT" statement handler (6C35H). Address ... 7766H

Address

@ (MSX only)

    This is the "LOCATE" statement handler. If a column coordinate is present it is evaluated (GETBYT, 521CH) and placed in register D, otherwise the current column is taken from CSRX. If a row coordinate is present it is evaluated (GETBYT, 521CH) and placed in register E, otherwise the current row is taken from CSRY. If a cursor switch operand exists it is evaluated (GETBYT, 521CH) and register A loaded with 78H for a zero operand (OFF) and 79H for any non-zero operand (ON). The cursor is then switched by outputting ESC, 78H/79H, "5" via the OUTDO standard routine. The row and column coordinates are placed in register pair HL and the cursor position set via the POSIT standard routine. Address ... 77A5H

Address

@ (MSX only)

    This is the "STOP ON/OFF/STOP" statement handler. The address of the device's TRPTBL status byte is placed in register pair HL and control transfers to the "ON/OFF/STOP" routine (77CFH). Address ... 77ABH

Address

@ (MSX only)

    This is the "SPRITE ON/OFF/STOP" statement handler. The address of the device's TRPTBL status byte is placed in register pair HL and control transfers to the "ON/OFF/STOP" routine (77CFH). Address ... 77B1H

Address

@ (MSX only)

    This is the "INTERVAL ON/OFF/STOP" statement handler. As there is no specific "INTERVAL" token (control transfers here when an "INT" token is found) a check is first made on the program text for the characters "E" and "R" then the "VAL" token (94H). The address of the device's TRPTBL status byte is placed in register pair HL and control transfers to the "ON/OFF/STOP" routine (77CFH). Address ... 77BFH

Address

@ (MSX only)

    This is the "STRIG ON/OFF/STOP" statement handler. The trigger number, from zero to four, is evaluated (7C08H) and the address of the device's TRPTBL status byte placed in register pair HL. The "ON/OFF/STOP" token is examined and the TRPTBL status byte modified accordingly (77FEH). Control then transfers directly to the Runloop (4612H) to avoid testing for pending interrupts until the end of the next statement. Address ... 77D4H

Address

@ (MSX only)

    This is the "KEY(n) ON/OFF/STOP" statement handler. The key number, from one to ten, is evaluated (GETBYT, 521CH) and the address of the device's TRPTBL status byte placed in register pair HL. The "ON/OFF/STOP" token is examined and the TRPTBL status byte modified accordingly (77FEH). Bit 0 of the TRPTBL status byte, the ON bit, is then copied into the corresponding entry in FNKFLG for use during the interrupt keyscan and control transfers directly to the Runloop (4612H). - 194 - 5. ROM BASIC INTERPRETER Address ... 77FEH

Address

@ (MSX only)

    This routine checks for the presence of one of the interrupt switching tokens and transfers control to the appropriate routine: "ON" (631BH), "OFF" (632BH) or "STOP" (6331H). If no token is present a "Syntax error" is generated (4055H). Address ... 7810H

Address

@ (MSX only)

    This routine is used by the "ON DEVICE GOSUB" statement handler (490DH) to check the program text for a device token. Unless none of the device tokens is present, in which case Flag C is returned, the device's TRPTBL entry number is returned in register B and the maximum allowable line number operand count in register C:

        DEVICE     TRPTBL#   LINE NUMBERS
        ------------------------------------
        KEY        00        10
        STOP       10        01
        SPRITE     11        01
        STRIG      12        05
        INTERVAL   17        01

    Additionally, for "INTERVAL" only, the interval operand is evaluated (FRMQNT, 542FH) and placed in INTVAL and INTCNT. Address ... 785CH

Address

@ (MSX only)

    This routine is used by the "ON DEVICE GOSUB" statement handler (490DH) to place the address of a program line in TRPTBL. The TRPTBL entry number, supplied in register B, is multiplied by three and added to the table base to point to the relevant entry. The address, supplied in register pair DE, is then placed there LSB first, MSB second. Address ... 786CH

Address

@ (MSX only)

    This is the "KEY" statement handler. If the following character is anything other than the "LIST" token (93H) control transfers to the "KEY n" statement handler (78AEH). Each of the ten function key strings is then taken from FNKSTR and displayed via the OUTDO standard routine with a CR,LF (7328H) after each one. The DEL character (7FH) or any control character smaller than 20H is replaced with a space. Address ... 78AEH

Address

@ (MSX only)

    This is the "KEY n", "KEY(n) ON/OFF/STOP", "KEY ON" and "KEY OFF" statement handler. If the next program text character is "(" control transfers to the "KEY(n) ON/OFF/STOP" statement handler (77D4H). If it is an "ON" token (95H) control transfers to the DSPFNK standard routine and if it is an "OFF" token (EBH) to the ERAFNK standard routine. Otherwise the function key number is evaluated (GETBYT, 521CH) and the key's FNKSTR address placed in register pair DE. The string operand is evaluated (4C64H) and its storage freed (FRESTR, 67D0H). Up to fifteen characters are copied from the string to FNKSTR and unused positions padded with zero bytes. If a zero byte is found in the operand string an "Illegal function call" error is generated (475AH). Control then transfers to the FNKSB standard routine to update the function key display if it is enabled. Address ... 7900H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "TIME" function. The contents of JIFFY are placed in DAC as a single precision number (3236H). Address ... 790AH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CSRLIN" function. The contents of CSRY are decremented and placed in DAC as an integer (2E9AH). Address ... 7911H

Address

@ (MSX only)

    This is the "TIME" statement handler. The operand is evaluated (FRMQNT, 542FH) and placed in JIFFY. Address ... 791BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "PLAY" function. The numeric channel selection operand is evaluated (7C08H). If this is zero the contents of MUSICF are placed in DAC as an integer of value zero or FFFFH. Otherwise the channel number is used to select the appropriate bit of MUSICF and this is then converted to an integer as before. Address ... 7940H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "STICK" function to an operand contained in DAC. The stick number is checked (521FH) and passed to the GTSTCK standard routine in register A. The result is placed in DAC as an integer (4FCFH ) . Address ... 794CH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "STRIG" function to an operand contained in DAC. The trigger number is checked (521FH) and passed to the GTTRIG standard routine in register A. The result is placed in DAC as an integer of value zero or FFFFH. Address ... 795AH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "PDL" function to an operand contained in DAC. The paddle number is checked (521FH) and passed to the GTPDL standard routine in register A. The result is placed in DAC as an integer (4FCFH). Address ... 7969H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "PAD" function to an operand contained in DAC. The pad number is checked (521F) and passed to the GTPAD standard routine in register A. The result is placed in DAC as an integer for pads 1, 2, 5 or 6. For pads 0, 3, 4 or 7 the result is placed in DAC as an integer of value zero or FFFFH. Address ... 7980H

Address

@ (MSX only)

    This is the "COLOR" statement handler. If a foreground colour operand exists it is evaluated (GETBYT, 521CH) and placed in register E, otherwise the current foreground colour is taken from FORCLR. If a background colour operand exists it is evaluated (GETBYT, 521CH) and placed in register D, otherwise the current background colour is taken from BAKCLR. If a border colour operand exists it is evaluated (GETBYT, 521CH) and placed in BDRCLR. The foreground colour is placed in FORCLR and ATRBYT, the background colour in BAKCLR and control transfers to the CHGCLR standard routine to modify the VDP. Address ... 79CCH

Address

@ (MSX only)

    This is the "SCREEN" statement handler. If a mode operand exists it is evaluated (GETBYT, 521CH) and passed to the CHGMOD standard routine in register A. If a sprite size operand exists it is evaluated (GETBYT, 521CH) and placed in bits 0 and 1 of RG1SAV, the Workspace Area copy of VDP Mode Register 1. The VDP sprite parameters are then cleared via the CLRSPR standard routine. If a key click operand exists it is evaluated (GETBYT, 521CH) and placed in CLIKSW, zero to disable the click and non-zero to enable it. If a baud rate operand exists it is evaluated and the baud rate set (7A2DH). If a printer mode operand exists it is evaluated (GETBYT, 521CH) and placed in NTMSXP, zero for an MSX printer and non- zero for a general purpose printer. Address ... 7A2DH

Address

@ (MSX only)

    This routine is used to set the cassette baud rate. The operand is evaluated (GETBYT, 521CH) and five bytes copied from CS1200 or CS2400 to LOW as appropriate. Address ... 7A48H

Address

@ (MSX only)

    This is the "SPRITE" statement handler. If the next character is anything other than a "$" control transfers to the "SPRITE ON/OFF/STOP" statement handler (77ABH). SCRMOD is then checked and an "Illegal function call" error generated (475AH) if the screen is in 40x24 Text Mode. The sprite pattern number is evaluated and its location in the VRAM Sprite Pattern Table obtained (7AA0H). The string operand is then evaluated (4C5FH) and its storage freed (FRESTR, 67D0H). The sprite size, obtained via the GSPSIZ standard routine, is compared with the string length and, if the string is shorter than the sprite, the Sprite Pattern Table entry is first filled with zeroes via the FILVRM standard routine. Characters are then copied from the string body to the Sprite Pattern Table via the LDIRVM standard routine until the string is exhausted or the sprite is full. If the string is longer than the sprite size any excess characters are ignored. Address ... 7A84H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "SPRITE$" function. The sprite pattern number is evaluated and its location in the VRAM Sprite Pattern Table obtained (7A9FH). The sprite size, obtained via the GSPSIZ standard routine, is then placed in register pair BC to control the number of bytes copied. After checking that sufficient space is available in the String Storage Area (6627H) the sprite pattern is copied from VRAM via the LDIRMV standard routine and the result descriptor created (6654H). Note that as no check is made on the screen mode during this function some interesting side effects can be found, see below. Address ... 7A9FH

Address

@ (MSX only)

    This routine is used by the "SPRITE$" statement and function to locate a sprite pattern in the VRAM Sprite Pattern Table. The pattern number operand is evaluated (7C08H) and passed to the CALPAT standard routine in register A. The pattern address is placed in register pair DE and the routine terminates.

    Note that no check is made on the pattern number magnitude for differing sprite sizes. Pattern numbers up to two hundred and fifty-five are accepted even in 16x16 sprite mode when the maximum pattern number should be sixty-three. As a result VRAM addresses greater than 3FFFH will be produced which will wrap around into low VRAM. With the "SPRITE$" statement this will corrupt the Character Generator Table, for example:

        10 SCREEN 3,2
        20 SPRITE$(0)=STRING$(32,255)
        30 PUT SPRITE 0,(0,0), ,0
        40 SPRITE$(65)=STRING$(32,255)
        50 GOTO 50

    The above puts a real sprite in the top left of the screen and then uses an illegal statement in line 40 to corrupt the VRAM just to the right of it. The "SPRITE$" function can also be manipulated in this way and, as there is no screen mode check, up to thirty-two bytes of the Name Table can be read in 40x24 - 198 - 5. ROM BASIC INTERPRETER Text Mode, for example:

        10 SCREEN 0,2
        20 PRINT"something"
        30 A$=SPRITE$(64)
        40 PRINT A$

Address

@ (MSX only)

    This is the "GET/PUT SPRITE" statement handler, control is transferred here from the general "GET/PUT" statement handler (775BH). Register B is first checked to make sure that the statement is "PUT" and an "Illegal function call" error generated (475AH) if otherwise. SCRMOD is then checked and an "Illegal function call" error generated (475AH) if the screen is in 40x24 Text Mode. The sprite number operand, from zero to thirty-one, is evaluated (GETBYT, 521CH) and passed to the CALATR standard routine to locate the four byte attribute block in the Sprite Attribute Table. If a coordinate operand exists it is evaluated and the X coordinate placed in register pair BC, the Y coordinate in register pair DE (579CH).

    The Y coordinate LSB is written to byte 0 of the attribute block in VRAM via the WRTVRM standard routine. Bit 7 of the X coordinate is then examined to determine whether it is negative, that is off the left hand side of the screen. If so thirty two is added to the X coordinate and register B is set to 80H to set the early clock bit in the attribute block. For example an X coordinate of -1 (FFFFH) would be changed to +31 with an early clock. The X coordinate LSB is then written to byte 1 of the attribute block via the WRTVRM standard routine. Byte 3 of the attribute block is read in via the RDVRM standard routine, the new early clock bit is mixed in and it is then written back to VRAM via the WRTVRM standard routine.

    If a colour operand is present it is evaluated (GETBYT, 521CH), byte 3 of the attribute block is read in via the RDVRM standard routine the new colour code is mixed into the lowest four bits and it is written back to VRAM via the WRTVRM standard routine. If a pattern number operand exists it is evaluated (GETBYT, 521CH) and checked for magnitude against the current sprite size provided by the GSPSIZ standard routine. The maximum allowable pattern number is two hundred and fifty-five for 8x8 sprites and sixty- three for 16x16 sprites. The pattern number is written to byte 2 of the attribute block via the WRTVRM standard routine and the handler terminates. Address ... 7B37H

Address

@ (MSX only)

    This is the "VDP" statement handler. The register number operand, from zero to seven, is evaluated (7C08H) followed by the data operand (GETBYT, 521CH). The register number is placed in register C, the data value in register B and control transferred to the WRTVDP standard routine. - 199 - 5. ROM BASIC INTERPRETER Address ... 7B47H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "VDP" function. The register number operand, from zero to eight, is evaluated (7C08H) and added to RG0SAV to locate the corresponding register image in the Workspace Area. The VDP register image is then read and placed in DAC as an integer (4FCFH). Address ... 7B5AH

Address

@ (MSX only)

    This is the "BASE" statement handler. The VDP table number operand, from zero to nineteen, is evaluated (7C08H) followed by the base address operand (4C64H). After checking that the base address is less than 4000H (7BFEH) the VDP table number is used to locate the associated entry in the masking table at 7BA3H. The base address is ANDed with the mask and an "Illegal function call" error generated (475AH) if any illegal bits are set. The VDP table number is then added to TXTNAM to locate the current base address in the Workspace Area and the new base address placed there. The VDP table number is divided by five to determine which of the four screen modes the table belongs to. If this is the same as the current screen mode the new base address is also written to the VDP (7B99H). Address ... 7B99H

Address

@ (MSX only)

    This routine is used by the "BASE" statement handler to update the VDP base addresses. The current screen mode, in register A, is examined and control transfers to the SETTXT, SETT32, SETGRP or SETMLT standard routine as appropriate. Note that this is not a full VDP initialization and that the four current table addresses (NAMBAS, CGPBAS, PATBAS and ATRBAS) which are the ones actually used by the screen routines, are not updated. This can be demonstrated with the following, where the Interpreter carries on outputting to the old VRAM Name Table:

        10 SCREEN 0
        20 BASE(0)=&H400
        30 PRINT"something"
        40 FOR N=1 TO 2000:NEXT
        50 BASE(0)=0

    Note also that this routine contains a bug. While SETTXT is correctly used for 40x24 Text Mode, SETGRP is used for 32x24 Text Mode and SETMLT for Graphics Mode and Multicolour Mode. Any "BASE" statement should therefore be immediately followed by a "SCREEN" statement to perform a full initialization. Address ... 7BA3H

Address

@ (MSX only)

    This masking table is used by the "BASE" statement handler to ensure that only legal VDP base addresses are accepted. The table number and corresponding Workspace Area variable are shown with each mask:

        MASK   TABLE
        ---------------------
        03FFH  00, TXTNAM
        003FH  01, TXTCOL
        07FFH  02, TXTCGP
        007FH  03, TXTATR
        07FFH  04, TXTPAT
        03FFH  05, T32NAM
        003FH  06, T32COL
        07FFH  07, T32CGP
        007FH  08, T32ATR
        07FFH  09, T32PAT
        03FFH  10, GRPNAM
        1FFFH  11, GRPCOL
        1FFFH  12, GRPCGP
        007FH  13, GRPATR
        07FFH  14, GRPPAT
        03FFH  15, MLTNAM
        003FH  16, MLTCOL
        07FFH  17, MLTCGP
        007FH  18, MLTATR
        07FFH  19, MLTPAT

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "BASE" function. The VDP table number operand, from zero to nineteen, is evaluated (7C08H) and added to TXTNAM to locate the required Workspace Area base address. This is then placed in DAC as a single precision number (3236H). Address ... 7BE2H

Address

@ (MSX only)

    This is the "VPOKE" statement handler. The VRAM address operand is evaluated (4C64H) and checked to ensure that it is less than 4000H (7BFEH). The data operand is then evaluated (GETBYT, 521CH) and passed to the WRTVRM standard routine in register A to write to the required address. Address ... 7BF5H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "VPEEK" function to an operand contained in DAC. The VRAM address operand is checked to ensure it is less than 4000H (7BFEH). VRAM is then read via the RDVRM standard routine and the result placed in DAC as an integer (4FCFH). Address ... 7BFEH

Address

@ (MSX only)

    This routine converts a numeric operand in DAC to an integer (2F8AH) and places it in register pair HL. If the operand is equal to or greater than 4000H, and thus outside the allowable VRAM range, an "Illegal function call" error is generated (475AH). Address ... 7C08H

Address

@ (MSX only)

    This routine evaluates (GETBYT, 521CH) a parenthesized numeric operand and returns it as an integer in register A. If the operand is greater than the maximum allowable value initially supplied in register A an "Illegal function call" error is generated (475AH). Address ... 7C16H

Address

@ (MSX only)

    This is the "DSKO$" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C1BH

Address

@ (MSX only)

    This is the "SET" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C20H

Address

@ (MSX only)

    This is the "NAME" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C25H

Address

@ (MSX only)

    This is the "KILL" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C2AH

Address

@ (MSX only)

    This is the "IPL" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C2FH

Address

@ (MSX only)

    This is the "COPY" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C34H

Address

@ (MSX only)

    This is the "CMD" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C39H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "DSKF" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. - 202 - 5. ROM BASIC INTERPRETER Address ... 7C3EH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "DSKI$" function. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C43H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "ATTR$" function. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C48H

Address

@ (MSX only)

    This is the "LSET" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C4DH

Address

@ (MSX only)

    This is the "RSET" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C52H

Address

@ (MSX only)

    This is the "FIELD" statement handler. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C57H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "MKI$" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C5CH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "MKS$" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C61H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "MKD$" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C66H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CVI" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. - 203 - 5. ROM BASIC INTERPRETER Address ... 7C6BH

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CVS" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C70H

Address

@ (MSX only)

    This routine is used by the Factor Evaluator to apply the "CVD" function to an operand contained in DAC. An "Illegal function call" error is generated (475AH) on a standard MSX machine. Address ... 7C76H

Address

@ (MSX only)

    This routine completes the power-up initialization. At this point the entire Workspace Area is zeroed and only EXPTBL and SLTTBL have been initialized. A temporary stack is set at F376H and all one hundred and twelve hooks (560 bytes) filled with Z80 RET opcodes (C9H). HIMEM is set to F380H and the lowest RAM location found (7D5DH) and placed in BOTTOM. The one hundred and forty-four bytes of data commencing at 7F27H are copied to the Workspace Area from F380H to F40FH The function key strings are initialized via the INIFNK standard routine, end of BUF and NLONLY are zeroed and a comma is placed in BUFMIN and a colon in KBFMIN. The address of the MSX ROM character set is taken from locations 0004H and 0005H and placed in CGPNT+1 and PRMPRV is set to point to PRMSTK. Dummy values are placed in STKTOP, MEMSIZ and VARTAB (their correct values are not known yet), one I/O buffer is allocated (7E6BH) and the Z80 SP set (62E5H). A zero byte is placed at the base of RAM, TXTTAB is set to the following location and a "NEW" executed (6287H).

    The VDP is then initialized via the INITIO, INIT32 and CLRSPR standard routines, the cursor coordinates are set to row 11, column 10 and the sign on message "MSX system etc." is displayed (6678H). After a three second delay a search is carried out for any extension ROMs (7D75H) and a further "NEW" executed (6287H) in case a BASIC program has been run from ROM.

    Finally the identification message "MSX BASIC etc." is displayed (7D29H) and control transfers to the Interpreter Mainloop "OK" point 411FH. Address ... 7D29H

Address

@ (MSX only)

    This routine is used during power-up to enable the function key display, place the screen in 40x24 Text Mode via the INITXT standard routine, and display (6678H) the identification message "MSX BASIC etc.". The amount of free memory is then computed by subtracting the contents of VARTAB from the contents of STKTOP and displayed (3412H) followed by the "Bytes free" message. - 204 - 5. ROM BASIC INTERPRETER Address ... 7D5DH

Address

@ (MSX only)

    This routine is used during power-up to find the lowest RAM location. Starting at EF00H each byte is tested until one is found that cannot be written to or an address of 8000H is reached. The base address, rounded upwards to the nearest 256 byte boundary, is returned in register pair HL. Address ... 7D75H

Address

@ (MSX only)

    This routine is used during power-up to perform an extension ROM search. Pages 1 and 2 (4000H to BFFFH) of each slot are examined and the results placed in SLTATR. An extension ROM has the two identification characters "AB" in the first two bytes to distinguish it from RAM. Information about its properties is also present in the first sixteen bytes as follows:

    +------------------------+
    | Reserved               |  Byte 10-15
    | BASIC Text Address MSB |  Byte 9
    | BASIC Text Address LSB |  Byte 8
    | DEVICE Address MSB     |  Byte 7
    | DEVICE Address LSB     |  Byte 6
    | STATEMENT Address MSB  |  Byte 5
    | STATEMENT Address LSB  |  Byte 4
    | INITIALIZE Address MSB |  Byte 3
    | INITIALIZE Address LSB |  Byte 2
    |       42H (`B')        |  Byte 1
    |       41H (`A')        |  Byte 0
    +------------------------+

Figure 48: ROM Header.

    Each page in a given slot is examined by reading the first two bytes (7E1AH) and checking for the "AB" characters. If a ROM is present the initialization address is read (7E1AH) and control passed to it via the CALSLT standard routine. With a games ROM there may be no return to BASIC from this point. The "CALL" extended statement handler address is then read (7E1AH) and bit 5 of register B set if it is valid, that is non-zero. The extended device handler address is read (7E1AH) and bit 6 of register B set if it is valid. Finally the BASIC program text address is read (7E1AH) and bit 7 of register B set if it is valid. Register B is then copied to the relevant position in SLTATR and the search continued until no more slots remain.

    SLTATR is then examined for any extension ROM flagged as containing BASIC program text. If one is found its position in SLTATR is converted to a Slot ID (7E2AH) and the ROM permanently switched in via the ENASLT standard routine. VARTAB is set to C000H, as it is not known how large the Program Text Area is, TXTTAB is set to 8008H and BASROM made non-zero to disable the CTRL-STOP key. The system is cleared (629AH) and control transfers to the Runloop (NEWSTT, 4601H) to execute the BASIC program. - 205 - 5. ROM BASIC INTERPRETER Address ... 7E1AH

Address

@ (MSX only)

    This routine is used to read two bytes from successive locations in an extension ROM. The initial address is supplied in register pair HL and the Slot ID in register C. The bytes are read via the RDSLT standard routine and returned in register pair DE. If both are zero FLAG Z is returned. Address ... 7E2AH

Address

@ (MSX only)

    This routine converts the SLTATR position supplied in register B into the corresponding Slot ID in register C and ROM base address in register H. The position is first modified so that it runs from 0 to 63 rather than from 64 to 1, so that the required information is present in the form:

     7     6     5     4     3     2     1     0
  +-----+-----+-----------+-----------+-----------+
  |  0  |  0  |  PSLOT #  |  SSLOT #  |   PAGE #  |
  +-----+-----+-----------+-----------+-----------+

Figure 49.

    Bits 0 and 1 are shifted into the highest two bits of register H to form the address. Bits 4 and 5 are shifted to bits 0 and 1 of register C to form the Primary Slot number. Bits 2 and 3 are shifted to bits 2 and 3 of register C to form the Secondary Slot number and bit 7 of the corresponding EXPTBL entry copied to bit 7 of register C. Address ... 7E4BH

Address

@ (MSX only)

    This is the "MAXFILES" statement handler. As control transfers here when a "MAX" token (CDH) is detected the program text is first checked for a trailing "FILES" token (B7H). The buffer count operand, from zero to fifteen, is then evaluated (GETBYT, 521CH) and any existing buffers closed (6C1CH). The required number of I/O buffers are allocated (7E6BH), the system is cleared (62A7H) and control transfers directly to the Runloop (NEWSTT, 4601H). Address ... 7E6BH

Address

@ (MSX only)

    This is the I/O buffer allocation routine. It is used during power-up and by the "MAXFILES" and "CLEAR" statement handlers to allocate storage for the number of I/O buffers supplied in register A. Two hundred and sixty-seven bytes are subtracted from the contents of HIMEM for every buffer to produce a new MEMSIZ value. The size of the existing String Storage Area (initially two hundred bytes) is computed by subtracting the old contents of STKTOP from the old contents of MEMSIZ, this is then subtracted from the new MEMSIZ value to produce the new STKTOP value. A further one hundred and forty bytes are subtracted for the Z80 stack and an "Out of memory" error generated (6275H) if this address is lower than the start of the Variable Storage Area. Otherwise the buffer count is placed in MAXFIL and MEMSIZ and STKTOP set to their new values. The caller's return address is popped, the Z80 SP set to the new position and the return address pushed back onto the stack. FILTAB is then set to the start of the I/O buffer pointer block and each pointer set to point to the associated FCB. Finally the address of I/O buffer 0, the Interpreter's "LOAD" and "SAVE" buffer, is placed in NULBUF and the routine terminates. Address ... 7ED8H

Address

@ (MSX only)

    This is the plain text message "MSX system" terminated by a zero byte. Address ... 7EE4H

Address

@ (MSX only)

    This is the plain text message "version 1.0" CR,LF terminated by a zero byte. Address ... 7EF2H

Address

@ (MSX only)

    This is the plain text message "MSX BASIC " terminated by a zero byte. Address ... 7EFDH

Address

@ (MSX only)

    This is the plain text message "Copyright 1983 by Microsoft" CR,LF terminated by a zero byte. Address ... 7F1BH

Address

@ (MSX only)

    This is the plain text message " Bytes free" terminated by a zero byte. Address ... 7F27H

Address

@ (MSX only)

    This block of one hundred and forty-four data bytes is used to initialize the Workspace Area from F380H to F40FH. Address ... 7FB7H

Address

@ (MSX only)

    This seven byte patch fixes a bug in the external device parsing routine (55F8H). It checks for a zero length device name in register A and changes it to one if necessary. Address ... 7FBEH

Address

@ (MSX only)

    This section of the ROM is unused and filled with zero bytes.

8. MEMORY USAGE MAP

    A maximum of 32 KB of RAM is available to the BASIC Interpreter to hold the program text, the BASIC Variables, the Z80 stack, the I/O buffers and the internal workspace. A memory map of these areas in the power-up state is shown below:

    Expanded Slot
    Register=FFFFH ---+-----------------------+
                      |     Workspace Area    |
    HIMEM=F380H-------|                       |
                      +-----------------------+ ---+
                      |     I/O Buffer 1      |    |
                      +-----------------------+    |
                      |         FCB 1         |    |
                      +-----------------------+    |
                      |     I/O Buffer 0      |    |
    NULBUF=F177H------|                       |    |
                      +-----------------------+    +--- Disk Work Area
                      |         FCB 0         |    |
                      +-----------------------+    |
                      |     F277H (FCB 1)     |    |
                      +-----------------------+    |
    FILTAB=F16AH------|     F16EH (FCB 0)     |    |
                      +-----------------------+    |
                      |          00H          |    |
                      +-----------------------+ ---+
    MEMSIZ=F168H---+--|                       |    |
    FRETOP=F168H---+  |  String Storage Area  |    |
                      |                       |    |
    STKTOP=F0A0H------|                       |    |
                      +-----------------------+    |
                      |       Z80 Stack       |    |
                      |                       |    |
                      +-----------------------+    |
                      .                       .    |
    STREND=8003H------.                       .    |
                      +-----------------------+    +--- User Work Area
                      |  Array Storage Area   |    |
    ARYTAB=8003H------|                       |    |
                      +-----------------------+    |
                      | Variable Storage Area |    |
    VARTAB=8003H------|                       |    |
                      +-----------------------+    |
                      |   Program Text Area   |    |
                      |                       |    |
    TXTTAB=8001H------+-----------------------+    |
                      |          00H          |    |
    BOTTOM=8000H------+-----------------------+ ---+

Figure 50: Memory Usage Map 8000H to FFFFH.

    The Program Text Area is composed of tokenized program lines stored in line number order and terminated by a zero end link, when in the "NEW" state only the end link is present. The zero byte at 8000H is a dummy end of line character needed to synchronize the Runloop at the start of a program. - 208 - 6. MEMORY MAP

    The Variable and Array Storage Areas are composed of string or numeric Variables and Arrays stored in the order in which they are found in the program text. Execution speed improves marginally if Variables are declared before Arrays in a program as this reduces the amount of memory to be moved upwards.

    The Z80 stack is positioned immediately below the String Storage Area, the structure of the stack top is shown below:

          STKTOP-------|     |
                       +-----+
                       | 00H |
     Mainloop SP-------| 00H |
                       +-----+
                       | 46H |
    Statement SP-------| 01H |
                       +-----+

Figure 51: Z80 Stack Top.

    Whenever the position of the stack is altered, as a result of a "CLEAR" or "MAXFILES" statement, two zero bytes are first pushed to function as a terminator during "FOR" or "GOSUB" parameter block searches. Assuming no parameter blocks are present the Z80 SP will therefore be at STKTOP-2 within the Interpreter Mainloop and at STKTOP-4 when control transfers from the Runloop to a statement handler.

    The String Storage Area is composed of the string bodies assigned to Variables or Arrays. During expression evaluation a number of intermediate string results may also be temporarily present under the permanent string heap. The zero byte following the String Storage Area is a temporary delimiter for the "VAL" function.

    The region between the String Storage Area and HIMEM is used for I/O buffer storage. I/O buffer 0, the "SAVE" and "LOAD" buffer, is always present but the number of user buffers is determined by the "MAXFILES" statement. Each I/O buffer consists of a 9 byte FCB, whose address is contained in the table under FCB 0, followed by a 256 byte data buffer. The FCB contains the status of the I/O buffer as below:

       0     1     2     3     4     5     6     7     8
    +-----+-----+-----+-----+-----+-----+-----+-----+-----+
    | Mod | 00H | 00H | 00H | DEV | 00H | POS | 00H | PPS |
    +-----+-----+-----+-----+-----+-----+-----+-----+-----+

Figure 52 : File Control Block.

    The MOD byte holds the buffer mode, the DEV byte the device code, the POS byte the current position in the buffer (0 to 255) and the PPS byte the "PRINT" position. The remainder of the FCB is unused on a standard MSX machine. - 209 - 6. MEMORY MAP Workspace Area

    The section of the Workspace Area from F380H to FD99H holds the BIOS/Interpreter variables. These are listed on the following pages in standard assembly language form:

    F380H RDPRIM: OUT (0A8H),A ; Set new Primary Slot
    F382H         LD E,(HL)    ; Read memory
    F383H         JR WRPRM1    ; Restore old Primary Slot

    This routine is used by the RDSLT standard routine to switch Primary Slots and read a byte from memory. The new Primary Slot Register setting is supplied in register A, the old setting in register D and the byte read returned in register E.

    F385H WRPRIM: OUT (0A8H),A ; Set new Primary Slot
    F387H         LD (HL),E    ; Write to memory
    F388H WRPRM1: LD A,D       ; Get old setting
    F389H         OUT (0A8H),A ; Restore old Primary Slot
    F38BH         RET

    This routine is used by the WRSLT standard routine to switch Primary Slots and write a byte to memory. The new Primary Slot Register setting is supplied in register A, the old setting in register D and the byte to write in register E.

    F38CH CLPRIM: OUT (0A8H),A ; Set new Primary Slot
    F38EH         EX AF,AF'    ; Swap to AF for call
    F38FH         CALL CLPRM1  ; Do it
    F392H         EX AF,AF'    ; Swap to AF
    F393H         POP AF       ; Get old setting
    F394H         OUT (0A8H),A ; Restore old Primary Slot
    F396H         EX AF,AF'    ; Swap to AF
    F397H         RET
    F398H CLPRM1: JP (IX)

    This routine is used by the CALSLT standard routine to switch Primary Slots and call an address. The new Primary Slot Register setting is supplied in register A, the old setting on the Z80 stack and the address to call in register pair IX.

    F39AH USRTAB: DEFW 475AH   ; USR 0
    F39CH         DEFW 475AH   ; USR 1
    F39EH         DEFW 475AH   ; USR 2
    F3A0H         DEFW 475AH   ; USR 3
    F3A2H         DEFW 475AH   ; USR 4
    F3A4H         DEFW 475AH   ; USR 5
    F3A6H         DEFW 475AH   ; USR 6
    F3A8H         DEFW 475AH   ; USR 7
    F3AAH         DEFW 475AH   ; USR 8
    F3ACH         DEFW 475AH   ; USR 9

    These ten variables contain the "USR" function addresses. Their values are set to the Interpreter "Illegal function call" error generator at power-up and thereafter only altered by the "DEFUSR" statement. F3AEH LINL40: DEFB 39 ; 39 in Japanese MSX, 37 in European.

    This variable contains the 40x24 Text Mode screen width. Its value is set at power-up and thereafter only altered by the "WIDTH" statement. F3AFH LINL32: DEFB 29

    This variable contains the 32x24 Text Mode screen width. Its value is set at power-up and thereafter only altered by the "WIDTH" statement. F3B0H LINLEN: DEFB 39 ; 39 in Japanese MSX, 37 in European.

    This variable contains the current text mode screen width. Its value is set from LINL40 or LINL32 whenever the VDP is initialized to a text mode via the INITXT or INIT32 standard routines. F3B1H CRTCNT: DEFB 24

    This variable contains the number of rows on the screen. Its value is set at power-up and thereafter unaltered. F3B2H CLMLST: DEFB 14

    This variable contains the minimum number of columns that must still be available on a line for a data item to be "PRINT"ed, if less space is available a CR,LF is issued first. Its value is set at power-up and thereafter only altered by the "WIDTH" and "SCREEN" statements.

    F3B3H TXTNAM: DEFW 0000H   ; Text Name Table Base
    F3B5H TXTCOL: DEFW 0000H   ; Text Colour Table Base
    F3B7H TXTCGP: DEFW 0800H   ; Text Character Generator Pattern Base
    F3B9H TXTATR: DEFW 0000H   ; Text Sprite Attribute Base
    F3BBH TXTPAT: DEFW 0000H   ; Text Sprite Pattern Base

    These five variables contain the 40x24 Text Mode VDP base addresses. Their values are set at power-up and thereafter only altered by the "BASE" statement.

    F3BDH T32NAM: DEFW 1800H   ; Text 32 Name Table Base
    F3BFH T32COL: DEFW 2000H   ; Text 32 Colour Table Base
    F3C1H T32CGP: DEFW 0000H   ; Text 32 Character Generator Pattern Base
    F3C3H T32ATR: DEFW 1B00H   ; Text 32 Sprite Attribute Base
    F3C5H T32PAT: DEFW 3800H   ; Text 32 Sprite Pattern Base

    These five variables contain the 32x24 Text Mode VDP base addresses. Their values are set at power-up and thereafter only altered by the "BASE" statement. - 211 - 6. MEMORY MAP

    F3C7H GRPNAM: DEFW 1800H   ; Graphic Name Table Base
    F3C9H GRPCOL: DEFW 2000H   ; Graphic Colour Table Base
    F3CBH GRPCGP: DEFW 0000H   ; Graphic Character Generator Pattern Base
    F3CDH GRPATR: DEFW 1B00H   ; Graphic Sprite Attribute Base
    F3CFH GRPPAT: DEFW 3800H   ; Graphic Sprite Pattern Base

    These five variables contain the Graphics Mode VDP base addresses. Their values are set at power-up and thereafter only altered by the "BASE" statement.

    F3D1H MLTNAM: DEFW 0800H   ; Multicolour Name Table Base
    F3D3H MLTCOL: DEFW 0000H   ; Multicolour Colour Table Base
    F3D5H MLTCGP: DEFW 0000H   ; Multicolour Character Generator Pattern Base
    F3D7H MLTATR: DEFW 1B00H   ; Multicolour Sprite Attribute Base
    F3D9H MLTPAT: DEFW 3800H   ; Multicolour Sprite Pattern Base

    These five variables contain the Multicolour Mode VDP base addresses. Their values are set at power-up and thereafter only altered by the "BASE" statement. F3DBH CLIKSW: DEFB 01H

    This variable controls the interrupt handler key click: 00H=Off, NZ=On. Its value is set at power-up and thereafter only altered by the "SCREEN" statement. F3DCH CSRY : DEFB 01H

    This variable contains the row coordinate (from 1 to CTRCNT) of the text mode cursor. F3DDH CSRX : DEFB 01H

    This variable contains the column coordinate (from 1 to LINLEN) of the text mode cursor. Note that the BIOS cursor coordinates for the home position are 1,1 whatever the screen width. F3DEH CNSDFG: DEFB FFH

    This variable contains the current state of the function key display: 00H=Off, NZ=On.

    F3DFH RG0SAV: DEFB 00H
    F3E0H RG1SAV: DEFB F0H
    F3E1H RG2SAV: DEFB 00H
    F3E2H RG3SAV: DEFB 00H
    F3E3H RG4SAV: DEFB 01H
    F3E4H RG5SAV: DEFB 00H
    F3E5H RG6SAV: DEFB 00H
    F3E6H RG7SAV: DEFB F4H

    These eight variables mimic the state of the eight write-only MSX VDP Mode Registers. The values shown are for 40x24 Text Mode. - 212 - 6. MEMORY MAP F3E7H STATFL: DEFB CAH

    This variable is continuously updated by the interrupt handler with the contents of the VDP Status Register. F3E8H TRGFLG: DEFB F1H

    This variable is continuously updated by the interrupt handler with the state of the four joystick trigger inputs and the space key. F3E9H FORCLR: DEFB 0FH ; White

    This variable contains the current foreground colour. Its value is set at power-up and thereafter only altered by the "COLOR" statement. The foreground colour is used by the CLRSPR standard routine to set the sprite colour and by the CHGCLR standard routine to set the 1 pixel colour in the text modes. It also functions as the graphics ink colour as it is copied to ATRBYT by the GRPPRT standard routine and used throughout the Interpreter as the default value for any optional colour operand. * Note: the default value may change with version. F3EAH BAKCLR: DEFB 04H ; Dark blue

    This variable contains the current background colour. Its value is set at power-up and thereafter only altered by the "COLOR" statement. The background colour is used by the CLS standard routine to clear the screen in the graphics modes and by the CHGCLR standard routine to set the 0 pixel colour in the text modes. * Note: the default value may change with version. F3EBH BDRCLR: DEFB 04H ; Dark blue

    This variable contains the current border colour. Its value is set at power-up and thereafter only altered by the "COLOR" statement. The border colour is used by the CHGCLR standard routine in 32x24 Text Mode, Graphics Mode and Multicolour Mode to set the border colour. * Note: the default value may change with version. F3ECH MAXUPD: DEFB C3H F3EDH DEFW 0000H

    These two bytes are filled in by the "LINE" statement handler to form a Z80 JP to the RIGHTC, LEFTC, UPC or DOWNC standard routines. F3EFH MINUPD: DEFB C3H F3F0H DEFW 0000H

    These two bytes are filled in by the "LINE" statement handler to form a Z80 JP to the RIGHTC, LEFTC, UPC or DOWNC standard routines. - 213 - 6. MEMORY MAP F3F2H ATRBYT: DEFB 0FH

    This variable contains the graphics ink colour used by the SETC and NSETCX standard routines. F3F3H QUEUES: DEFW F959H

    This variable contains the address of the control blocks for the three music queues. Its value is set at power-up and thereafter unaltered. F3F5H FRCNEW: DEFB FFH

    This variable contains a flag to distinguish the two statements in the "CLOAD/CLOAD?" statement handler: 00H=CLOAD, FFH=CLOAD?. F3F6H SCNCNT: DEFB 01H

    This variable is used as a counter by the interrupt handler to control the rate at which keyboard scans are performed. F3F7H REPCNT: DEFB 01H

    This variable is used as a counter by the interrupt handler to control the key repeat rate. F3F8H PUTPNT: DEFW FBF0H

    This variable contains the address of the put position in KEYBUF. F3FAH GETPNT: DEFW FBF0H

    This variable contains the address of the get position in KEYBUF.

    F3FCH CS1200: DEFB 53H     ; LO cycle 1st half
    F3FDH         DEFB 5CH     ; LO cycle 2nd half
    F3FEH         DEFB 26H     ; HI cycle 1st half
    F3FFH         DEFB 2DH     ; HI cycle 2nd half
    F400H         DEFB 0FH     ; Header cycle count

    These five variables contain the 1200 baud cassette parameters. Their values are set at power-up and thereafter unaltered.

    F401H CS2400: DEFB 25H     ; LO cycle 1st half
    F402H         DEFB 2DH     ; LO cycle 2nd half
    F403H         DEFB 0EH     ; HI cycle 1st half
    F404H         DEFB 16H     ; HI cycle 2nd half
    F405H         DEFB 1FH     ; Header cycle count

    These five variables contain the 2400 baud cassette parameters. Their values are set at power-up and thereafter unaltered. - 214 - 6. MEMORY MAP

    F406H LOW   : DEFB 53H     ; LO cycle 1st half
    F407H         DEFB 5CH     ; LO cycle 2nd half
    F408H HIGH  : DEFB 26H     ; HI cycle 1st half
    F409H         DEFB 2DH     ; HI cycle 2nd half
    F40AH HEADER: DEFB 0FH     ; Header cycle count

    These five variables contain the current cassette parameters. Their values are set to 1200 baud at power-up and thereafter only altered by the "CSAVE" and "SCREEN" statements. F40BH ASPCT1: DEFW 0100H

    This variable contains the reciprocal of the default "CIRCLE" aspect ratio multiplied by 256. Its value is set at power-up and thereafter unaltered. F40DH ASPCT2: DEFW 01C0H

    This variable contains the default "CIRCLE" aspect ratio multiplied by 256. Its value is set at power-up and thereafter unaltered. The aspect ratio is present in two forms so that the "CIRCLE" statement handler can select the appropriate one immediately rather than needing to examine and possibly reciprocate it as is the case with an operand in the program text.

    F40FH ENDPRG: DEFB ":"
    F410H         DEFB 00H
    F411H         DEFB 00H
    FE12H         DEFB 00H
    F413H         DEFB 00H

    These five bytes form a dummy program line. Their values are set at power-up and thereafter unaltered. The line exists in case an error occurs in the Interpreter Mainloop before any tokenized text is available in KBUF. If an "ON ERROR GOTO" is active at this time then it provides some text for the "RESUME" statement to terminate on. F414H ERRFLG: DEFB 00H

    This variable is used by the Interpreter error handler to save the error number. F415H LPTPOS: DEFB 00H

    This variable is used by the "LPRINT" statement handler to hold the current position of the printer head. F416H PRTFLG: DEFB 00H

    This variable determines whether the OUTDO standard routine directs its output to the screen or to the printer: 00H=Screen, 01H=Printer. - 215 - 6. MEMORY MAP F417H NTMSXP: DEFB 00H

    This variable determines whether the OUTDO standard routine will replace headered graphics characters directed to the printer with spaces: 00H=Graphics, NZ=Spaces. Its value is set at power-up and thereafter only altered by the "SCREEN" statement. F418H RAWPRT: DEFB 00H

    This variable determines whether the OUTDO standard routine will modify control and headered graphics characters directed to the printer: 00H=Modify, NZ=Raw. Its value is set at power- up and thereafter unaltered. F419H VLZADR: DEFW 0000H F41BH VLZDAT: DEFB 00H

    These variables contain the address and value of any character temporarily removed by the "VAL" function. F41CH CURLIN: DEFW FFFFH

    This variable contains the current Interpreter line number. A value of FFFFH denotes direct mode. F41EH KBFMIN: DEFB ":"

    This byte provides a dummy prefix to the tokenized text contained in KBUF. Its function is similar to that of ENDPRG but is used for the situation where an error occurs within a direct statement. F41FH KBUF : DEFS 318

    This buffer contains the tokenized form of the input line collected by the Interpreter Mainloop. When a direct statement is executed the contents of this buffer form the program text. F55DH BUFMIN: DEFB ","

    This byte provides a dummy prefix to the text contained in BUF. It is used to synchronize the "INPUT" statement handler as it starts to analyze the input text. F55EH BUF : DEFS 259

    This buffer contains the text collected from the console by the INLIN standard routine. F661H TTYPOS: DEFB 00H

    This variable is used by the "PRINT" statement handler to hold the current screen position (Teletype!). - 216 - 6. MEMORY MAP F662H DIMFLG: DEFB 00H

    This variable is normally zero but is set by the "DIM" statement handler to control the operation of the variable search routine. F663H VALTYP: DEFB 02H

    This variable contains the type code of the operand currently contained in DAC: 2=integer, 3=String, 4=Single Precision, 8=Double Precision. F664H DORES : DEFB 00H

    This variable is normally zero but is set to prevent the tokenization of unquoted keywords following a "DATA" token. F665H DONUM : DEFB 00H

    This variable is normally zero but is set when a numeric constant follows one of the keywords GOTO, GOSUB, THEN, etc., and must be tokenized to the special line number operand form. F666H CONTXT: DEFW 0000H

    This variable is used by the CHRGTR standard routine to save the address of the character following a numeric constant in the program text. F668H CONSAV: DEFB 00H

    This variable is used by the CHRGTR standard routine to save the token of a numeric constant found in the program text. F669H CONTYP: DEFB 00H

    This variable is used by the CHRGTR standard routine to save the type of a numeric constant found in the program text. F66AH CONLO : DEFS 8

    This buffer is used by the CHRGTR standard routine to save the value of a numeric constant found in the program text. F672H MEMSIZ: DEFW F168H

    This variable contains the address of the top of the String Storage Area. Its value is set at power-up and thereafter only altered by the "CLEAR" and "MAXFILES" statements. F674H STKTOP: DEFW F0A0H

    This variable contains the address of the top of the Z80 stack. Its value is set at power-up to MEMSIZ-200 and thereafter only altered by the "CLEAR" and "MAXFILES" statements. F676H TXTTAB: DEFW 8001H

    This variable contains the address of the first byte of the Program Text Area. Its value is set at power-up and thereafter unaltered. F678H TEMPPT: DEFW F67AH

    This variable contains the address of the next free location in TEMPST. F67AH TEMPST: DEFS 30

    This buffer is used to store string descriptors. It functions as a stack with string producers pushing their results and string consumers popping them. F698H DSCTMP: DEFS 3

    This buffer is used by the string functions to hold a result descriptor while it is being constructed. F69BH FRETOP: DEFW F168H

    This variable contains the address of the next free location in the String Storage Area. When the area is empty FRETOP is equal to MEMSIZ. F69DH TEMP3 : DEFW 0000H

    This variable is used for temporary storage by various parts of the Interpreter. F69FH TEMP8 : DEFW 0000H

    This variable is used for temporary storage by various parts of the Interpreter. F6A1H ENDFOR: DEFW 0000H

    This variable is used by the "FOR" statement handler to hold the end of statement address during construction of a parameter block. F6A3H DATLIN: DEFW 0000H

    This variable contains the line number of the current "DATA" item in the program text. F6A5H SUBFLG: DEFB 00H

    This variable is normally zero but is set by the "ERASE", "FOR", "FN" and "DEF FN" handlers to control the processing of subscripts by the variable search routine. F6A6H FLGINP: DEFB 00H

    This variable contains a flag to distinguish the two statements in the "READ/INPUT" statement handler: 00H=INPUT, NZ=READ. F6A7H TEMP : DEFW 0000H

    This variable is used for temporary storage by various parts of the Interpreter. F6A9H PTRFLG: DEFB 00H

    This variable is normally zero but is set if any line number operands in the Program Text Area have been converted to pointers. F6AAH AUTFLG: DEFB 00H

    This variable is normally zero but is set when "AUTO" mode is turned on. F6ABH AUTLIN: DEFW 0000H

    This variable contains the current "AUTO" line number. F6ADH AUTINC: DEFW 0000H

    This variable contains the current "AUTO" line number increment. F6AFH SAVTXT: DEFW 0000H

    This variable is updated by the Runloop at the start of every statement with the current location in the program text. It is used during error recovery to set ERRTXT for the "RESUME" statement handler and OLDTXT for the "CONT" statement handler. F6B1H SAVSTK: DEFW F09EH

    This variable is updated by the Runloop at the start of every statement with the current Z80 SP for error recovery purposes. F6B3H ERRLIN: DEFW 0000H

    This variable is used by the error handler to hold the line number of the program line generating an error. F6B5H DOT : DEFW 0000H

    This variable is updated by the Mainloop and the error handler with the current line number for use with the "." parameter. - 219 - 6. MEMORY MAP F6B7H ERRTXT: DEFW 0000H

    This variable is updated from SAVTXT by the error handler for use by the "RESUME" statement handler. F6B9H ONELIN: DEFW 0000H

    This variable is set by the "ON ERROR GOTO" statement handler with the address of the program line to execute when an error occurs. F6BBH ONEFLG: DEFB 00H

    This variable is normally zero but is set by the error handler when control transfers to an "ON ERROR GOTO" statement. This is to prevent a loop developing if an error occurs inside the error recovery statements. F6BCH TEMP2 : DEFW 0000H

    This variable is used for temporary storage by various parts of the Interpreter. F6BEH OLDLIN: DEFW 0000H

    This variable contains the line number of the terminating program line. It is set by the "END" and "STOP" statement handlers for use with the "CONT" statement. F6C0H OLDTXT: DEFW 0000H

    This variable contains the address of the terminating program statement. F6C2H VARTAB: DEFW 8003H

    This variable contains the address of the first byte of the Variable Storage Area. F6C4H ARYTAB: DEFW 8003H

    This variable contains the address of the first byte of the Array Storage Area. F6C6H STREND: DEFW 8003H

    This variable contains the address of the byte following the Array Storage Area. F6C8H DATPTR: DEFW 8000H

    This variable contains the address of the current "DATA" item in the program text. - 220 - 6. MEMORY MAP

    F6CAH DEFTBL: DEFB 08H     ; A
    F6CBH         DEFB 08H     ; B
    F6CCH         DEFB 08H     ; C
    F6CDH         DEFB 08H     ; D
    F6CEH         DEFB 08H     ; E
    F6CFH         DEFB 08H     ; F
    F6D0H         DEFB 08H     ; G
    F6D1H         DEFB 08H     ; H
    F6D2H         DEFB 08H     ; I
    F6D3H         DEFB 08H     ; J
    F6D4H         DEFB 08H     ; K
    F6D5H         DEFB 08H     ; L
    F6D6H         DEFB 08H     ; M
    F6D7H         DEFB 08H     ; N
    F6D8H         DEFB 08H     ; O
    F6D9H         DEFB 08H     ; P
    F6DAH         DEFB 08H     ; Q
    F6DBH         DEFB 08H     ; R
    F6DCH         DEFB 08H     ; S
    F6DDH         DEFB 08H     ; T
    F6DEH         DEFB 08H     ; U
    F6DFH         DEFB 08H     ; V
    F6E0H         DEFB 08H     ; W
    F6E1H         DEFB 08H     ; X
    F6E2H         DEFB 08H     ; Y
    F6E3H         DEFB 08H     ; Z

    These twenty-six variables contain the default type for each group of BASIC Variables. Their values are set to double precision at power-up, "NEW" and "CLEAR" and thereafter altered only by the "DEF" group of statements. F6E4H PRMSTK: DEFW 0000H

    This variable contains the base address of the previous "FN" parameter block on the Z80 stack. It is used during string garbage collection to travel from block to block on the stack. F6E6H PRMLEN: DEFW 0000H

    This variable contains the length of the current "FN" parameter block in PARM1. F6E8H PARM1 : DEFS 100

    This buffer contains the local Variables belonging to the "FN" function currently being evaluated. F74CH PRMPRV: DEFW F6E4H

    This variable contains the address of the previous "FN" parameter block. It is actually a constant used to ensure that string garbage collection commences with the current parameter block before proceeding to those on the stack. - 221 - 6. MEMORY MAP F74EH PRMLN2: DEFW 0000H

    This variable contains the length of the "FN" parameter block being constructed in PARM2 F750H PARM2 : DEFS 100

    This buffer is used to construct the local Variables owned by the current "FN" function. F7B4H PRMFLG: DEFB 00H

    This variable is used during a Variable search to indicate whether local or global Variables are being examined. F7B5H ARYTA2: DEFW 0000H

    This variable is used during a Variable search to hold the termination address of the storage area being examined. F7B7H NOFUNS: DEFB 00H

    This variable is normally zero but is set by the "FN" function handler to indicate to the variable search routine that local Variables are present. F7B8H TEMP9 : DEFW 0000H

    This variable is used for temporary storage by various parts of the Interpreter. F7BAH FUNACT: DEFW 0000H

    This variable contains the number of currently active "FN functions. F7BCH SWPTMP: DEFS 8

    This buffer is used to hold the first operand in a "SWAP" statement. F7C4H TRCFLG: DEFB 00H

    This variable is normally zero but is set by the "TRON" statement handler to turn on the trace facility. F7C5H FBUFFR: DEFS 43

    This buffer is used to hold the text produced during numeric output conversion. F7F0H DECTMP: DEFW 0000H

    This variable is used for temporary storage by the double precision division routine. F7F2H DECTM2: DEFW 0000H

    This variable is used for temporary storage by the double precision division routine. F7F4H DECCNT: DEFB 00H

    This variable is used by the double precision division routine to hold the number of non-zero bytes in the mantissa of the second operand. F7F6H DAC : DEFS 16 ; Decimal Accumulator

    This buffer functions as the Interpreter's primary accumulator during expression evaluation. F806H HOLD8 : DEFS 65

    This buffer is used by the double precision multiplication routine to hold the multiples of the first operand. F847H ARG : DEFS 16

    This buffer functions as the Interpreter's secondary accumulator during expression evaluation. F857H RNDX : DEFS 8

    This buffer contains the current double precision random number. F85FH MAXFIL: DEFB 01H

    This variable contains the number of currently allocated user I/O buffers. Its value is set to 1 at power-up and thereafter only altered by the "MAXFILES" statement. F860H FILTAB: DEFW F16AH

    This variable contains the address of the pointer table for the I/O buffer FCBs. F862H NULBUF: DEFW F177H

    This variable contains the address of the first byte of the data buffer belonging to I/O buffer 0. F864H PTRFIL: DEFW 0000H

    This variable contains the address of the currently active I/O buffer FCB. - 223 - 6. MEMORY MAP F866H FILNAM: DEFS 11

    This buffer holds a user-specified filename. It is eleven characters long to allow for disc file specs such as "FILENAME.BAS". F871H FILNM2: DEFS 11

    This buffer holds a filename read from an I/O device for comparison with the contents of FILNAM. F87CH NLONLY: DEFB 00H

    This variable is normally zero but is set during a program "LOAD". Bit 0 is used to prevent I/O buffer 0 being closed during loading and bit 7 to prevent the user I/O buffers being closed if auto-run is required. F87DH SAVEND: DEFW 0000H

    This variable is used by the "BSAVE" statement handler to hold the end address of the memory block to be saved. F87FH FNKSTR: DEFS 160

    This buffer contains the ten sixteen-character function key strings. Their values are set at power-up and thereafter only altered by the "KEY" statement. F91FH CGPNT : DEFB 00H ; Slot ID F920H DEFW 1BBFH ; Address

    These variables contain the location of the character set copied to the VDP by the INITXT and INIT32 standard routines. Their values are set to the MSX ROM character set at power-up and thereafter unaltered. F922H NAMBAS: DEFW 0000H

    This variable contains the current text mode VDP Name Table base address. Its value is set from TXTNAM or T32NAM whenever the VDP is initialized to a text mode via the INITXT or INIT32 standard routines. F924H CGPBAS: DEFW 0800H

    This variable contains the current text mode VDP Character Pattern Table base address. Its value is set from TXTCGP or T32CGP whenever the VDP is initialized to a text mode via the INITXT or INIT32 standard routines. F926H PATBAS: DEFW 3800H

    This variable contains the current VDP Sprite Pattern Table base address. Its value is set from T32PAT, GRPPAT or MLTPAT whenever the VDP is initialized via the INIT32, INIGRP or INIMLT standard routines. F928H ATRBAS: DEFW 1B00H

    This variable contains the current VDP Sprite Attribute Table base address. Its value is set from T32ATR, GRPATR or MLTATR whenever the VDP is initialized via the INIT32, INIGRP or INIMLT standard routines. F92AH CLOC : DEFW 0000H ; Pixel location F92CH CMASK : DEFB 80H ; Pixel Mask

    These variables contain the current pixel physical address used by the RIGHTC, LEFTC, UPC, TUPC, DOWNC, TDOWNC, FETCHC, STOREC, READC, SETC, NSETCX, SCANR and SCANL standard routines. CLOC holds the address of the byte containing the current pixel and CMASK defines the pixel within that byte. F92DH MINDEL: DEFW 0000H

    This variable is used by the "LINE" statement handler to hold the minimum difference between the end points of the line. F92FH MAXDEL: DEFW 0000H

    This variable is used by the "LINE" statement handler to hold the maximum difference between the end points of the line. F931H ASPECT: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the current aspect ratio. This is stored as a single byte binary fraction so an aspect ratio of 0.75 would become 00C0H. The MSB is only required if the aspect ratio is exactly 1.00, that is 0100H. F933H CENCNT: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the point count of the end angle. F935H CLINEF: DEFB 00H

    This variable is used by the "CIRCLE" statement handler to hold the two line flags. Bit 0 is set if a line is required from the start angle to the centre and bit 7 set if one is required from the end angle. F936H CNPNTS: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the number of points within a forty-five degree segment. - 225 - 6. MEMORY MAP F938H CPLOTF: DEFB 00H

    This variable is normally zero but is set by the "CIRCLE" statement handler if the end angle is smaller than the start angle. It is used to determine whether the pixels should be set "inside" the angles or "outside" them. F939H CPCNT : DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the point count within the current forty-five degree segment, this is in fact the Y coordinate. F93BH CPCNT8: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the total point count of the present position. F93DH CRCSUM: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler as the point computation counter. F93FH CSTCNT: DEFW 0000H

    This variable is used by the "CIRCLE" statement handler to hold the point count of the start angle. F941H CSCLXY: DEFB 00H

    This variable is used by the "CIRCLE" statement handler as a flag to determine in which direction the elliptic squash is to be applied: 00H=Y, 01H=X. F942H CSAVEA: DEFW 0000H

    This variable is used for temporary storage by the SCANR standard routine. F944H CSAVEM: DEFB 00H

    This variable is used for temporary storage by the SCANR standard routine. F945H CXOFF : DEFW 0000H

    This variable is used for temporary storage by the "CIRCLE" statement handler. F947H CYOFF : DEFW 0000H

    This variable is used for temporary storage by the "CIRCLE" statement handler. - 226 - 6. MEMORY MAP F949H LOHMSK: DEFB 00H

    This variable is used by the "PAINT" statement handler to hold the leftmost position of a LH excursion. F94AH LOHDIR: DEFB 00H

    This variable is used by the "PAINT" statement handler to hold the new paint direction required by a LH excursion. F94BH LOHADR: DEFW 0000H

    This variable is used by the "PAINT" statement handler to hold the leftmost position of a LH excursion. F94DH LOHCNT: DEFW 0000H

    This variable is used by the "PAINT" statement handler to hold the size of a LH excursion. F94FH SKPCNT: DEFW 0000H

    This variable is used by the "PAINT" statement handler to hold the skip count returned by the SCANR standard routine. F951H MOVCNT: DEFW 0000H

    This variable is used by the "PAINT" statement handler to hold the movement count returned by the SCANR standard routine. F953H PDIREC: DEFB 00H

    This variable is used by the "PAINT" statement handler to hold the current paint direction: 40H=Down, C0H=Up, 00H=Terminate. F954H LFPROG: DEFB 00H

    This variable is normally zero but is set by the "PAINT" statement handler if there has been any leftwards progress. F955H RTPROG: DEFB 00H

    This variable is normally zero but is set by the "PAINT" statement handler if there has been any rightwards progress. F956H MCLTAB: DEFW 0000H

    This variable contains the address of the command table to be used by the macro language parser. The "DRAW" table is at 5D83H and the "PLAY" table at 752EH. F958H MCLFLG: DEFB 00H

    This variable is zero if the macro language parser is being used by the "DRAW", statement handler and non-zero if it is being used by "PLAY".

    F959H QUETAB: DEFB 00H     ; AQ Put position
    F95AH         DEFB 00H     ; AQ Get position
    F95BH         DEFB 00H     ; AQ Putback flag
    F95CH         DEFB 7FH     ; AQ Size
    F95DH         DEFW F975H   ; AQ Address

    F95FH         DEFB 00H     ; BQ Put position
    F960H         DEFB 00H     ; BQ Get position
    F961H         DEFB 00H     ; BQ Putback flag
    F962H         DEFB 7FH     ; BQ Size
    F963H         DEFW F9F5H   ; BQ Address

    F965H         DEFB 00H     ; CQ Put position
    F966H         DEFB 00H     ; CQ Get position
    F967H         DEFB 00H     ; CQ Putback flag
    F968H         DEFB 7FH     ; CQ Size
    F969H         DEFW FA75H   ; CQ Address

    F96BH         DEFB 00H     ; RQ Put position
    F96CH         DEFB 00H     ; RQ Get position
    F96DH         DEFB 00H     ; RQ Putback flag
    F96EH         DEFB 00H     ; RQ Size
    F96FH         DEFW 0000H   ; RQ Address

    These twenty-four variables form the control blocks for the three music queues (VOICAQ, VOICBQ and VOICCQ) and the RS232 queue. The three music control blocks are initialized by the GICINI standard routine and thereafter maintained by the interrupt handler and the PUTQ standard routine. The RS232 control block is unused in the current MSX ROM.

    F971H QUEBAK: DEFB 00H     ; AQ Putback character
    F972H         DEFB 00H     ; BQ Putback character
    F973H         DEFB 00H     ; CQ Putback character
    F974H         DEFB 00H     ; RQ Putback character

    These four variables are used to hold any unwanted character returned to the associated queue. Although the putback facility is implemented in the MSX ROM it is currently unused.

    F975H VOICAQ: DEFS 128     ; Voice A queue
    F9F5H VOICBQ: DEFS 128     ; Voice B queue
    FA75H VOICCQ: DEFS 128     ; Voice C queue
    FAF5H RS2IQ : DEFS 64      ; RS232 queue

    These four buffers contain the three music queues and the RS232 queue, the latter is unused in MSX and has been redefined in MSX2. FB35H PRSCNT: DEFB 00H

    This variable is used by the "PLAY" statement handler to count the number of completed operand strings. Bit 7 is also set after each of the three operands has been parsed to prevent repeated activation of the STRTMS standard routine. FB36H SAVSP : DEFW 0000H

    This variable is used by the "PLAY" statement handler to save the Z80 SP before control transfers to the macro language parser. Its value is compared with the SP on return to determine whether any data has been left on the stack because of a queue-full termination by the parser. FB38H VOICEN: DEFB 00H

    This variable contains the current voice number being processed by the "PLAY" statement handler. The values 0, 1 and 2 correspond to PSG channels A, B and C. FB39H SAVVOL: DEFW 0000H

    This variable is used by the "PLAY" statement "R" command handler to save the current volume setting while a zero- amplitude rest is generated. FB3BH MCLLEN: DEFB 00H

    This variable is used by the macro language parser to hold the length of the string operand being parsed. FB3CH MCLPTR: DEFW 0000H

    This variable is used by the macro language parser to hold the address of the string operand being parsed. FB3EH QUEUEN: DEFB 00H

    This variable is used by the interrupt handler to hold the number of the music queue currently being processed. The values 0, 1 and 2 correspond to PSG channels A, B and C. FB3FH MUSICF: DEFB 00H

    This variable contains three bit flags set by the STRTMS standard routine to initiate processing of a music queue by the interrupt handler. Bits 0, 1 and 2 correspond to VOICAQ, VOICBQ and VOICCQ. FB40H PLYCNT: DEFB 00H

    This variable is used by the STRTMS standard routine to hold the number of "PLAY" statement sequences currently held in the music queues. It is examined when all three end of queue marks have been found for one sequence to determine whether dequeueing should be restarted. - 229 - 6. MEMORY MAP FB41H VCBA : DEFW 0000H ; Duration counter FB43H DEFB 00H ; String length FB44H DEFW 0000H ; String address FB46H DEFW 0000H ; Stack data address FB48H DEFB 00H ; Music packet length FB49H DEFS 7 ; Music packet FB50H DEFB 04H ; Octave FB51H DEFB 04H ; Length FB52H DEFB 78H ; Tempo FB53H DEFB 88H ; Volume FB54H DEFW 00FFH ; Envelope period FB56H DEFS 16 ; Space for stack data

    This thirty-seven byte buffer is used by the "PLAY" statement handler to hold the current parameters for voice A. FB66H VCBB : DEFS 37

    This buffer is used by the "PLAY" statement handler to hold the current parameters for voice B, its structure is the same as VCBA. FB8BH VCBC : DEFS 37

    This buffer is used by the "PLAY" statement handler to hold the current parameters for voice C, its structure is the same as VCBA. FBB0H ENSTOP: DEFB 00H

    This variable determines whether the interrupt handler will execute a warm start to the Interpreter upon detecting the keys CODE, GRAPH, CTRL and SHIFT depressed together: 00H=Disable, NZ=Enable. FBB1H BASROM: DEFB 00H

    This variable determines whether the ISCNTC and INLIN standard routines will respond to the CTRL-STOP key: 00H=Enable, NZ=Disable. It is used to prevent termination of a BASIC ROM located during the power-up ROM search. FBB2H LINTTB: DEFS 24

    Each of these twenty-four variables is normally non-zero but is zeroed if the contents of the corresponding screen row have overflowed onto the next row. They are maintained by the BIOS but only actually used by the INLIN standard routine (the screen editor) to discriminate between logical and physical lines. FBCAH FSTPOS: DEFW 0000H

    This variable is used to hold the cursor coordinates upon entry to the INLIN standard routine. Its function is to restrict the extent of backtracking performed when the text is collected from the screen at termination. FBCCH CURSAV: DEFB 00H

    This variable is used to hold the screen character replaced by the text cursor. FBCDH FNKSWI: DEFB 00H

    This variable is used by the CHSNS standard routine to determine whether the shifted or unshifted function keys are currently displayed: 00H=Shifted, 01H=Unshifted. FBCEH FNKFLG: DEFS 10

    Each of these ten variables is normally zero but is set to 01H if the associated function key has been turned on by a "KEY(n) ON" statement. They are used by the interrupt handler to determine whether, in program mode only, it should return a character string or update the associated entry in TRPTBL. FBD8H ONGSBF: DEFB 00H

    This variable is normally zero but is incremented by the interrupt handler whenever a device has achieved the conditions necessary to generate a program interrupt. It is used by the Runloop to determine whether any program interrupts are pending without having to search TRPTBL. FBD9H CLIKFL: DEFB 00H

    This variable is used internally by the interrupt handler to prevent spurious key clicks when returning multiple characters from a single key depression such as a function key. FBDAH OLDKEY: DEFS 11

    This buffer is used by the interrupt handler to hold the previous state of the keyboard matrix, each byte contains one row of keys starting with row 0. FBE5H NEWKEY: DEFS 11

    This buffer is used by the interrupt handler to hold the current state of the keyboard matrix. Key transitions are detected by comparison with the contents of OLDKEY after which OLDKEY is updated with the current state. FBF0H KEYBUF: DEFS 40

    This buffer contains the decoded keyboard characters produced by the interrupt handler. Note that the buffer is organized as a circular queue driven by GETPNT and PUTPNT and consequently has no fixed starting point. FC18H LINWRK: DEFS 40

    This buffer is used by the BIOS to hold a complete line of screen characters. FC40H PATWRK: DEFS 8

    This buffer is used by the BIOS to hold an 8x8 pixel pattern. FC48H BOTTOM: DEFW 8000H

    This variable contains the address of the lowest RAM location used by the Interpreter. Its value is set at power-up and thereafter unaltered. FC4AH HIMEM : DEFW F380H

    This variable contains the address of the byte following the highest RAM location used by the Interpreter. Its value is set at power-up and thereafter only altered by the "CLEAR" statement.

    FC4CH TRPTBL: DEFS 3       ; KEY 1
    FC4FH         DEFS 3       ; KEY 2
    FC52H         DEFS 3       ; KEY 3
    FC55H         DEFS 3       ; KEY 4
    FC58H         DEFS 3       ; KEY 5
    FC5BH         DEFS 3       ; KEY 6
    FC5EH         DEFS 3       ; KEY 7
    FC61H         DEFS 3       ; KEY 8
    FC64H         DEFS 3       ; KEY 9
    FC67H         DEFS 3       ; KEY 10
    FC6AH         DEFS 3       ; STOP
    FC6DH         DEFS 3       ; SPRITE
    FC70H         DEFS 3       ; STRIG 0
    FC73H         DEFS 3       ; STRIG 1
    FC76H         DEFS 3       ; STRIG 2
    FC79H         DEFS 3       ; STRIG 3
    FC7CH         DEFS 3       ; STRIG 4
    FC7FH         DEFS 3       ; INTERVAL
    FC82H         DEFS 3       ; Unused
    FC85H         DEFS 3       ; Unused
    FC88H         DEFS 3       ; Unused
    FC8BH         DEFS 3       ; Unused
    FC8EH         DEFS 3       ; Unused
    FC91H         DEFS 3       ; Unused
    FC94H         DEFS 3       ; Unused
    FC97H         DEFS 3       ; Unused

    These twenty-six three byte variables hold the current state of the interrupt generating devices. The first byte of each entry contains the device status (bit 0=On, bit 1=Stop, bit 2=Event active) and is updated by the interrupt handler, the Runloop interrupt processor and the "DEVICE 0=ON/OFF/STOP" and "RETURN" statement handlers. The remaining two bytes of each entry are set by the "ON DEVICE GOSUB" statement handler and contain the address of the program line to execute upon a program interrupt. FC9AH RTYCNT: DEFB 00H

    This variable is unused by the current MSX ROM. FC9BH INTFLG: DEFB 00H

    This variable is normally zero but is set to 03H or 04H if the CTRL-STOP or STOP keys are detected by the interrupt handler. FC9CH PADY : DEFB 00H

    This variable contains the Y coordinate of the last point detected by a touchpad. FC9DH PADX : DEFB 00H

    This variable contains the X coordinate of the last point detected by a touchpad. FC9EH JIFFY : DEFW 0000H

    This variable is continually incremented by the interrupt handler. Its value may be set or read by the "TIME" statement or function. FCA0H INTVAL: DEFW 0000H

    This variable holds the interval duration set by the "ON INTERVAL" statement handler. FCA2H INTCNT: DEFW 0000H

    This variable is continually decremented by the interrupt handler. When zero is reached its value is reset from INTVAL and, if applicable, a program interrupt generated. Note that this variable always counts irrespective of whether an "INTERVAL ON" statement is active. FCA4H LOWLIM: DEFB 31H

    This variable is used to hold the minimum allowable start bit duration as determined by the TAPION standard routine. FCA5H WINWID: DEFB 22H

    This variable is used to hold the LO/HI cycle discrimination duration as determined by the TAPION standard routine. - 233 - 6. MEMORY MAP FCA6H GRPHED: DEFB 00H

    This variable is normally zero but is set to 01H by the CNVCHR standard routine upon detection of a graphic header code. FCA7H ESCCNT: DEFB 00H

    This variable is used by the CHPUT standard routine ESC sequence processor to count escape parameters. FCA8H INSFLG: DEFB 00H

    This variable is normally zero but is set to FFH by the INLIN standard routine when insert mode is on. FCA9H CSRSW : DEFB 00H

    If this variable is zero the cursor is only displayed while the CHGET standard routine is waiting for a keyboard character. If it is non-zero the cursor is permanently displayed via the CHPUT standard routine. FCAAH CSTYLE: DEFB 00H

    This variable determines the cursor style: 00H=Block, NZ=Underline. FCABH CAPST : DEFB 00H

    This variable is used by the interrupt handler to hold the current caps lock status: 00H=Off, NZ=On. FCACH KANAST: DEFB 00H

    This variable is used to hold the keyboard Kana lock status on Japanese machines and the DEAD key status on European machines. FCADH KANAMD: DEFB 00H

    This variable holds a keyboard mode on Japanese machines only. FCAEH FLBMEM: DEFB 00H

    This variable is set by the file I/O error generators but is otherwise unused. FCAFH SCRMOD: DEFB 00H

    This variable contains the current screen mode: 0=40x24 Text Mode, 1=32x24 Text Mode, 2=Graphics Mode, 3=Multicolour Mode. FCB0H OLDSCR: DEFB 00H

    This variable holds the screen mode of the last text mode set. - 234 - 6. MEMORY MAP FCB1H CASPRV: DEFB 00H

    This variable is used to hold any character returned to an I/O buffer by the cassette putback function. FCB2H BDRATR: DEFB 00H

    This variable contains the boundary colour for the "PAINT" statement handler. Its value is set by the PNTINI standard routine and used by the SCANR and SCANL standard routines. FCB3H GXPOS : DEFW 0000H

    This variable is used for temporary storage of a graphics X coordinate. FCB5H GYPOS : DEFW 0000H

    This variable is used for temporary storage of a graphics Y coordinate. FCB7H GRPACX: DEFW 0000H

    This variable contains the current graphics X coordinate for the GRPPRT standard routine. FCB9H GRPACY: DEFW 0000H

    This variable contains the current graphics Y coordinate for the GRPPRT standard routine. FCBBH DRWFLG: DEFB 00H

    Bits 6 and 7 of this variable are set by the "DRAW" statement "N" and "B" command handlers to turn the associated mode on. FCBCH DRWSCL: DEFB 00H

    This variable is used by the "DRAW" statement "S" command handler to hold the current scale factor. FCBDH DRWANG: DEFB 00H

    This variable is used by the "DRAW" statement "A" command handler to hold the current angle. FCBEH RUNBNF: DEFB 00H

    This variable is normally zero but is set by the "BLOAD" statement handler when an auto-run "R" parameter is specified. FCBFH SAVENT: DEFW 0000H

    This variable contains the "BSAVE" and "BLOAD" entry address. - 235 - 6. MEMORY MAP FCC1H EXPTBL: DEFB 00H ; Primary Slot 0 FCC2H DEFB 00H ; Primary Slot 1 FCC3H DEFB 00H ; Primary Slot 2 FCC4H DEFB 00H ; Primary Slot 3

    Each of these four variables is normally zero but is set to 80H during the power-up RAM search if the associated Primary Slot is found to be expanded. FCC5H SLTTBL: DEFB 00H ; Primary Slot 0 FCC6H DEFB 00H ; Primary Slot 1 FCC7H DEFB 00H ; Primary Slot 2 FCC8H DEFB 00H ; Primary Slot 3

    These four variables duplicate the contents of the four possible Secondary Slot Registers. The contents of each variable should only be regarded as valid if EXPTBL shows the associated Primary Slot to be expanded. FCC9H SLTATR: DEFS 4 ; PS0, SS0 FCCDH DEFS 4 ; PS0, SS1 FCD1H DEFS 4 ; PS0, SS2 FCD5H DEFS 4 ; PS0, SS3 FCD9H DEFS 4 ; PS1, SS0 FCDDH DEFS 4 ; PS1, SS1 FCE1H DEFS 4 ; PS1, SS2 FCE5H DEFS 4 ; PS1, SS3 FCE9H DEFS 4 ; PS2, SS0 FCEDH DEFS 4 ; PS2, SS1 FCF1H DEFS 4 ; PS2, SS2 FCF5H DEFS 4 ; PS2, SS3 FCF9H DEFS 4 ; PS3, SS0 FCFDH DEFS 4 ; PS3, SS1 FD01H DEFS 4 ; PS3, SS2 FD05H DEFS 4 ; PS3, SS3

    These sixty-four variables contain the attributes of any extension ROMs found during the power-up ROM search. The characteristics of each 16 KB ROM are encoded into a single byte so four bytes are required for each possible slot. The encoding is:

        Bit 7 set=BASIC program
        Bit 6 set=Device handler
        Bit 5 set=Statement handler

    Note that the entries for page 0 (0000H to 3FFFH) and page 3 (C000H to FFFFH) will always be zero as only page 1 (4000H to 7FFFH) and page 2 (8000H to BFFFH) are actually examined. The MSX convention is that machine code extension ROMs are placed in page 1 and BASIC program ROMs in page 2. FD09H SLTWRK: DEFS 128

    This buffer provides two bytes of local workspace for each of the sixty-four possible extension ROMs. FD89H PROCNM: DEFS 16

    This buffer is used to hold a device or statement name for examination by an extension ROM. FD99H DEVICE: DEFB 00H

    This variable is used to pass a device code, from 0 to 3, to an extension ROM.

The Hooks

RST 30H
DEFB Slot ID
DEFW Address
RET


    The hooks are listed on the following pages together with the address they are called from and a brief note as to their function.

    FD9AH HKEYI : DEFS 5       ; 0C4AH Interrupt handler
    FD9FH HTIMI : DEFS 5       ; 0C53H Interrupt handler
    FDA4H HCHPU : DEFS 5       ; 08C0H CHPUT standard routine
    FDA9H HDSPC : DEFS 5       ; 09E6H Display cursor
    FDAEH HERAC : DEFS 5       ; 0A33H Erase cursor
    FDB3H HDSPF : DEFS 5       ; 0B2BH DSPFNK standard routine
    FDB8H HERAF : DEFS 5       ; 0B15H ERAFNK standard routine
    FDBDH HTOTE : DEFS 5       ; 0842H TOTEXT standard routine
    FDC2H HCHGE : DEFS 5       ; 10CEH CHGET standard routine
    FDC7H HINIP : DEFS 5       ; 071EH Copy character set to VDP
    FDCCH HKEYC : DEFS 5       ; 1025H Keyboard decoder
    FDD1H HKYEA : DEFS 5       ; 0F10H Keyboard decoder
    FDD6H HNMI  : DEFS 5       ; 1398H NMI standard routine
    FDDBH HPINL : DEFS 5       ; 23BFH PINLIN standard routine
    FDE0H HQINL : DEFS 5       ; 23CCH QINLIN standard routine
    FDE5H HINLI : DEFS 5       ; 23D5H INLIN standard routine
    FDEAH HONGO : DEFS 5       ; 7810H "ON DEVICE GOSUB"
    FDEFH HDSKO : DEFS 5       ; 7C16H "DSKO$"
    FDF4H HSETS : DEFS 5       ; 7C1BH "SET"
    FDF9H HNAME : DEFS 5       ; 7C20H "NAME"
    FDFEH HKILL : DEFS 5       ; 7C25H "KILL"
    FE03H HIPL  : DEFS 5       ; 7C2AH "IPL"
    FE08H HCOPY : DEFS 5       ; 7C2FH "COPY"
    FE0DH HCMD  : DEFS 5       ; 7C34H "CMD"
    FE12H HDSKF : DEFS 5       ; 7C39H "DSKF"
    FE17H HDSKI : DEFS 5       ; 7C3EH "DSKI$"
    FE1CH HATTR : DEFS 5       ; 7C43H "ATTR$"
    FE21H HLSET : DEFS 5       ; 7C48H "LSET"
    FE26H HRSET : DEFS 5       ; 7C4DH "RSET"
    FE2BH HFIEL : DEFS 5       ; 7C52H "FIELD"
    FE30H HMKI$ : DEFS 5       ; 7C57H "MKI$"
    FE35H HMKS$ : DEFS 5       ; 7C5CH "MKS$"
    FE3AH HMKD$ : DEFS 5       ; 7C61H "MKD$"
    FE3FH HCVI  : DEFS 5       ; 7C66H "CVI"
    FE44H HCVS  : DEFS 5       ; 7C6BH "CVS"
    FE49H HCVD  : DEFS 5       ; 7C70H "CVD"
    FE4EH HGETP : DEFS 5       ; 6A93H Locate FCB
    FE53H HSETF : DEFS 5       ; 6AB3H Locate FCB
    FE58H HNOFO : DEFS 5       ; 6AF6H "OPEN"
    FE5DH HNULO : DEFS 5       ; 6B0FH "OPEN"
    FE62H HNTFL : DEFS 5       ; 6B3BH Close I/O buffer 0
    FE67H HMERG : DEFS 5       ; 6B63H "MERGE/LOAD"
    FE6CH HSAVE : DEFS 5       ; 6BA6H "SAVE"
    FE71H HBINS : DEFS 5       ; 6BCEH "SAVE"
    FE76H HBINL : DEFS 5       ; 6BD4H "MERGE/LOAD"
    FE7BH HFILE : DEFS 5       ; 6C2FH "FILES"
    FE80H HDGET : DEFS 5       ; 6C3BH "GET/PUT"
    FE85H HFILO : DEFS 5       ; 6C51H Sequential output
    FE8AH HINDS : DEFS 5       ; 6C79H Sequential input
    FE8FH HRSLF : DEFS 5       ; 6CD8H "INPUT$"
    FE94H HSAVD : DEFS 5       ; 6D03H "LOC", 6D14H "LOF",
                               ; 6D25H "EOF", 6D39H "FPOS"
    FE99H HLOC  : DEFS 5       ; 6D0FH "LOC"
    FE9EH HLOF  : DEFS 5       ; 6D20H "LOF"
    FEA3H HEOF  : DEFS 5       ; 6D33H "EOF"
    FEA8H HFPOS : DEFS 5       ; 6D43H "FPOS"
    FEADH HBAKU : DEFS 5       ; 6E36H "LINE INPUT#"
    FEB2H HPARD : DEFS 5       ; 6F15H Parse device name
    FEB7H HNODE : DEFS 5       ; 6F33H Parse device name
    FEBCH HPOSD : DEFS 5       ; 6F37H Parse device name
    FEC1H HDEVN : DEFS 5       ; This hook is not used.
    FEC6H HGEND : DEFS 5       ; 6F8FH I/O function dispatcher
    FECBH HRUNC : DEFS 5       ; 629AH Run-clear
    FED0H HCLEA : DEFS 5       ; 62A1H Run-clear
    FED5H HLOPD : DEFS 5       ; 62AFH Run-clear
    FEDAH HSTKE : DEFS 5       ; 62F0H Reset stack
    FEDFH HISFL : DEFS 5       ; 145FH ISFLIO standard routine
    FEE4H HOUTD : DEFS 5       ; 1B46H OUTDO standard routine
    FEE9H HCRDO : DEFS 5       ; 7328H CR,LF to OUTDO
    FEEEH HDSKC : DEFS 5       ; 7374H Mainloop line input
    FEF3H HDOGR : DEFS 5       ; 593CH Line draw
    FEF8H HPRGE : DEFS 5       ; 4039H Program end
    FEFDH HERRP : DEFS 5       ; 40DCH Error handler
    FF02H HERRF : DEFS 5       ; 40FDH Error handler
    FF07H HREAD : DEFS 5       ; 4128H Mainloop "OK"
    FF0CH HMAIN : DEFS 5       ; 4134H Mainloop
    FF11H HDIRD : DEFS 5       ; 41A8H Mainloop direct statement
    FF16H HFINI : DEFS 5       ; 4237H Mainloop finished
    FF1BH HFINE : DEFS 5       ; 4247H Mainloop finished
    FF20H HCRUN : DEFS 5       ; 42B9H Tokenize
    FF25H HCRUS : DEFS 5       ; 4353H Tokenize
    FF2AH HISRE : DEFS 5       ; 437CH Tokenize
    FF2FH HNTFN : DEFS 5       ; 43A4H Tokenize
    FF34H HNOTR : DEFS 5       ; 44EBH Tokenize
    FF39H HSNGF : DEFS 5       ; 45D1H "FOR"
    FF3EH HNEWS : DEFS 5       ; 4601H NEWSTT, Runloop new statement
    FF43H HGONE : DEFS 5       ; 4646H Runloop execute
    FF48H HCHRG : DEFS 5       ; 4666H CHRGTR standard routine
    FF4DH HRETU : DEFS 5       ; 4821H "RETURN"
    FF52H HPRTF : DEFS 5       ; 4A5EH "PRINT"
    FF57H HCOMP : DEFS 5       ; 4A54H "PRINT"
    FF5CH HFINP : DEFS 5       ; 4AFFH "PRINT"
    FF61H HTRMN : DEFS 5       ; 4B4DH "READ/INPUT" error
    FF66H HFRME : DEFS 5       ; 4C6DH Expression Evaluator
    FF6BH HNTPL : DEFS 5       ; 4CA6H Expression Evaluator
    FF70H HEVAL : DEFS 5       ; 4DD9H Factor Evaluator
    FF75H HOKNO : DEFS 5       ; 4F2CH Factor Evaluator
    FF7AH HFING : DEFS 5       ; 4F3EH Factor Evaluator
    FF7FH HISMI : DEFS 5       ; 51C3H Runloop execute
    FF84H HWIDT : DEFS 5       ; 51CCH "WIDTH"
    FF89H HLIST : DEFS 5       ; 522EH "LIST"
    FF8EH HBUFL : DEFS 5       ; 532DH Detokenize
    FF93H HFRQI : DEFS 5       ; 543FH Convert to integer
    FF98H HSCNE : DEFS 5       ; 5514H Line number to pointer
    FF9DH HFRET : DEFS 5       ; 67EEH Free descriptor
    FFA2H HPTRG : DEFS 5       ; 5EA9H Variable search
    FFA7H HPHYD : DEFS 5       ; 148AH PHYDIO standard routine
    FFACH HFORM : DEFS 5       ; 148EH FORMAT standard routine
    FFB1H HERRO : DEFS 5       ; 406FH Error handler
    FFB6H HLPTO : DEFS 5       ; 085DH LPTOUT standard routine
    FFBBH HLPTS : DEFS 5       ; 0884H LPTSTT standard routine
    FFC0H HSCRE : DEFS 5       ; 79CCH "SCREEN"
    FFC5H HPLAY : DEFS 5       ; 73E5H "PLAY" statement

    The Workspace Area from FFCAH to FFFEH is unused in MSX, but in MSX2, MSX2+ and turbo R. Here forth are texts containing data not presented in the original version of this book. FFFFH ScSltR: DEFB 00H ; Secondary Slot Register

    This address is directly controled by slot expansion devices, if any is active, otherwise it has no function. For more information, see Section 1 (Programmable Peripheral Interface).

10. APPENDICES

A0. INTRODUCTION

    The following appendix texts are a compilation from several sources, including "MSX Datapack" by Japan ASCII Co.'s and "Japan MSX Magazine". The contents of each appendix are:

    A0. This presentation ..........................  240
    A1. MSX2 and MSX2+ VDP .........................  ???
    A2. MSX BIOS Complement Data ...................  ???
    A3. MSX2, MSX2+ and turbo R BIOS ...............  ???
    A4. MSX2, MSX2+ and turbo R Sub-BIOS ...........  ???
    A5. MSX2, MSX2+ and turbo R Word Area ..........  ???
    A6. MSX2, MSX2+ and turbo R I/O Port Map .......  ???

A1. MSX2 and MSX2+ VDP

A2. MSX BIOS Complementary Data

    This section presents several data that has been simply ignored by the original version of this book. Many of them have already been included in the original text, mainly label names and entry and exit data, but it was impossible to include everything without changing all the page numbers.

    The topics covered by this section are:

• MSX ROM Version Identification Bytes;
• MSX Math-Pack.

7.A4. MSX2, MSX2+ and turbo R Sub-BIOS

7.A5. Work Area

    The Work Area of MSX2, MSX2+ and turbo R are very similar of MSX's one. The only differences are in the RS2IQ area (from FAF5H to FB34H), which has been divided in two, from FAF5H to FB02 for new MSX features and from FB03 to FB34H for RS-232C and some system variables, and the last bytes of the RAM, after the hooks (from FFCAH to FFFEH).

    The MSX-Disk work area has the following entries: F323 DISKVE: DEFS 2 F325 BREAKV: DEFS 2 F341 RAMAD0: DEFS 1 F342 RAMAD1: DEFS 1 F343 RAMAD2: DEFS 1 F344 RAMAD3: DEFS 1 F348 MASTERS: DEFS 1 F37D BDOS: DEFS 2 ; MSX-Disk BIOS Function Call. F24F HPROMPT: DEFS 5 FAF5 DPPAGE: DEFB 0 ; Current Display Page.

    The number of the VDP page currently displayed (MSX2, MSX2+, turbo R). FAF6 ACPAGE: DEFB 0 ; Active Access Page.

    The number of the VDP page actually accessed (MSX2, MSX2+, turbo R). FAF7 AVCSAV: DEFS 1 ; AV Control Board Save state.

    Used by MSX AV Control boards (MSX, MSX2, MSX2+, turbo R). FAF8 EXBRSA: DEFS 1 ; Extended BIOS Sub-ROM slot Id. Save.

    Slot identification of the extended BIOS Sub-ROM (MSX2, MSX2+, turbo R). FAF9 CHRCNT: DEFS 1 ; Character Counter.

    "Romaji" to "kana" character conversion counter. May range from 0 to 2. MSX2, MSX2+ and turbo R. FAFA ROMA : DEFS 2 ; "Romaji" to "kana" character buffer.

    This buffer contains the characters to a "romaji" to "kana" conversion. MSX2, MSX2+ and turbo R. FAFC MODE : DEFS 1 ; "Romaji" to "kana" conversion Mode.

    MSX2, MSX2+ and turbo R. Each bit has a special meaning:

      7    6    5    4    3    2    1    0
    +----+----+----+----+----+---------+----+
    |Kata|D2SJ|RGB |Clip|Mask|VRAM Size|Conv|
    +----+----+----+----+----+---------+----+

    Conv     : 0 = don't convert "romaji", 1 = convert.
    VRAM Size: 00 = 16 KBytes, 01 = 64 KBytes, 10 = 128 KBytes.
    Mask     : 0 = don't mask, 1 = mask (VRAM address of SCREEN 0 to 3).
    Clip     : 0 = clipping active, 1 = don't clip.
    RGB      : 0 = SCREEN 10, 1 = SCREEN 11 (RGB flag of MSX2+).
    D2SJ     : Dai 2 Suijyun Kanji ROM, 0 = not available, 1 = available.
    Kata     : 0 = "hiragana" ("romaji" conversion), 1 = "katakana".

    MSX2, MSX2+ and turbo R. FAFE XSAVE : DEFS 2 ; Light-Pen X Save. FB00 YSAVE : DEFS 2 ; Light-Pen Y Save.

    MSX2, MSX2+ and turbo R. This variables have the light-pen coordinates:

                  XSAVE

     F E D C B A 9 8 7 6 5 4 3 2 1 0
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |I|O|O|O|O|O|O|O|X|X|X|X|X|X|X|X|
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                  YSAVE

     F E D C B A 9 8 7 6 5 4 3 2 1 0
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |x|O|O|O|O|O|O|O|Y|Y|Y|Y|Y|Y|Y|Y|
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

    I: 1 = light-pen interruption.
    O: offset.
    X: X coordinate.
    Y: Y coordinate.
    x: anything.

    This variable contains the logical operation to be used by VDP (MSX2, MSX2+, turbo R) in commands like COPY and LINE.

    The following are the RS-232C work area.

    FB03  RSTMP : DEFS 50      ; RS-232C and MSX-Disk Temporary work area.
    FB03  TOCNT : DEFS 1       ; Used by RS-232C access routines.
    FB04  RSFCB : DEFS 1       ; RS-232C low address.
                  DEFS 1       ; RS-232C high address.
    FB06  RSIQLN: DEFS 1       ; Used by RS-232C access routines.
    FB07  MEXBIH: RST  30H
                  DEFS 1       ; Byte datum.
                  DEFS 1       ; (Low).
                  DEFS 1       ; (High).
                  RET
    FB0C  OLDSTT: RST  30H
                  DEFS 1       ; Byte datum.
                  DEFS 1       ; (Low).
                  DEFS 1       ; (High).
                  RET
    FB11  OLDINT: RST  30H
                  DEFS 1       ; Byte datum.
                  DEFS 1       ; (Low).
                  DEFS 1       ; (High).
                  RET
    FB16  DEVNUM: DEFS 1       ; Used by RS-232C access routines.
    FB17  DATCNT: DEFS 1       ; Data Counter (byte datum).
                  DEFS 1       ; Byte pointer.
                  DEFS 1       ; Byte pointer.

    FB1A  ERRORS: DEFS 1
    FB1B  FLAGS : DEFS 1
    FB1C  ESTBLS: DEFS 1
    FB1D  COMMSK: DEFS 1
    FB1E  LSTCOM: DEFS 1
    FB1F  LSTMOD: DEFS 1

    These variables are used by the RS-232C access routines.

    The following are some new system variables. FB20 HOKVLD: DEFS 1 ; Used by the Extended BIOS ROM. FB21 DRVTBL: DEFS 8 ; Drive Table.

    This variable contains the MSX-Disk ROM slot identification.

    FFCAH FCALL : DEFS 5       ; Used by BIOS extensions.
    FFCFH DISINT: DEFS 5       ; Used by DOS.
    FFD4H ENAINT: DEFS 5       ; Used by DOS.

    FFE7H RG8SAV: DEFS 1       ; VDP Register 8 Save copy.
    FFE8H RG9SAV: DEFS 1       ; VDP Register 9 Save copy.
    FFE9H RG10SA: DEFS 1       ; VDP Register 10 Save copy.
    FFEAH RG11SA: DEFS 1       ; VDP Register 11 Save copy.
    FFEBH RG12SA: DEFS 1       ; VDP Register 12 Save copy.
    FFECH RG13SA: DEFS 1       ; VDP Register 13 Save copy.
    FFEDH RG14SA: DEFS 1       ; VDP Register 14 Save copy.
    FFEEH RG15SA: DEFS 1       ; VDP Register 15 Save copy.
    FFEFH RG16SA: DEFS 1       ; VDP Register 16 Save copy.
    FFF0H RG17SA: DEFS 1       ; VDP Register 17 Save copy.
    FFF1H RG18SA: DEFS 1       ; VDP Register 18 Save copy.
    FFF2H RG19SA: DEFS 1       ; VDP Register 19 Save copy.
    FFF3H RG20SA: DEFS 1       ; VDP Register 20 Save copy.
    FFF4H RG21SA: DEFS 1       ; VDP Register 21 Save copy.
    FFF5H RG22SA: DEFS 1       ; VDP Register 22 Save copy.
    FFF6H RG23SA: DEFS 1       ; VDP Register 23 Save copy.

    FFF7H         DEFS 3       ; Reserved for the system.

    FFFAH RG25SA: DEFS 1       ; VDP Register 25 Save copy.
    FFFBH RG26SA: DEFS 1       ; VDP Register 26 Save copy.
    FFFCH RG27SA: DEFS 1       ; VDP Register 27 Save copy.

    FFFDH         DEFS 2       ; Reserved for the system.

A6. MSX2, MSX2+ and turbo R I/O Port Map

Port Range    Device and Optional Description

00H-3FH       User Definition ***.

40H-4FH       Expansion I/O Ports ****.

50H-6FH       Reserved for the system *.

70H-73H       MIDI Saurus (BIT2).

74H-7BH       Reserved for the system *.

7CH-7DH       MSX-Music.

7EH-7FH       Reserved for the system *.

80H-87H       RS-232C.
        80H   8251 Data.
        81H   8251 Status - Command.
        82H   Status Lead - Interrupt Mask.
        83H   Unused.
        84H   8253.
        85H   8253.
        86H   8253.
        87H   8253.

88H-8BH       For the MSX2 VDP (V9938) Adapter I/O ports **.

8CH-8DH       MSX Modem.

8EH-8FH       Reserved for the system *.

90H-91H       Printer ports.
        90H   Bit 0 on writing: /STROBE, 0=sending datum.
              Bit 1 on reading: 0=Ready, 1=Busy.
        91H   Data port.

92H-97H       Reserved for the system *.

98H-9BH       VDP (V9938) of MSX2. Also most MSX and all MSX2+ and turbo R.
        98H   VRAM access.
        99H   Command register access.
        9AH   Palette register access (write only).
        9BH   Register indirect assignment (write only).

9CH-9FH       Reserved for the system *.

A0H-A3H       PSG (AY-3-8910) access ports.
        A0H   Address latch.
        A1H   Data Write.
        A2H   Data Read.

A4H-A7H       Reserved for the system *.

A8H-ABH       Parallel port (PPI, 8255).
        A8H   Port A.
        A9H   Port B.
        AAH   Port C.
        ABH   Mode Set.

ACH-AFH       MSX Engine (One-Chip-MSX I/O Ports).

B0H-B3H       Memory Expansion (Sony method, using 8255).
        B0H   Port A, address (A0-A7).
        B1H   Port B, address (A8-A10, A13-A15), control, Read/Write.
        B2H   Port C, address (A11-A12), datum (D0-D7).
        B3H   Mode Set.

B4H-B5H       Clock-IC (RP-5C01).
        B4H   Address latch.
        B5H   Datum.

B6H-B7H       Reserved for the system *.

B8H-BBH       Light-pen control (Sanyo method).
        B8H   Read/Write.
        B9H   Read/Write.
        BAH   Read/Write.
        BBH   Write only.

BCH-BFH       VHD control (JVC, 8255).
        BCH   Port A.
        BDH   Port B.
        BEH   Port C.
        BFH   Mode Set.

C0H-C1H       MSX-Audio.

C2H-C7H       Reserved for the system *.

C8H-CFH       MSX Interface (Asynchronous Serial Communication Interface).

D0H-D7H       Floppy Disk Controller (FDC). According to the Datapack, the
                use of these ports to access disks will limit its access
                time. Translator notes: because the I/O ports have a fix
                transfer rate to keep compatibility, newer and faster data
                storage devices won't be able to work properly using this
                access. Standard FDC use memory registers, which can also
                be accessed faster than by I/O ports. The use of these
                ports also brings some access conflict problems. Only some
                brazilian interfaces are known to use them.

D8H-D9H       Kanji ROM (Toshiba method) Dai Itchi Suijyun (Level 1).
        D8H   b5-b0 lower address (Write Only).
        D9H   b5-b0 upper address (Write).
              b7-b0 datum (Read).

DAH-DBH       Kanji ROM Dai Ni Suijyun (Level 2).

DCH-F2H       Reserved for the system *.

F3H           VDP Display Mode (MSX2+).
        b0    M3
        b1    M4
        b2    M5
        b3    M2
        b4    M1
        b5    TP
        b6    YUV
        b7    YAE

F4H     MSB   Hardware Reset (MSX2+).

F5H           System Control (Write Only). Bit=1 means device available.
        b0    Kanji ROM Dai Itchi Suijyun.
        b1    Kanji ROM Dai Ni Suijyun.
        b2    MSX-Audio.
        b3    Superimposer.
        b4    MSX Interface.
        b5    RS-232C.
        b6    Light-pen.
        b7    Clock-IC (only in MSX2, MSX2+).

F6H           Colour Bus I/O.

F7H           A/V Control.
        b0    Audio R                         Mixing On    (Write)
        b1    Audio L                         Mixing Off   (Write)
        b2    Video Input Selection           21 Pin RGB   (Write)
        b3    Video Input Detection           No Input     (Read)
        b4    AV Control                      TV           (Write)
        b5    Ym Control                      TV           (Write)
        b6    VDP Register 9, Bit 4 Shift                  (Write)
        b7    VDP Register 9, Bit 5 Shift                  (Write)

F8H-FBH       Reserved for the system *.

FCH-FFH       Memory Mapper (Write Only). Translator note:
        FCH   Page 3 Segment Select Register.
        FDH   Page 2 Segment Select Register.
        FEH   Page 1 Segment Select Register.
        FFH   Page 0 Segment Select Register.

    Notes:

    ** The new VDP is connected to different ports of MSX, avoiding conflict with the current built-in VDP. one may have it connected to alternative ones.

    * Reserved for the system: this basically means that a hardware maker should consult the standard definer (Japan ASCII Co.) for permission to use some port. Of course, most hardware makers today simply ignores the simple existance of any standard definer and goes on filling up the not many left unused ports with not very much reliable devices. Anyway, it is a good idea to check twice the possibilities of memory access registers instead of I/O ports.

    ** VDP (V9938) Adapter I/O Ports: when MSX2 had just been released, the number of MSX already in use was large, so a special cartridge had been released with the MSX2 VDP IC inside, the V9938. That cartridge could be simply inserted in an MSX slot and it would become an MSX2. Those expansion cartridges contain a V9938 and the MSX2 BIOS. The video access ports, of course, couldn't be shared by both internal MSX's and external MSX2's VDP, so new ports were defined to the second. Thus, a program shouldn't assume VDP as being at fixed ports. The addresses 0006H and 0007H of the Main BIOS contain the read and write port numbers of the VDP, as already explained in the VDP chapter, and they should be used.

    *** User Definition: these ports are for home-made devices, for personal use only, that is, hardware that won't be commercialized or distributed in any way. This is to avoid the proliferation of hardware with unauthorized usage of ports, but still allowing simple home-made hardware to be easily done.

    **** Expansion I/O Ports: if you must or want anyway to use I/O ports in your project, Japan ASCII has defined the Expansion I/O Ports, from 40H to 4FH. A special selector must be connected to these ports in the new device. The selector only allows access to the device if the correct signature is sent to it.

    The schematic circuitry diagram is presented here, not specifying componets, just logical elements and functions:

      CPU                                                         Device
  --------+                                                     +--------
          |                                                     |
   A0~A15 |=====+==============================================>| A0~A15
          |     |                                               |
          |     |                                               |
          |     |                           +--------+          |
          |     |                           | Data   |          |
    D0~D7 |<====|===========+==============>| Bus    |<========>| D0~D7
          |     |           |               | Buffer |    ^     |
          |     V           |               +--------+    |     +--------
          |  +------+       V                  ^  ^       |
          |  |xx40H?|     +---+                |  |       |
          |  +------+     |ID?|  Enable F/F    |  |       |
          |     |         +---+                |  |e    +--------+
          |     |   +---+   |      +---+       |  |n    |        |
          |     |   |   |   |      |   |       |  |a    | Buffer |.....
          |     +-->|   |   ++---->|D Q|-------+  |b    |        |    .
          |     |   |   |          |   | enable   |l    +--------+    .
     IORQ |O--+-|--O|AND|--------->|T  |          |e       ^          .
          |   | |   |   |  write   | R |          |        |          .
       WR |O--|-|--O|   |          +---+          |        |    complement
          |   | |   |   |            O            |        |      of ID
          |   | |   +---+            |       +---------+   |
          |   | |                    |       | Access  |   |
          |   | |   +---+            |       | Control |   |
          |   | |   |   |            |       +---------+   |
          |   | +-->|   |            |                     |
          |   |     |   |  read      |                     |
          |   +----O|AND|----------------------------------+
          |         |   |            |
       RD |O-------O|   |            |
          |         |   |            |
          |         +---+            |
          |                          |
    RESET |O-------------------------+
          |
  --------+

    This circuitry has the following effects:

•     It keeps the device and the MSX Bus disconnected until it receives a valid ID code through port 40H, so a program that wants to use the device must enable it first, writing the proper ID byte to the I/O port 40H. The ID byte may range from 128 to 254 for a specific type of device or from 1 to 127 to define a specific hardware maker. The 0 and 255 values are reserved because most systems returns one of those two when an empty or disconnected port is accessed.

•     If an ID byte different from the expected is received, connection between bus and device is instantly broken. Any and all access are ignored until another matching ID value is received from port 40H.

•     If port 40H is read, the accessing program will receive the ID byte of the currently active device.

MSX Maker List

    MSX standard has not defined any device that uses the expansion I/O ports, so no table of devices exist, but each hardware maker has its ID byte (most names are written in EUC [Enhanced Unix Code] Japanese characters, but don't worry if you cannot view them or simply cannot understand them, the reading and translation follows):

        1 = アスキー (Japan ASCII)
        2 = キャノン (Canon)
        3 = カシオ計算機 (Casio Keisanki = Casio Calculators)
        4 = 富士通 (Fujitsu)
        5 = 富士通ゼネラル (Fujitsu General)
        6 = 日立製作所 (Hitatchi Seisakushyo = Hitachi Industrial Plant)
        7 = 京セラ (Kyosera)
        8 = 松下電器産業 (Matsushita Denkisangyou = Electric Industrial [this includes all Matsushita enterprises, Panasonic, National, Technics and Quasar])
        9 = 三菱電機 (Mitsubishi Denki = Electrical Machines)
       10 = 日本電気 (Nihon Denki = Japan Electric)
       11 = ヤマハ (Yamaha)
       12 = 日本ビクター (Nihon Victor = Japan Victor Company [JVC])
       13 = フィリップス (Philips)
       14 = パイオニア (Pioneer)
       15 = 三洋電機 (Sanyo Denki = Electric Machines)
       16 = シャープ (Sharp)
       17 = ソニー (Sony)
       18 = スペクトラビデオ (Spectravideo)
       19 = 東芝 (Toshiba)
       20 = ミツミ電機 (Mitsumi Denki = Electrical Machines)
       21 = テレマティカ (Telematica)
       22 = グラディエンテ (Gradiente)
       23 = シャープドブラジル (Sharp do Brasil = Brazilian Sharp)
       24 = Goldstar (now known as "LG")
       25 = Daewoo
       26 = Samsung
       27 \
      ...  > Reserved
      127 /
      128 \
      ...  > Device numbers
      255 /

    The Disk Operating System has a BIOS, from 0000H till 003AH, and one entry point at F37DH:

    0000H     Warm Bootstrap routine.
    0005H     BDOS (same as F37DH, but available only from MSX-DOS).
    000CH     RDSLT (same as MSX-BIOS RDSLT).
    0014H     WRSLT (same as MSX-BIOS WRSLT).
    001CH     CALSLT (same as MSX-BIOS CALSLT).
    0024H     ENASLT (same as MSX-BIOS ENASLT).
    0030H     CALLF (same as MSX-BIOS CALLF).
    0038H     INTRPT (same as MSX-BIOS INTRPT).
    F37DH     BDOS (same as 0005H, but available from both MSX-DOS and MSX-Disk BASIC).

    The MSX-DOS memory is mapped as follows:

  0000H-003BH    Disk-BIOS
  005CH-00FFH    System Scratch Area
  0080H          Parameter string length
  0081H-00FFH    Command line parameter string
  0100H          Transient Program Area (TPA)
                 COMMAND.COM
                 MSXDOS.SYS
                 MSX-DOS Work Area
  F37DH          BDOS alternative entry
        FFFFH    MSX Work Area

    The Disk-DOS and System Scratch Area are mapped as follows:

0000H		JP WBOOT     JP BDOS         JP RDSLT
0010H		JP WRSLT           JP CALSLT
0020H		JP ENASLT
0030H		JP CALLF           JP INTRPT
0040H
0050H
0060H
0070H
0080H		PL Command line parameter string
00F0H
0100H		Transient Program Area

    The BDOS entry contains several functions, which are selected by the value given in register C. Other input parameters required by the specific functions must be loaded on the other registers. DOS-BIOS functions may change any registers and return results in specific ones.

    The following functions are available on DOS-BIOS:

   00H        System Reset
   01H        Get Character from Console
   02H        send one character to console
   03H        get one character from auxiliary device
   04H        send one character to auxiliary device
   05H        send one character to printer
   06H        get one character from console (no input wait, no echo back, no control code check)/ one character output
   07H        get one character from console (input wait, no echo back, no control code check)
   08H        get one character from console (input wait, no echo back, control code check)
   09H        send string
   0AH        get string
   0BH        check input from console
   0CH        get version number
   0DH        disk reset
   0EH        select default drive
   0FH        Open File
   10H        Close File
   11H        Find First
   12H        Find Next
   13H        delete file
   14H        read sequential file
   15H        write sequential file
   16H        create file
   17H        rename file
   18H        get login vector
   19H        get default drive name
   1AH        set DMA address
   1BH        get disk information
   1CH-20H     no function
   21H        write random file
   22H        read random file
   23H        get file size
   24H        set random record field
   25H        no function
   26H        write random block
   27H        read random block
   28H        write random file (00H is set to unused portion)
   29H        no function
   2AH        get date
   2BH        set date
   2CH        get time
   2DH        set time
   2EH        set verify flag
   2FH        read logical sector
   30H        write logical sector

* Console input

Function:	01H
Setup:		none
Return value:	A register <-- one character from console

When there is no input (no key pressed and input buffer empty), an input is
wait for. Input characters are echoed back to the console. The following
control character input is allowed: Ctrl-C causes program execution to be
halted and a return to the MSX-DOS command level; Ctrl-P causes any sucessive
input to also echoed to the printer until Ctrl-N is accepted; Ctrl-S causes
the display to stop until any key is pressed.

	Ctrl-C ........ system reset
	Ctrl-P ........ printer echo
	Ctrl-N ........ halt printer echo
	Ctrl-S ........ pause display


* Console output

Function:	02H
Setup:		E register <-- character code to be sent out
Return value:	none

This system call displays the character specified by the E register on the
screen. It also checks the four control characters, listed above.


* External input

Function:	03H
Setup:		none
Return value:	A register <-- one character from AUX device

This system call checks four control characters.


* External output

Function:	04H
Setup:		E register <-- character code to send to AUX device
Return value:	none

This system call checks four control characters.


* Printer output

Function:	05H
Setup:		A register <-- one character from console

This system call does not echo back. It treats control characters in the same
way as function 01H.


* Direct console input/output

Function:	06H
Setup:		E register <-- character code to be send to the console
		When 0FFH is specified, the character will be input
		from the console.
Return value:	When the E register is set to 0FFH (input), the result of
		input is in the A register. The value set in the A register
		is the character code of the key, if it was pressed;
		otherwise, the value is 00H. When the E register is set to
		a value other than 0FFH (output), there is no return value.

This system call does not support control characters and does not echo back
input. This system call checks four control characters.


* Direct console input - 1

Function:	07H
Setup:		none
Return value:	A register <-- one character from console

This system call does not support control characters, nor echo back.


* Direct console input - 2

Function:	08H
Setup:		none
Return value:	A register <-- one character from console

This system call does not echo back. It treats control characters in the same
way as function 01H.


* String output

Function:	09H
Setup:		DE register <-- starting address of string, prepared on
				memory, to be sent to the console.
Return value:	none

24H ("$") is appended to the end of the string as the end symbol. This system
call checks and performs four control character functions, as listed
previously.


* String input

Function:	0AH
Setup:		The address of memory where the maximum number of input
		characters (1 to 0FFH) is set should be set in the DE
		register.
Return value:	Number of characters actually sent from console is set in the
		address, one added to the address indicated by the DE
		register; string sent from console is set in the area from
		the address, two added to the address indicated by the DE
		register.

Return key input is considered as the end of console input. However, when the
number of input characters exceeds the specified number of characters
(contents indicated by DE register, 1 to 255), characters within the
specified number of characters will be treated as an input string and set in
memory, and the operation ends. The rest of characters including the return
key are ignored. Editing with the template is available to string input using
this system call. This system call checks and performs four control character
function, as listed previously.


* Console status check

Function:	0BH
Setup:		none
Return value:	0FFH is set in the A register when the keyboard is being
		pressed; otherwise, 00H is set.

This system call checks and performs four control character function, as
listed previously.


4.2 Environment Setting and Readout

The following system calls set the MSX system environment; for example,
changing the default drive, or setting various default values of the system


* System reset

Function:	00H
Setup:		none
Return value:	none

When this is called in MSX-DOS, the system is reset by jumping to 0000H. When
MSX DISK-BASIC call this, it is "warm started". That is, it returns to BASIC
command level without destroying programs currently loaded.


* Version number acquisition

Function:	0CH
Setup:		none
Return value:	HL register <-- 0022H

This system call is for acquiring various CP/M version numbers, on MSX-DOS,
however, 0022H is always returned.


* Disk reset

Function:	0DH
Setup:		none
Return value:	none

If there is a sector which has been changed but not written to the disk, this
system call writes it to the disk, then it sets the default drive to drive A
and sets DMA to 0080H.


* Default drive setting

Funtion:	0EH
Setup:		E register <-- default drive number (A = 00H, B = 01H, ...)
Return value:	none

Disk access by the system calls are made to the drive indicated by the
default drive number, unless otherwise specified. Note that, when the drive
number, which is set in the FCB specified upon calling the system call, is
other than 00H, the default drive setting made by this system call is
ignored.


* Login vector acquisition

Function:	18H
Setup:		none
Return value:	HL register <-- online drive information

The online drive is the drive connected to MSX normally when the disk system
is booted up. Executing this system call causes each drive to be examined
whether it is online, and the result is returned in the HL register as shown
in Figure 3.22. When the bit is "1", the corresponding drive is online;
otherwise it is not.


   Figure 3.22	Login vector

------------------------------------------------------------------------
| register name  |	      H 	    |		 L	       |
|----------------+--------------------------+--------------------------|
|   bit number	 |  7  6  5  4	3  2  1  0  |  7  6  5	4  3  2  1  0  |
|----------------+--------------------------+--------------------------|
|   drive name	 |  meaningless on MSX-DOS  |  H: G: F: E: D: C: B: A: |
|----------------+-----------------------------------------------------|
| online/offline |  1 is set for online and 0 for offline in each bit  |
------------------------------------------------------------------------


* Default drive acquisition

Function:	19H
Setup:		none
Return value:	A register <-- defaut drive number (A = 00H, B = 01H, ...)


* Setting of address to be transferred to

Function:	1AH
Setup:		DE register <-- address setting to be transferred to
				(DMA address)
Return value:	none

Though DMA address is initialized to 0080H at system reset, it can be reset
to any address by using this system call.


* Disk information acquisition

Function:	1BH
Setup:		E register <-- number of the objective drive
			       (default drive = 00H, A = 01H, B = 02H, ...)
Return value:	A register <-- number of logical sectors per one cluster
			       (FFH if E register is set inappropriate)
		BC register <-- logical sector size
		DE register <-- amount of clusters
		IX register <-- DPB starting address
		IY register <-- FAT starting address on memory

This system call gets the information about the disk in the specified drive.
Specifying 00H for the drive number specifies the default drive. For other
than that, specify 01H for drive A, 02H for drive B, and so on.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Date acquisition

Function:	2AH
Setup:		none
Return value:	HL register <-- year
		D register <-- month
		E register <-- day of month
		A register <-- day of week

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Date setting

Function:	2BH
Setup:		HL register <-- year
		D register <-- month
		E register <-- day of month
Return value:	A indicates whether the system call has done succesfully. If
		successful, the A register is set to 00H; otherwise, 0FFH.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Time acquisition

Function:	2CH
Setup:		none
Return value:	H register <-- hour
		L register <-- minute
		D register <-- second
		E register <-- 1/100 second

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Time setting

Function:	2DH
Setup:		H register <-- hour
		L register <-- minute
		D register <-- second
		E register <-- 1/100 second
Return value:	If successful, the A register is set to 00H; otherwise, 0FFH

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Verify flag setting

Function:	2EH
Setup:		E register <-- 00H, when resetting verify flag
		E register <-- value other than 00H, when setting
			       the verify flag
Return value:	none

Setting the verify flag causes successive writing to the disk to be done in
mode "verify on". That is, by reading the contents written on the disk, the
check is made to compare them with the contents to be written.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


4.3 Absolute READ/WRITE (direct access to sectors)

MSX manages the disk with the "logical sector" as a minimum unit. The logical
sector is defined independent of the physical secotrs of the disk, and is
numbered from 0 to the maximum logical sector (maximum number depends on the
kind of the disks).

Logical sectors enable users of MSX-DOS or MSX DISK-BASIC to access the disk
without being concerned about the number of physical sectors per track, where
that number depends on the media type of the disk. In fact, by utilizing
system calls which use FCB (file control block), the file can be easily
handled in detail even without considering logical sectors, so the user does
not even need to use logical sectors. But for some purposes, access using
logical sectors is desirable, so MSX-DOS ans MSX DISK-BASIC offer system
calls which can access logical sectors.

This section describes the system calls which access the disk by use of
logical sectors.


* Reading from the disk using logical sectors

Function:	2FH
Setup:		The logical sector number to be read from (for more than
		one logical sector, the starting logical sector number)
		should be set in the DE register. The number of logical
		sectors to be read should be set in the H register, and
		the drive number (00H for drive A, 01H for drive B, and
		so on. The same follows for function 30H below) to be used
		to read should be set in the L register.
Return value:	The contents read are set in the DMA buffer.

This system call reads out a specified number of continuous logical sectors
from the specified logical sector of the specified drive and stores the
contents in memory after DMA. It then stores the contents of what it has read
in memory after DMA. Function 1AH (which specifies the address it is to be
transferred to) assures that there is enough available space in memory.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Writing to the disk using logocal sectors

Function:	30H
Setup:		Contents to be written should be set in memory area after
		the address indicated by DMA. The logical sector number from
		where the writing begins should be set in the DE register.
		The number of logical sectors to be written should be set
		in the H register. The drive number to be written to should
		be set in the L register.
Return value:	none

This system call has been created for MSX-DOS and is not compatible with
CP/M.



4.4 File Access Using FCB

Since accessing a file is difficult when using the system calls described in
the previous section (which read and write logical sectors directly), system
calls using FCB are needed to easier access the disk by specifying files.

There are three categories of system calls using FCB. First is sequential
file access and second is random file access, both are offered to maintain
CP/M compatibility. The third is what gives MSX-DOS its power: random block
access. This method is not available in CP/M. Random block access has the
following features:

	* Any record size can be specified
	* Random access can be made to multiple records
	* File size can be controlled in bytes

This section describes system calls for file access using FCB, including
random block access. Note that the following three functions do not work
correctly when FCB is in the address range 4000H to 7FFFH:

	1. Function call 11H
	2. Function call 12H
	3. Input/output for devices (CON, PRN, NUL, AUX)


* Opening files

Function:	0FH
Setup:		DE register <-- starting address of FCB which is not opened
Return value:	00H is set in the A register when a file is opened
		succeessfully; otherwise 0FFH is set. When a file is opened
		successfully, each field of the FCB is set.

When 00H is specified for a drive number, the default drive set by function
0EH (default drive setting) is used. To open a file on another drive, specify
01H for drive A, 02H for drive B and so on.

When a file is opened by this system call, all FCB fields except record size,
current block, current record, and random record are set using information
from the directory area on the disk. Fields which are not set should be set
by the user after executing this system call, if needed. The state that each
field of FCB is set is "the state that file is opened" when using system
calls using FCB, and, in this case, system calls which access the file using
FCB, described below, can be used.


* Closing files

Function:	10H
Setup:		DE register <-- starting address of opened FCB
Return value:	00H is set in the A register when file is closed
		scuccessfully; otherwise, 0FFH is set.

By writing the current contents of FCB in memory to the corresponding
directory area on the disk, file information can be kept current. When the
file is only read, it does not need to be closed by using this system call.


* File search - 1

Function:	11H
Setup:		DE register <-- starting address of FCB which is not opened
Return value:	00H is set in the A register when the file is found;
		otherwise 0FFH is set. When the file is found, the directory
		entry (32 bytes) of the file on the disk is set in the area
		indicated by DMA, and FCB drive number is set (thus, 33 bytes
		are used).

Wildcard characters can be used in the name of the file. For example, a
specification "????????.c" causes any file name with an extension of "c" to
be searched for, and the directory information of the file first found is
written in after DMA. To find all matching files or to see whether there is
only one matching file, use function 12H described below.


* File search - 2

Function:	12H
Setup:		none
Return value:	00H is set in the A register when the file is found;
		otherwise 0FFH is set. When the file is found, the directory
		entry (32 bytes) of the file on the disk is set in the area
		indicated by DMA, and the FCB drive number is set
		(thus, 33 bytes are used).

This system call should be used to search for multiple files meeting the file
name specification by wildcard characters in function 11H. So this function
should not be used by itself.

This system call allows the directory information of files meeting the
specifications in function 11H to be listed in order, one by one.


* Deleting files

Function:	13H
Setup:		DE register <-- starting address of FCB which is not opened
Return value:	00H is set in the A register when file is successfully
		deleted., otherwise 0FFH is set.

Using wildcard characters for the file name may cause more than one file to
be deleted. Exercise caution when using wildcards to delete files.


* Sequential readout

Function:	14H
Setup:		DE register <-- starting address of opened FCB
		FCB current block <-- starting block for readout
		FCB current record <-- starting record for readout
Return value:	00H is set in the A register when readout is successful;
		otherwise 01H is set. When successful, one record which has
		been read is set in the area indicated by DMA.

The FCB current block and record will be updated automatically after the
readout. That is, in successive readouts, the current block and record do not
need to be set. The record size for readout is fixed at 128 bytes.


* Sequential writing to the disk

Function:	15H
Setup:		DE register <-- starting address of opened FCB
		FCB current block <-- starting block for writing
		FCB current record <-- starting record for writing
		128 bytes starting from DMA <-- data to be written
Return value:	00H is set in the A register when writing is successful;
		otherwise 01H is set.

The FCB current block and record will be updated automatically after the
readout.


* Creating files

Function:	16H
Setup:		DE register <-- starting address of FCB which is not opened
Retu4rn value:	00H is set in the A register when the file is created
		successfully; otherwise 0FFH is set.

The record size, current block and record, and the random record of the FCB
should be set after executing this system call.


* Renaming files

Function:	17H
Setup:		New file name should be set within 11 bytes after the
		18th byte of the FCB (2nd byte of file size field
		of FCB = 16 bytes after old file name) corresponding to
		old file name (that is, it should be set in 18th to 28th
		byte), the FCB address should be set in the DE register.
Return value:	00H is set in the A register when the file name is renamed
		successfully; otherwise 0FFH is set.

Wildcard characters can be used for both the new and old file names. For
example, specifying "????????.o" for the old file name and "????????.obj" for
the new file name causes the extension of all files having ".o" to be changed
to ".obj".


* Random reading from the disk

Function:	21H
Setup:		DE register <-- starting address of opened FCB
		random record in FCB <-- record for readout
Return value:	00H is set in the A register when readout is successful;
		otherwise 01H is set. When successful, the contents of one
		record which has been read are set in the area indicated
		by DMA.

The lenght of the record is fixed to 128 bytes.


* Random writing to the disk

Function:	22H
Setup:		DE register <-- starting address of opened FCB
		random record in FCB <-- record to be written to
		128 bytes starting from DMA <-- data to be written
Return value:	00H is set in the A register when writing is successful;
		otherwise 01H is set.

The lenght of the record is fixed to 128 bytes.


* File size acquisition

Function:	23H
Setup:		DE register <-- starting address of FCB which is not opened
Return value:	00H is set in the A register when the function is successful;
		otherwise 00H is set. When successful, the size of the
		specified file is set in increments of 128 bytes, in the
		first three bytes of the random record field.

The file size is calculated in increments of 128 bytes. That is, 2 would be
set for files ranging in size from 129 bytes to 256 bytes. Thus a file with
257 bytes would return a value of 3.


* Random record field setting

Function:	24H
Setup:		DE register <-- starting address of opened FCB
		FCB current block <-- objective block
		FCB current record <-- objective record
Return value:	Current record position, calculated from the current block
		and record fields of specified FCB, is set in the random
		record field.


* Random writing to the disk - 2 (random block access)

Function:	26H
Setup:		DE register <-- the starting address from the FCB
		FCB record size <-- size of record to be written
		FCB random record <-- the record ID number
		HL register <-- the number of records to be written
		DMA memory buffer <-- the data to be written
Return value:	00H is set in the A register when writing is successful;
		otherwise 01H is set.

After writing to the disk, the value of the random record field is
automatically updated and points to the next record. The size of one record
can be set to any value from 1 byte to 65535 bytes by setting the desired
value in the FCB record size field. When 0 records are to be written, the
file lenght is calculated at the record size multiplied by the record number.
The rest is discarded.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Random readout - 2 (random block access)

Function:	27H
Setup:		DE register <-- starting address of opened FCB
		FCB record size <-- record size to be read
		FCB random record <-- record to start reading
		HL register <-- number of records to be read
Return value:	00H is set in the A register when data is read successfully;
		otherwise 01H is read. The number of records actually read
		is set back in the HL register. When this number is almost
		one, the data which has been read is set in the area
		indicated by DMA.

After readout, the random record field is automatically updated. After
executing this system call, the total number of records actually read is set
in the HL register. That is, if the end of file is reached before the
specified number of records have been read, the actual number of records read
will be returned in the HL register.

This system call has been created for MSX-DOS and is not compatible with
CP/M.


* Random writing - 3

Function:	28H
Setup:		DE register <-- starting address of opened FCB
		FCB random record <-- record to be written
		128 bytes in DMA buffer <-- data to be written
Return value:	00H is set in the A register when writing is successful;
		otherwise, 01H is set.

The lenght of records is fixed at 128 bytes.

This system call is the same as 22H except for one point. When the file
becomes large, 00H is written to the added records coming before the
specified record.

- A -                                "CLEAR" 163                          .
                                     CLIKFL 57, 231                       .
"ABS" 97                             CLIKSW 57, 197, 212                  .
Angles 152                           CLINEF 152, 225                      .
ARG 90, 223                          Clipping 70                          .
Array storage 155, 162, 208          CLMLST 126, 211                      .
ARYTAB 137, 208, 220                 "CLOAD" 180                          .
"ASC" 168                            CLOC 70, 225                         .
ASPCT1 77, 215                       "CLOSE" 173                          .
ASPEC2 77, 215                       CLPRIM 31, 210                       .
ASPECT 150, 225                      "CLS" 41                             .
"ATN" 93                             CMASK 70, 225                        .
ATRBAS 34, 225                       "CMD" 202                            .
ATRBYT 69, 72, 76, 144, 214          CNPNTS 150, 225                      .
"ATTR$" 203                          CNSDFG 49, 212                       .
AUTFLG 116, 219                      Coincidence 9, 52                    .
AUTINC 116, 219                      "COLOR" 197                          .
AUTLIN 116, 219                      Colours 12, 39, 40, 77               .
"AUTO" 125                           CONLO 121, 217                       .
                                     Control codes 44                     .
- B -                                CONSAV 121, 217                      .
                                     "CONT" 162                           .
BAKCLR 39, 213                       CONTXT 121, 217                      .
"BASE" 200, 201                      CONTYP 121, 217                      .
BASROM 32, 86, 230                   Coordinates, graphic 70, 143         .
Baud rate 79, 197                    Coordinates, text 42, 47, 51         .
BDRATR 77, 235                       "COPY" 202                           .
BDRCLR 40, 213                       "COS" 92                             .
"BEEP" 60                            CPCNT8 152, 226                      .
"BIN$" 165                           CPLOTF 150, 226                      .
"BLOAD" 177                          Crash 30                             .
BOTTOM 204, 232                      CRCSUM 151, 226                      .
Boundary inflections 148             CRTCNT 47, 51, 211                   .
"BSAVE" 177                          CS1200 197, 214                      .
BUF 85, 127, 176, 216                CS2400 197, 214                      .
                                     "CSAVE" 179                          .
- C -                                CSAVEA 78, 149, 226                  .
                                     CSAVEM 78, 149, 226                  .
"CALL" 140                           CSCLXY 150, 226                      .
Caps LED 6, 57                       "CSNG" 100                           .
CAPST 57, 234                        "CSRLIN" 196                         .
CASPRV 184, 235                      CSRSW 45, 234                        .
Cassette input 24, 81                CSRX 43, 212                         .
Cassette motor 6, 65, 78             CSRY 43, 212                         .
Cassette output 6, 79                CSTCNT 150, 226                      .
"CDBL" 100                           CSTYLE 45, 85, 234                   .
CENCNT 150, 225                      CTRL-STOP 32, 33, 42, 57, 59         .
CGPBAS 34, 224                       CURLIN 56, 115, 116, 216             .
CGPNT 38, 204, 224, 256              CURSAV 46, 231                       .
Character set 38, 84, 256            Cursor 13, 32, 42, 43, 46            .
"CHR$" 168                           "CVD" 204                            .
"CINT" 99                            "CVI" 203                            .
"CIRCLE" 150                         "CVS" 204                            .
                                                                          .
                                                                          .
                                                                          .
- i1 -                                                                    .
                                                                          .
                                                                          .
INDEX                                                                     .
                                                                          .
                                                                          .
                                                                          .
CXOFF 152, 226                       Expander 4                           .
CYOFF 152, 226                       Extension ROM 30, 140, 178, 205      .
                                     Expression Evaluator 128             .
- D -                                EXPTBL 5, 32, 236                    .
                                                                          .
DAC 90, 223                          - F -                                .
"DATA" 123                                                                .
Data areas 26                        Factor Evaluator 129                 .
DATLIN 115, 128, 218                 FBUFFR 106, 137, 222                 .
DATPTR 161, 220                      "FIELD" 203                          .
Dead key 56, 58                      File Control Block 66, 209           .
"DEFDBL" 121                         "FILES" 173                          .
"DEFFN" 134                          Filespec 171                         .
"DEFINT" 121                         FILNAM 171, 180, 224                 .
"DEFSNG" 121                         FILNM2 181, 224                      .
"DEFSTR" 12I                         FILTAB 208, 223                      .
DEFTBL 121, 221                      "FIX" 101                            .
"DEFUSR" 134                         FLGINP 127, 219                      .
"DELETE" 137                         "FN" 134                             .
Dequeueing 60                        FNKFLG 56, 231                       .
DEVICE 142, 237                      FNKSTR 65, 224                       .
Device 140, 171, 178                 "FOR" 119                            .
"DIM" 155                            FORCLR 40, 213                       .
DIMFLG 155, 217                      "FPOS" 175                           .
DOT 115, 219                         "FRE" 171                            .
"DRAW" 153                           FRETOP 159, 208, 218                 .
DRWANG 154, 235                      FSTPOS 84, 88, 230                   .
DRWFLG 154, 235                      FUNACT 159, 222                      .
DRWSCL 154, 235                      Function addresses 110, 130          .
"DSKF" 202                           Function key display 49, 53          .
"DSKI$" 203                          Function keys 56, 58, 65             .
"DSKO$" 202                                                               .
                                     - G -                                .
- E -                                                                     .
                                     "GET" 193                            .
Editor 84                            GETPNT 33, 59, 214                   .
Edit keys 85                         "GOSUB" 122                          .
"ELSE" 123, 125                      "GOTO" 123                           .
"END" 161                            Graphic characters 42                .
ENDFOR 119, 218                      Graphic output 69                    .
ENSTOP 53, 230                       GRPACX 69, 143, 235                  .
"EOF" 175                            GRPACY 69, 143, 235                  .
"ERASE" 162                          GRPATR 36, 212                       .
"ERL" 130                            GRPHED 42, 234                       .
"ERR" 130                            GRPNAM 36, 212                       .
ERRFLG 115, 130, 215                 GRPPAT 36, 212                       .
ERRLIN 115, 125, 130, 219            GXPOS 144, 235                       .
"ERROR" 125                          GYPOS 144, 235                       .
Error generators 115, 176                                                 .
Error handler 115                    - H -                                .
Error messages 113                                                        .
ERRTXT 115, 125, 220                 HEADER 79, 215                       .
ESCCNT 43, 234                       "HEX$" 165                           .
"EXP" 94                             HIMEM 163, 204, 206, 232             .
                                                                          .
                                                                          .
                                                                          .
- i2 -                                                                    .
                                                                          .
                                                                          .
INDEX                                                                     .
                                                                          .
                                                                          .
                                                                          .
Hooks 30, 204, 237                   Links 117                            .
                                     LINL32 34, 211                       .
- I -                                LINL40 34, 211                       .
                                     LINLEN 34, 211                       .
ID Bytes A0                          LINTTB 39, 84, 230                   .
"IF" 125                             LINWRK 46, 232                       .
"INKEY$" 187                         "LIST" 137                           .
"INP" 114                            "LLIST" 136                          .
"INPUT" 127                          "LOAD" 172                           .
"INPUT$" 174                         "LOC" 174                            .
Input, keyboard 53, 59               "LOCATE" 193                         .
INSFLG 85, 234                       "LOF" 174                            .
"INSTR" 170                          "LOG" 93                             .
"INT" 101                            LOWLIM 81, 233                       .
INTCNT 52, 233                       "LPOS" 132                           .
Interrupt mode 32                    "LPRINT" 125                         .
Interrupts 9, 52, 56, 161            LPTPOS 83, 125, 132, 215             .
"INTERVAL" 194                       "LSET" 203                           .
INTFLG 32, 57, 233                                                        .
INTVAL 52, 233                       - M -                                .
I/O Buffer 66, 125, 130, 171                                              .
I/O Dispatcher 178                   Macro parser 142                     .
"IPL" 202                            Mainloop 116                         .
                                     Math constants 97                    .
- J -                                Math operators 112                   .
                                     MAXDEL 147, 225                      .
Japanese 4, 13, 25, 26, 58, 88,      MAXFIL 171, 223                      .
  211, 234                           "MAXFILES" 206                       .
JIFFY 52, 196, 233                   MAXUPD 147, 213                      .
Joystick 24, 52, 63                  MCLFLG 142, 188, 227                 .
                                     MCLLEN 142, 188, 229                 .
- K -                                MCLPTR 142, 188, 229                 .
                                     MCLTAB 142, 188, 227                 .
KANAMD 33, 234                       MEMSIZ 159, 163, 208, 217            .
KANAST 58, 234                       "MERGE" 172                          .
Kansas City 79                       "MID$" 169, 170                      .
KBUF 116, 118, 216                   MINDEL 147, 225                      .
"KEY" 195                            MINUPD 147, 213                      .
Keyboard 5, 53, 55, 66               "MKD$" 203                           .
KEYBUF 33, 231                       "MKI$" 203                           .
Key click 6, 57, 58                  "MKS$" 203                           .
Key numbers 54                       MLTATR 36, 212                       .
Keywords, BASIC 110                  MLTNAM 36, 212                       .
"KILL" 202                           MLTPAT 36, 212                       .
                                     "MOTOR" 187                          .
- L -                                MOVCNT 149, 227                      .
                                     MSX2 228, 239, 240                   .
"LEFT$" 169                          MSX2+ 239, 240                       .
"LEN" 168                            MUSICF 52, 62, 229                   .
"LET" 123                            Music packet 61, 193                 .
"LFILES" 173                                                              .
"LINE" 145                           - N -                                .
"LINE INPUT" 126                                                          .
Line numbers 117, 118, 119, 122      NAMBAS 34, 224                       .
                                                                          .
                                                                          .
                                                                          .
- i3 -                                                                    .
                                                                          .
                                                                          .
INDEX                                                                     .
                                                                          .
                                                                          .
                                                                          .
"NAME" 202                           Printer 33, 41, 83                   .
"NEW" 159                            PRMLEN 159, 221                      .
NEWKEY 53, 231                       PRMLN2 159, 222                      .
Newton-Raphson 93                    PRMSTK 159, 221                      .
"NEXT" 164                           PROCNM 140, 237                      .
NLONLY 159, 224                      Program storage 117, 208             .
NOFUNS 135, 156, 222                 PRSCNT 188 228                       .
NTMSXP 83, 197, 216                  PRTFLG 83, i25, 215                  .
NULBUF 208, 223                      "PSET" 144                           .
Numeric output 106                   PSG 21, 33, 34, 60, 190              .
Numeric types 104                    PTRFIL 66, 125, 223                  .
                                     "PUT" 193                            .
- O -                                PUTPNT 33, 59, 214                   .
                                                                          .
"OCT$" 164                           - Q -                                .
OLDKEY 33, 53, 231                                                        .
OLDLIN 115, 162, 220                 QUEBAK 68, 228                       .
OLDSCR 34, 136, 234                  QUETAB 34, 67, 228                   .
OLDTXT 60, 115, 159, 162, 220        Queue 60, 67, 68, 189                .
"ON" 124                             QUEUEN 62, 229                       .
ONEFLG 124, 159, 220                 QUEUES 68, 214                       .
ONELIN 115, 124, 159, 220                                                 .
ONGSBF 56, 160, 231                  - R -                                .
"OPEN" 172                                                                .
"OUT" 114                            R800 244                             .
Output, Interpreter 82               RDPRIM 30 210                        .
Output, screen 43, 69                "READ" 104, 115, 127, 128, 175, 219  .
                                     "REM" 123                            .
- P -                                "RENUM" 138                          .
                                     REPCNT 53, 214                       .
"PAD" 197                            "RESTORE" 161                        .
Paddle 25, 64                        "RESUME" 124                         .
PADX 65, 233                         "RETURN" 123                         .
PADY 65, 233                         RG0SAV (till RG7SAV) 35, 212         .
Page 3                               RG8SAV (till RG27SAV) ?              .
"PAINT" 148                          "RIGHT$" 169                         .
PARM1 135 156, 221                   "RND" 94                             .
PARM2 135 222                        RNDX 94, 159, 223                    .
PATBAS 34, 224                       RS2IQ 228                            .
PATWRK 69, 232                       "RSET" 203                           .
"PDL" 196                            "RUN" 122                            .
"PEEK" 138                           RUNBNF 177, 235                      .
"PLAY" 188, 196                      Runloop 120                          .
PLYCNT 62, 189, 229                                                       .
"POINT" 144                          - S -                                .
Pointers 118, 139                                                         .
"POKE" 138                           "SAVE" 172                           .
Polynomial 95                        SAVENT 179, 235                      .
"POS" 132                            SAVSTK 115, 120, 159, 219            .
Power-up 32, 204                     SAVTXT 120, 219                      .
Precedence 112, 129                  SCNCNT 52, 214                       .
"PRESET" 144                         "SCREEN" 197                         .
Primary Slot 3, 30, 32, 66           SCRMOD 34, 136, 234                  .
"PRINT" 125                          Secondary Slot 4, 30, 32, 239        .
                                                                          .
                                                                          .
                                                                          .
- i4 -                                                                    .
                                                                          .
                                                                          .
INDEX                                                                     .
                                                                          .
                                                                          .
                                                                          .
"SET" 202                            - U -                                .
"SGN" 98                                                                  .
"SIN" 92                             "USR" 133                            .
SKPCNT 149, 227                      USRTAB 133, 210                      .
Slot ID 29, 140, 205                                                      .
SLTATR 140, 205, 236                 - V -                                .
SLTTBL 32, 236                                                            .
"SOUND" 188                          "VAL" 169                            .
"SPACE$" 169                         VALTYP 104, 217                      .
"SPRITE" 194, 197, 198               Variable storage 135, 155, 208       .
Sprites 9, 18, 37, 199               "VARPTR" 130                         .
"SQR" 93                             VARTAB 137, 208, 220                 .
Stack space 159                      VCBA 34, 61, 67, 230                 .
Standard routines 26                 "VDP" 199, 200                       .
Statement addresses 109, 136         VDP Address Register 8, 40           .
STATFL 52, 213                       VDP Mode Registers 9, 35             .
"STICK" 196                          VDP Modes 10, 13, 34, 41             .
STKTOP 159, 163, 209, 217            VDP Status Register 9, 52, 65        .
"STOP" 161, 194                      VDP Timing 39                        .
"STR$" 165                           "VPEEK" 201                          .
STREND 137, 208, 220                 "VPOKE" 201                          .
"STRIG" 194, 196                     VLZADR 115, 216                      .
"STRING$" 168                        VLZDAT 115, 216                      .
String storage 133, 163, 166, 168                                         .
SUBFLG 155, 218                      - W -                                .
"SWAP" 162                                                                .
SWPTMP 162, 222                      "WAIT" 114                           .
                                     Wait state 81                        .
- T -                                Warm start 53                        .
                                     "WIDTH" 136                          .
T32ATR 34, 211                       WINWID 81, 233                       .
T32CGP 34, 211                       Workspace Area 32, 208               .
T32NAM 34, 211                       WRPRIM 30, 210                       .
T32PAT 34, 211                                                            .
"TAN" 93                             - Z -                                .
TEMPPT 159, 218                                                           .
TEMPST 124, 218                      Z80 Clock 81                         .
"TIME" 196                                                                .
Tokens 111, 117, 118, 121, 137                                            .
Touchpad 25, 64                                                           .
TRCFLG 120, 222                                                           .
TRGFLG 52, 213                                                            .
"TROFF" 162                                                               .
"TRON" 162                                                                .
TRPTBL 32, 56, 160, 194, 232                                              .
TTYPOS 43, 125, 132, 216                                                  .
turbo R 65, 239, 240                                                      .
TXTCGP 34, 211                                                            .
TXTNAM 34 211                                                             .
TXTTAB 117, 204, 208, 218                                                 .
Types 140                                                                 .
                                                                          .
                                                                          .
                                                                          .
                                                                          .
                                                                          .
                                                                          .
- i5 -