Intel .hex and Motorola S-record files are similar standard file formats which can be very useful, but they do have a couple of limitations. For example, consider an 8192 byte trig table that could be loaded anywhere. Generating this file in one these standard formats means you must specify a load address. Changing the load address generally means generating a new file (assuming you have the program that generated it in the first place!), since editing it manually is very tedious. (Each line, typically containing 16 or 32 bytes of data, has both an address and a checksum.) Another limitation is due to their line-and-field-oriented nature. Addresses widths are fixed, due to the field widths being fixed and you must use an exteneded version of these formats for larger address sizes.
Here is an alternate format with the following features:
The characters used are:
The "number" (variable) is initialized to 0. This (along with the "reset number to zero") rules above mean that you can omit leading zeros. (Or omit all digits if the number is zero, i.e. 1,,2, is equivalent to 1,0,2, though your humble author prefers to use the former form only to indicate a "don't care" value rather than a zero value.)
Note that you can have an (or more than one) embedded [ or ] in a comment, by manually balancing the brackets at the beginning or end of the comment. For example:
[ a comment with a [ embedded ]] [[ a comment with a ] embedded ]
To use it, as with the bootstrap code, use the X register to specify which zero page locations to use. You may initialize the (16-bit) address variable to a default value if you wish, but none of the zero page locations is required to be initialized. Again the first 6 bytes and the data after the RTS are relocation info that can be omitted. As with the bootstrap code, the "get" routine is an input routine that should return one character in the accumulator. It is called in only one place, so it could even be patched manually if necessary.
Here is the object code in "bootstrap encoded" form:
<h<|,h=x9)<xx(lx<xxh|m=y|m=xxhx98-l(=yx8l(98,x<x<xx99)l(m,xx|(i= x9(||(8|lm=x|y<xm,<x|(x-m,<ylhx)x9<8|(9xlm=x|m<xlm=y|m<ylhy989,h x99x|h9|8y,xx9,9l(=l(9xxx9i(|hil(y9}lx<|=l=x-l=yxx|(h)<m=x8h|hh| ,8=y<i<x<x<x<x!
Here is the source and object code:
[ 10 @ 'BPL' ]
[ 20 @ 'JSR' ]
[ 30 @ 'BMI' ]
[ 60 @ 'RTS' ]
[ 90 @ 'BCC' ]
[ A0 @ 'LDY #' ]
[ B0 @ 'BCS' ]
[ D0 @ 'BNE' ]
[ F0 @ 'BEQ' ]
[ 81 @ 'STA ,X)' ]
[ 94 @ 'STY ,X' ]
[ 15 @ 'ORA ,X' ]
[ 95 @ 'STA ,X' ]
[ B5 @ 'LDA ,X' ]
[ 16 @ 'ASL ,X' ]
[ 36 @ 'ROL ,X' ]
[ F6 @ 'INC ,X' ]
[ 18 @ 'CLC' ]
[ 88 @ 'DEY' ]
[ 98 @ 'TYA' ]
[ C8 @ 'INY' ]
[ D8 @ 'CLD' ]
[ 09 @ 'ORA #' ]
[ 29 @ 'AND #' ]
[ 49 @ 'EOR #' ]
[ 69 @ 'ADC #' ]
[ C9 @ 'CMP #' ]
[ ]
[ 0000 @ 'get' ]
[ ]
[ 00 @ 'ADDRESS' ]
[ 01 @ 'ADDRESS+1' ]
[ 02 @ 'NUMBERL' ]
[ 03 @ 'NUMBERH' ]
[ ]
[ 1000 @ ]
[ ]
(10,04, )[ 'BPL',00*, ]
(30,02, )[ 'BMI',02*, ]
(5B,00, )[ (DW) 005B,, (RELDATA-PROGRAM) ]
[ 00=>02=> ]
[ (PROGRAM) ]
(D8, )[ 'CLD', ]
(A0,00, )[ 'LDY #',00, ]
(98, )[ 00* 'TYA', ]
(95,03, )[ 'STA ,X','NUMBERH', ]
(95,02, )[ 02* 'STA ,X','NUMBERL', ]
(98, )[ 04* 'TYA', ]
(C9,5D, )[ 06* 'CMP #',5D, ]
(F0,03, )[ 'BEQ',0A*, ]
(C8, )[ 08* 'INY', ]
(F0,4A, )[ 'BEQ',14*, ]
(20,00,00, )[ 0A=> 'JSR','get',, ]
(C9,5B, )[ 'CMP #',5B, ]
(F0,F6, )[ 'BEQ',08=< ]
(88, )[ 'DEY', ]
(D0,EF, )[ 'BNE',06=< ]
(C9,2C, )[ 'CMP #',2C, ]
(D0,0C, )[ 'BNE',0C*, ]
(B5,02, )[ 'LDA ,X','NUMBERL', ]
(81,00, )[ 'STA ,X)','ADDRESS', ]
(F6,00, )[ 'INC ,X','ADDRESS', ]
(D0,DD, )[ 'BNE',00=< ]
(F6,01, )[ 'INC ,X','ADDRESS+1', ]
(B0,D9, )[ 'BCS',00=< ]
(C9,40, )[ 0C=> 'CMP #',40, ]
(D0,0A, )[ 'BNE',0E*, ]
(B5,02, )[ 'LDA ,X','NUMBERL', ]
(95,00, )[ 'STA ,X','ADDRESS', ]
(B5,03, )[ 'LDA ,X','NUMBERH', ]
(95,01, )[ 'STA ,X','ADDRESS+1', ]
(B0,CB, )[ 'BCS',00=< ]
(49,30, )[ 0E=> 'EOR #',30, ]
(C9,0A, )[ 'CMP #',0A, ]
(90,0E, )[ 'BCC',10*, ]
(09,20, )[ 'ORA #',20, ]
(C9,61, )[ 'CMP #',61, ]
(F0,16, )[ 'BEQ',16*, ]
(69,88, )[ 'ADC #',88, ]
(C9,FA, )[ 'CMP #',FA, ]
(90,BE, )[ 'BCC',04=< ]
(29,0F, )[ 'AND #',0F, ]
(A0,04, )[ 10=> 'LDY #',04, ]
(16,02, )[ 12* 'ASL ,X','NUMBERL', ]
(36,03, )[ 'ROL ,X','NUMBERH', ]
(88, )[ 'DEY', ]
(D0,F9, )[ 'BNE',12=< ]
(15,02, )[ 'ORA ,X','NUMBERL', ]
(18, )[ 'CLC', ]
(90,AC, )[ 'BCC',02=< ]
[ 14=> ]
(60, )[ 16=> 'RTS', ]
[ (RELDATA) ]
(03,11,00,00,00,)[ (DB) 03, (DW) 0011,, (DW) 'get',, ]
(00, )[ (DB) 00, ]
(Q )[ _ ]
The format described above is (intentionally) not specific to the 6502 family. Your humble author has developed an "extended" version of this format which adds three additional characters to the original format, which generate the 8-bit signed displacement for 6502 branches. (Though this feature was designed specifically for the 6502, it could be use with other microprocessors that use the same 8-bit signed displacements.) The three additional characters are:
Here is an example of how these new characters can be used:
[ calculate checksum ] A9,00, [ LDA #0 ] 0* 18, [ .1 CLC ] 61,00, [ ADC (0,X) ] F6,00, [ INC 0,X ] D0,2*, [ BNE .2 ] F6,01, [ INC 1,X ] 2> 88, [ .2 DEY ] D0,0< [ BNE .1 ] 60, [ RTS ] Q
The extended format routine is used in the much the same way as before. The "labels" are all stored on a single page, so you must initialize the high byte of the label variable (i.e. 5,X) to indicate which page to store labels. Note that labels can be reused. After the 2> the label 2 is available again. For additional backward branches you can use 0< as many times as you like. Otherwise, you can reuse the label 0 after the last 0<. Note that you must use separate labels for multiple forward branches to the same destination. Also note that the routine does not check whether the branch is too long. Use only even labels.
Here is the object code in "bootstrap encoded" form:
<h<|,h=x|h<xx(hy<x|m=y|m=xlx<xxhx98-l(=yx8l(9,,x<x<xx99)l(m,xx|( i=ll=xx9<8|(8x|l<xlm=y|m<ylhx(x9(|l(<m|l<|x9)xl((mx9)ll(<=x9(l|( =h8hly<|l-<xhx(hxh|y<xm,<x|(hhm,<ylhll89,hx99x|h9|8y,xx9,9l(=l(9 xxx9i(|hhx(y9}lx<|=l=x-l=yxx|(h)<m=x8h|h}l,8lm<x|y<|m,<|lm<y|y<| lh|}ly<||m=xm,<|ly<||m=y8hlm<xl-=x|y=x(hlhii=y=h<x<x<x<x!
Here is the source and object code:
[ 10 @ 'BPL' ]
[ 20 @ 'JSR' ]
[ 30 @ 'BMI' ]
[ 60 @ 'RTS' ]
[ 90 @ 'BCC' ]
[ A0 @ 'LDY #' ]
[ B0 @ 'BCS' ]
[ D0 @ 'BNE' ]
[ F0 @ 'BEQ' ]
[ 81 @ 'STA ,X)' ]
[ A1 @ 'LDA ,X)' ]
[ 94 @ 'STY ,X' ]
[ B4 @ 'LDY ,X' ]
[ 15 @ 'ORA ,X' ]
[ 95 @ 'STA ,X' ]
[ B5 @ 'LDA ,X' ]
[ F5 @ 'SBC ,X' ]
[ 16 @ 'ASL ,X' ]
[ 36 @ 'ROL ,X' ]
[ F6 @ 'INC ,X' ]
[ 18 @ 'CLC' ]
[ 38 @ 'SEC' ]
[ 88 @ 'DEY' ]
[ 98 @ 'TYA' ]
[ A8 @ 'TAY' ]
[ C8 @ 'INY' ]
[ D8 @ 'CLD' ]
[ 09 @ 'ORA #' ]
[ 29 @ 'AND #' ]
[ 49 @ 'EOR #' ]
[ 69 @ 'ADC #' ]
[ A9 @ 'LDA #' ]
[ C9 @ 'CMP #' ]
[ ]
[ 0000 @ 'get' ]
[ ]
[ 00 @ 'ADDRESS' ]
[ 01 @ 'ADDRESS+1' ]
[ 02 @ 'NUMBER' ]
[ 03 @ 'NUMBER+1' ]
[ 04 @ 'LABELS' ]
[ ]
[ 1000 @ ]
[ ]
(10,04, )[ 'BPL',00*, ]
(30,02, )[ 'BMI',02*, ]
(90,00, )[ (DW) 0090,, (RELDATA-PROGRAM) ]
[ 00=>02=> ]
[ (PROGRAM) ]
(D8, )[ 'CLD', ]
(A9,00, )[ 00* 'LDA #',00, ]
(95,03, )[ 'STA ,X','NUMBER+1', ]
(95,02, )[ 02* 'STA ,X','NUMBER', ]
(A0,00, )[ 04* 'LDY #',00, ]
(98, )[ 'TYA', ]
(C9,5D, )[ 06* 'CMP #',5D, ]
(F0,03, )[ 'BEQ',0A*, ]
(C8, )[ 08* 'INY', ]
(F0,5E, )[ 'BEQ',18*, ]
(20,00,00, )[ 0A=> 'JSR','get',, ]
(C9,5B, )[ 'CMP #',5B, ]
(F0,F6, )[ 'BEQ',08=< ]
(88, )[ 'DEY', ]
(D0,EF, )[ 'BNE',06=< ]
(B4,02, )[ 'LDY ,X','NUMBER', ]
(C9,40, )[ 'CMP #',40, ]
(D0,08, )[ 'BNE',0C*, ]
(94,00, )[ 'STY ,X','ADDRESS', ]
(B5,03, )[ 'LDA ,X','NUMBER+1', ]
(95,01, )[ 'STA ,X','ADDRESS+1', ]
(B0,D8, )[ 'BCS',00=< ]
(C9,2C, )[ 0C=> 'CMP #',2C, ]
(F0,15, )[ 'BEQ',0E*, ]
(94,04, )[ 'STY ,X','LABELS', ]
(C9,2A, )[ 'CMP #',2A, ]
(F0,3D, )[ 'BEQ',1C*, ]
(C9,3E, )[ 'CMP #',3E, ]
(F0,45, )[ 'BEQ',1E*, ]
(C9,3C, )[ 'CMP #',3C, ]
(D0,12, )[ 'BNE',12*, ]
(18, )[ 'CLC', ]
(A1,04, )[ 'LDA ,X)','LABELS', ]
(F5,00, )[ 'SBC ,X','ADDRESS', ]
(A8, )[ 'TAY', ]
(38, )[ 'SEC', ]
(98, )[ 0E=> 'TYA', ]
(81,00, )[ 'STA ,X)','ADDRESS', ]
(F6,00, )[ 'INC ,X','ADDRESS', ]
(D0,B8, )[ 'BNE',00=< ]
(F6,01, )[ 'INC ,X','ADDRESS+1', ]
(B0,B4, )[ 10* 'BCS',00=< ]
(49,30, )[ 12=> 'EOR #',30, ]
(C9,0A, )[ 'CMP #',0A, ]
(90,0E, )[ 'BCC',14*, ]
(09,20, )[ 'ORA #',20, ]
(C9,61, )[ 'CMP #',61, ]
(F0,16, )[ 'BEQ',1A*, ]
(69,88, )[ 'ADC #',88, ]
(C9,FA, )[ 'CMP #',FA, ]
(90,A8, )[ 'BCC',04=< ]
(29,0F, )[ 'AND #',0F, ]
(A0,04, )[ 14=> 'LDY #',04, ]
(16,02, )[ 16* 'ASL ,X','NUMBER', ]
(36,03, )[ 'ROL ,X','NUMBER+1', ]
(88, )[ 'DEY', ]
(D0,F9, )[ 'BNE',16=< ]
(15,02, )[ 'ORA ,X','NUMBER', ]
(18, )[ 'CLC', ]
(90,96, )[ 'BCC',02=< ]
[ 18=> ]
(60, )[ 1A=> 'RTS', ]
(B5,00, )[ 1C=> 'LDA ,X','ADDRESS', ]
(81,04, )[ 'STA ,X)','LABELS', ]
(F6,04, )[ 'INC ,X','LABELS', ]
(B5,01, )[ 'LDA ,X','ADDRESS+1', ]
(81,04, )[ 'STA ,X)','LABELS', ]
(B0,85, )[ 'BCS',00=< ]
(A1,04, )[ 1E=> 'LDA ,X)','LABELS', ]
(95,02, )[ 'STA ,X','NUMBER', ]
(F6,04, )[ 'INC ,X','LABELS', ]
(A1,04, )[ 'LDA ,X)','LABELS', ]
(95,03, )[ 'STA ,X','NUMBER+1', ]
(18, )[ 'CLC', ]
(B5,00, )[ 'LDA ,X','ADDRESS', ]
(F5,02, )[ 'SBC ,X','NUMBER', ]
(81,02, )[ 'STA ,X)','NUMBER', ]
(38, )[ 'SEC', ]
(B0,BB, )[ 'BCS',10=< ]
[ (RELDATA) ]
(03,12,00,00,00,)[ (DB) 03, (DW) 0012,, (DW) 'get',, ]
(00, )[ (DB) 00, ]
(Q )[ _ ]
Your humble author developed a version of these formats that had checksums, where you control what (addresses, data, or both) and how much was accumulated in the checksum (e.g. you could have one checksum for one section, another checksum for another section, and no checksum for a third section, and in addition, you could comment out checksums at your convenience.) However, the only time checksums have ever been anything other than completely superfluous with the standard hex file formats for your humble author was for file type checking, i.e. on those infrequent occasions when mistakenly loading a "raw binary" file instead of a (text) "hex" file. It seems to your humble author that merely using a "signature" (e.g. [Hex File]) at the beginning of the file would fulfill this purpose, and that the checksum feature is an unnecessary complication. Your humble author is not trying to be overly critical of the standard formats, but is trying to learn from them (what they do well and where there is room for improvement).