Welcome to the premier issue of the Apple Assembly Line!
This new monthly newsletter is dedicated to the many Apple owners using assembly language, or who would like to learn how. Articles will include commented disassemblies of Apple ROM routines, DOS, and other commercial software; how to augment and modify existing products; beginner's lessons in assembly language; handy subroutines every programmer needs in his tool kit; and many more.
In this issue you will find a tutorial on efficient ways to increment and decrement multiple-byte values, a very powerful subroutine for formatting messages on the screen, and patch code for the S-C ASSEMBLER II Version 4.0 to "adapt" it to the Paymar Lower-Case Adapter. There is also an article describing a recently reported error found in ALL 6502 chips, and a brief announcement of some new products from S-C SOFTWARE.
Since there will be a lot of source code printed in this and forthcoming issues of the Apple Assembly Line, I plan to offer quarterly diskettes containing all published source code (in the format of the S-C ASSEMBLER II Version 4.0) at a nominal price. How does $15 per quarter sound? Of course, you can always type it in.... The articles should be considered copy-righted, but feel free to use the code in any way you can. It is printed here for your enlightenment, entertainment, and for your USE. I hope you find it all helpful.
I do not know all there is to know about the 6502, or the Apple, or about anything! Nor do I have an infinite amount of time. Therefore, I will be happy to accept articles and programs from you. I may print them exactly as you write them, or I may modify them first. In any case, you will get credit, and the satisfaction of knowing you are helping many others in their conquest of the computer.
If you know others who should be receiving this newsletter, spread the word! If you are not subscribing yet, then send your $12 today! If you have any comments about the content, format, or whatever, write now! Or, you can call me during reasonable hours at (ddd) ddd-dddd.
| Sincerely, <<signature>> Bob Sander-Cederlof |
In This Issue...
I suppose there are as many ways to do it as there are programmers. Some are short and fast, some long and slow, some neat, some not so neat.
Adding one to a number is called “incrementing”, and subtracting one is called “decrementing”. The 6502 has two instructions for these two functions: INC and DEC. (For the moment I will overlook the four instructions for doing the same to the X and Y registers: INX, INY, DEX, and DEY.) It is easy to see how to use INC and DEC on single-byte values; with a little more trouble we can also use them for values of two or more bytes.
Adding one using ADC
CLC
LDA VALL Add 1 to low
ADC #1 byte
STA VALL
LDA VALH Add carry to
ADC #0 high byte
STA VALH
|
Adding one using INC
INC VALL Add 1 to low byte
BNE .1 Skip if no carry
INC VALH Add carry
.1 ....
|
Of course, there are many other variations. Not indicated here are the various address modes possible in addressing VALL and VALH. They may be in page zero, or elsewhere. They may be directly addressed, as shown above; or, you may use the indirect indexed, indexed indirect, and plain indexed modes.
It is easy to see how to extend both methods above to triple precision, or more. Here is a three-byte version using INC:
INC VALL Increment low byte
BNE .1 Unless zero, no carry
INC VALM Increment middle byte
BNE .1 Unless zero, no further carry
INC VALH Increment high byte
.1 ....
|
Believe it or not, there is one disadvantage to using INC code like this. Sometimes code is required to have a constant running time regardless of the data values. Then, you either must use straight line code with ADC instructions, or add complicated padding code to the INC form.
How about subtracting one? Here are two ways to do it to two-byte values:
Decrementing with SBC
SEC
LDA VALL Low byte - 1
SBC #1
STA VALL
LDA VALH High byte - borrow
SBC #0
STA VALH
|
Decrementing with DEC
LDA VALL Need to borrow?
BNE .1 No
DEC VALH Yes, hi-byte - 1
.1 DEC VALL Low byte - 1
|
Which one do you like better? It is still a matter of taste, unless the amount of memory used or time consumed is very important. There are also different side effects, such as the final state of the Carry status. INC and DEC do not change the Carry status, while of course ADC and SBC do. You may wish to preserve Carry through the process, making the INC/DEC code preferable. Or, you may wish to know the resulting Carry status after incrementing or decrementing for some reason; then you should use the ADC/SBC code.
Back to subtracting one...how about doing it to a three-byte value? We just add three more lines:
LDA VALL See if need to borrow
BNE .2 No
LDA VALM See if need to borrow again
BNE .1 No
DEC VALH Borrow from high byte
.1 DEC VALM Borrow from middle byte
.2 DEC VALL Decrement low byte
|
Easier than you thought, right? You would not believe the many strange ways I have seen this operation coded in commercial software (even some released by Apple themselves!). Yet it seems to me that this method is the same way we would do it with pencil and paper in decimal arithmetic. Think how you would do this:
123040
-1
------
xxxxxx
|
If you think of each digit as though it were a byte...isn't the algorithm the same?
Now it is time for all of us to go back over the programs we wrote during the past three years for the Apple, and replace a lot of old code!
Bob Sander-Cederlof, October, 1980.
New Products from S-C SOFTWARE
As many of you know, because you have already bought it, version 4.0 of the S-C Assembler II is now on the market. With this new version, the price has gone up from $35 to $55. An upgrade kit for owners of previous versions is only $22.50
Now another new version is available, for those of you without disks! Tape Version 4.0 requires only 16K RAM and a cassette drive. The price is $45 for the complete package, or $22.50 for an upgrade kit from the previous tape version. All of the new features of Disk Version 3.2 and 4.0 are included, except those which require a disk drive. For the time being, the manual consists of a copy of the disk version 4.0 manuals, with a single sheet describing the differences in the tape version. Purchasers of tape version 4.0 will be able to upgrade to the disk version when they get a disk drive, for only $12.50.
And still another version of the assembler! This one is a cross assembler for the Motorola 6800, 6801, and 6802 microprocessors. It has all the features of the S-C Assembler II Disk Version 4.0, but the source language accepted is that of the 6800 family rather than the 6502. The price for this package is only $300, which is less than a month of time-sharing services for an equivalent capability would cost! An Apple, a ROM blower from Mountain Hardware, and the S-C Assembler II-6800 are all you need for a full-blown development system.
Formatting a series of nice messages or screens-full of messages is hard enough to do in Applesoft...but in assembly language it can really be a difficult job. And it seems to take so much memory to do the equivalent of VTAB, HTAB, HOME, and PRINT. I was recently motivated to do something about this for a large, verbose program. I designed a general subroutine for printing text, which can print all 128 chracters of ASCII, plus do some fancy footwork on the way.
Embedded control codes in the text to be printed perform such handy functions as HTAB, VTAB, HOME, NORMAL, INVERSE, Clear to End of LIne, Clear to End of Page, Two-Second Delay, and Repeat. All characters to be printed directly are entered with the high-order bit set to one; bytes with the high order bit zero are control codes. Comments in lines 1250-1350 of the listing show what the codes are.
To simplify the calling sequence, a table of message addresses is built along with the messages themselves. To print a specific message, merely load the message index number into the A-register (LDA #0 for the first message, LDA #1 for the second, etc.), and JSR MESSAGE.PRINTER. Some sample messages are given in the listing, starting at line 2240.
There are a lot of unused control codes, which you can use to augment the subroutine. I am planning to add a code to switch to a HI-RES TEXT driver, for writing text on either of the two Hi-Res screens. You can probably think of a lot of useful ones yourself. The point is that this type of subroutine can simplify programming of an interactive program, and save memory too.
1000 *---------------------------------
1010 MON.CH .EQ $24
1020 MON.CV .EQ $25
1030 MON.VTAB .EQ $FC22
1040 MON.CLREOP .EQ $FC42
1050 MON.HOME .EQ $FC58
1060 MON.CLREOL .EQ $FC9C
1070 MON.WAIT .EQ $FCA8
1080 MON.COUT .EQ $FDED
1090 MON.NORMAL .EQ $FE84
1100 MON.INVERSE .EQ $FE80
1110 *---------------------------------
1120 MSG.PNTR .EQ $18,19
1130 MSG.SCANNER .EQ $1A
1140 *---------------------------------
1150 * MESSAGE PRINTER
1160 *
1170 * CALL:
1180 * (A) = MESSAGE # (0-N)
1190 * JSR MESSAGE.PRINTER
1200 *
1210 * ACTION:
1220 * 1. FINDS SPECIFIED MESSAGE
1230 * 2. PRINTS ON THE SCREEN
1240 * 3. INTERPRETS CHARACTERS AS FOLLOWS:
1250 * $00 END OF MESSAGE
1260 * $01-28 HTAB 1-40
1270 * $40-57 VTAB 1-24
1280 * $60 CLEAR SCREEN, HOME CURSOR
1290 * $61XXYY REPEAT CHARACTER YY, XX TIMES
1300 * $62 DELAY ABOUT TWO SECONDS
1310 * $63 NORMAL MODE
1320 * $64 INVERSE MODE
1330 * $65 CLEAR TO END OF LINE
1340 * $66 CLEAR TO END OF SCREEN
1350 * $80-FF PRINT AS IS
1360 *
1370 *---------------------------------
1380 MESSAGE.PRINTER
1390 ASL DOUBLE MSG NUMBER TO GET INDEX
1400 TAY
1410 LDA MESSAGE.ADDRESS.TABLE,Y
1420 STA MSG.PNTR
1430 LDA MESSAGE.ADDRESS.TABLE+1,Y
1440 STA MSG.PNTR+1
1450 LDA #0
1460 STA MSG.SCANNER
1470 .1 JSR GET.NEXT.CHAR.FROM.MESSAGE
1480 BNE .3
1490 RTS $00: EOM
1500 .3 BPL .5 SPECIAL ACTION
1510 JSR MON.COUT PRINT THE CHARACTER
1520 .4 JMP .1
1530 *---------------------------------
1540 .5 CMP #$40 CHECK FOR VTAB
1550 BCS .6 YES
1560 CMP #$29 IN RANGE FOR HTAB?
1570 BCS .4 NO, IGNORE
1580 STA MON.CH
1590 DEC MON.CH
1600 BCC .4 ...ALWAYS
1610 *---------------------------------
1620 .6 CMP #$58 IN RANGE FOR VTAB?
1630 BCS .7 NO
1640 AND #$1F MASK VALUE
1650 STA MON.CV YES
1660 JSR MON.VTAB
1670 JMP .4
1680 *---------------------------------
1690 .7 EOR #$60 CHECK FOR TOKENS
1700 CMP #7 $60 THROUGH $66
1710 BCS .4 NOT TOKEN, SO IGNORE
1720 ASL MAKE DUBLE INDEX
1730 TAX
1740 LDA /.4-1 PUT RETURN ON STACK
1750 PHA TO SIMULATE A JSR ADDR,X
1760 LDA #.4-1
1770 PHA
1780 LDA MSGTKNTBL+1,X
1790 PHA
1800 LDA MSGTKNTBL,X
1810 PHA
1820 RTS
1830 *---------------------------------
1840 MSGTKNTBL
1850 .DA MON.HOME-1
1860 .DA MSG.REPEAT-1
1870 .DA LONG.DELAY-1
1880 .DA MON.NORMAL-1
1890 .DA MON.INVERSE-1
1900 .DA MON.CLREOL-1
1910 .DA MON.CLREOP-1
1920 *---------------------------------
1930 MSG.REPEAT
1940 JSR GET.NEXT.CHAR.FROM.MESSAGE
1950 TAX NUMBER OF MULTIPLES
1960 JSR GET.NEXT.CHAR.FROM.MESSAGE
1970 .1 JSR MON.COUT
1980 DEX
1990 BNE .1
2000 RTS
2010 *---------------------------------
2020 LONG.DELAY
2030 LDY #12
2040 .1 JSR MON.WAIT DELAY 167309 CYCLES
2050 DEY
2060 BNE .1
2070 RTS
2080 *---------------------------------
2090 GET.NEXT.CHAR.FROM.MESSAGE
2100 LDY MSG.SCANNER
2110 LDA (MSG.PNTR),Y
2120 INC MSG.SCANNER
2130 BNE .1
2140 INC MSG.PNTR+1
2150 .1 CMP #0
2160 RTS
2170 *---------------------------------
2180 MESSAGE.ADDRESS.TABLE
2190 .DA MSG0
2200 .DA MSG1
2210 .DA MSG2
2220 .DA MSG3
2230 *---------------------------------
2240 MSG0 .HS 60 HOME SCREEN
2250 * CELL 1 -- VOCABULARY CHECK
2260 .HS 64 INVERSE MODE
2270 .HS 6129AD 4A DASHES
2280 .HS 28ADAD 2 DASHES
2290 .HS 28ADAD
2300 .HS 28ADAD 2 DASHES
2310 .HS 28ADAD 2 DASHES
2320 .HS 28ADAD 2 DASHES
2330 .HS 28ADAD 2 DASHES
2340 .HS 286129AD 41 DASHES
2350 .HS 63 NORMAL MODE
2360 .HS 4205 VTAB 3, HTAB 5
2370 .AS -/DEMONSTRATION OF MESSAGE PRINTER/
2380 .HS 440F VTAB 5, HTAB 15
2390 .AS -/S-C SOFTWARE/ [This business closed in 1988]
2400 .HS 450E VTAB 6, HTAB 14
2410 .AS -/P. O. BOX 5537/ [no longer a valid address]
2420 .HS 460B VTAB 7, HTAB 11
2430 .AS -/RICHARDSON, TX 75080/
2440 .HS 4A VTAB 11
2450 .HS 00
2460 *---------------------------------
2470 MSG1 .HS 490166 VTAB 10, HTAB 1, CLR EOP
2480 .AS -/SELECT ONE: /
2490 .HS 00
2500 *---------------------------------
2510 MSG2 .HS 570165 VTAB 24, HTAB 1, CLR EOL
2520 .HS 64 INVERSE MODE
2530 .AS -/ <SPACE> FOR MENU, <RETURN> FOR MORE /
2540 .HS 6300 NORMAL MODE, EOM
2550 *---------------------------------
2560 MSG3 .HS 87878D
2570 .AS -/***SYNTAX ERROR/
2580 .HS 8D00 CARRIAGE RETURN AND END OF MESSAGE
|
Since purchasing the Paymar adapter, I have spent a lot of time adapting software to effectively use it! The program given here will adapt the version 4.0 of Bob Sander-Cederlof's assembler to allow lower-case comments.
The two patches at lines 1340 and 1390 have to be entered, and the body of the patch loaded at $300. Once installed, typing a control-A will toggle the shift-lock; control-S will perform a single-character upper-case shift; control-K, -L, and -O give access to the characters normally missing from the Apple keyboard.
Only comments can be entered in lower-case. Further modification to the assembler would be required to allow commands, labels, and opcodes to be entered in lower- or mixed-case.
1000 *---------------------------------
1010 * Lower case conversion for
1020 * S-C ASSEMBLER II Version 4.0
1030 * Copyright 1980 by S-C SOFTWARE
1040 * Complete with 126 ASCII characters
1050 *---------------------------------
1060 * The CTRL-A and CTRL-S keys are used similar to
1070 * shift and lock keys on a standard typewriter.
1080 *
1090 * CTRL-A is the shift-lock key.
1100 * Each time CTRL-A is pressed the case
1110 * will toggle to the opposite mode.
1120 *
1130 * CTRL-S makes the following character
1140 * enter in upper-case.
1150 *---------------------------------
1160 * REMEMBER!
1170 * All commands and mnemonic entries
1180 * must be in UPPER case!
1190 * Use lower case only for comments!
1200 *---------------------------------
1210 CTRLA .EQ $81 SHIFT LOCK
1220 CTRLK .EQ $8B [ or {
1230 CTRLL .EQ $8C \ or |
1240 CTRLO .EQ $8F _ or rubout
1250 CTRLS .EQ $93 SHIFT
1260 *---------------------------------
1270 * Remember:
1280 * shift M yields ] or }
1290 * shift N yields ^ or ~
1300 * shift P yields @ or `
1310 RDKEY .EQ $FD0C
1320 *---------------------------------
1330 .OR $1380
1340 .TF LC.PATCH1
1350 JSR LC
1360 *---------------------------------
1370 .OR $139A
1380 .TF LC.PATCH2
1390 AND #$FF
1400 *---------------------------------
1410 .OR $300
1420 * CAUTION: Do not assemble your programs into
1430 * $0300 up. You will destroy this routine!!!
1440 LC JSR RDKEY
1450 CMP #CTRLA
1460 BEQ LOCK
1470 CMP #CTRLS
1480 BNE CHECK
1490 SHIFT LDA #0
1500 STA LCKFLG
1510 SHIFT1 LDA #0
1520 STA CASE
1530 BEQ LC ...ALWAYS
1540 LOCK LDA LCKFLG
1550 EOR #1
1560 STA LCKFLG
1570 BNE SHIFT1
1580 LDA #$20
1590 STA CASE
1600 BNE LC ...ALWAYS
1610 CHECK CMP #CTRLK
1620 BEQ SPEC
1630 CMP #CTRLL
1640 BEQ SPEC
1650 CMP #CTRLO
1660 BNE CONV
1670 SPEC ORA #$50
1680 CONV CMP #$C0
1690 BCC RETURN
1700 ORA CASE
1710 RETURN PHA
1720 LDA LCKFLG
1730 BNE OUT
1740 LDA #$20
1750 STA CASE
1760 OUT PLA
1770 RTS
1780 LCKFLG .DA #0
1790 CASE .DA #$20
1800 *---------------------------------
1810 * Written by Bob Matzinger
1820 * September 6, 1980
1830 *---------------------------------
|
INTERFACE, the newsletter of Rockwell International (P. O. Box 3669, RC 55, Anaheim, CA 92803), Issue No. 2, is the source for the following information. It should be noted by all Apple owners working in assembly language, because it could cause an almost unfindable bug!
There is an error in the JUMP INDIRECT instruction of ALL 6500 family CPU chips, no matter where they were made. This means the error is present in ALL APPLES. This fatal error occurs only when the low byte of the indirect pointer location happens to be $FF, as in JMP ($08FF). Normally, the processor should fetch the low-order address byte from location $08FF, increment the program counter to $0900, and then fetch the high-order address byte from $0900. Instead, the high-order byte of the program counter never gets incremented! The high-order address byte gets loaded from $0800 instead of $0900! For this reason, your program should NEVER include an instruction of the type JMP ($xxFF).
Try this example to satisfy yourself that you understand the problem: insert the following data from the monitor.
*800:09 the actual hi-byte used
*810:6C FF 08 this is JMP ($08FF) [indirect through the address at $08FF]
*8FF:50 0A the lo-byte and the expected hi-byte jump address
*A50:00 BRK instruction we SHOULD reach
*950:00 BRK instruction we DO reach!
|
Execute the instruction at $0810 by typing 810G. If the JMP indirect worked correctly, it would branch to location $0A50 and execute the BRK instruction there. However, since the JMP indirect instruction has this serious flaw, it will actually branch to the BRK instruction at $0950!
Since it is very difficult to predict the final address of all pointers in a large assembly language program, unless they are all grouped in a block at the beginning of the program, I suggest that you take special measures to protect yourself against this hardware problem. (One measure, of course, was suggested in that sentence.) My favorite method is to avoid using the JMP indirect instruction. It takes too long to set it up in most cases anyway. I prefer to push the branch address (less one) onto the stack, and RTS to effect the branch. This allows me to create the effect of an indexed JMP. For example, suppose a command character is being decoded. I process it into a value in the A-register between 0 and N-1 (for N commands), and do the following:
ASL Double to create index
TAX for address table
LDA JUMP.TABLE+1,X High order byte
PHA of branch address
LDA JUMP.TABLE,X Low order byte
PHA of branch address
RTS
|
The jump table looks like this:
JUMP.TABLE
.DA COMMANDA-1 The "-1" is
.DA COMMANDB-1 on each line
.DA COMMANDC-1 because the RTS
.DA COMMANDD-1 adds one before
et cetera branching.
|
This trick was described by Steve Wozniak in an article in BYTE magazine back in 1977 or 1978. It is also used by him in the Apple monitor code, and in SWEET-16. In both of these cases, he has arranged all the command processors to be in the same page, so that the high order byte of the address can be loaded into the A-register with a load-A-immediate, and the jump table can be only one-byte-per-command. See your Apple ROMs at locations $FFBE-FFCB (jump table at $FFE3-FFF9) and in SWEET-16 at $F69E, F6A0, F684-F6B8 (jump table at $F6E3-F702).
You can extend this idea of an indexed JMP instruction into a simulated indexed JSR instruction. All you have to do is first push onto the stack the return address (less one), and then the branch address (less one). I use this trick in the Message.Printer program described elsewhere in this issue.