izaya/collapseos
1
0
Fork 0
mirror of https://github.com/hsoft/collapseos.git synced 2024-11-15 04:48:06 +11:00
Code Issues Releases Wiki Activity
4da8b68623
collapseos/blk.fs

2959 lines
93 KiB
Forth
Raw Normal View History

Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 000 )
MASTER INDEX
005 Z80 assembler 030 8086 assembler
050 AVR assembler 70-99 unused
100 Block editor 120 Visual Editor
Add a new "Remote Shell" application This allows a COS shell to remotely control another shell from a serial line. It's very crude now, but the plan is to add tools much like the POSIX tools, but for COS.
2021-01-04 06:01:34 +11:00
150 Remote Shell
sms: generate TMR SEGA signature in Collapse OS itself Having the signature generation code in /tools prevents self-hosting on the SMS.
2020-11-26 13:42:07 +11:00
160 AVR SPI programmer 165 Sega ROM signer
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
170-259 unused 260 Cross compilation
280 Z80 boot code 350 Core words
400 AT28 EEPROM driver 401 Grid subsystem
410 PS/2 keyboard subsystem 418 Z80 SPI Relay driver
420 SD Card subsystem 440 8086 boot code
470 Z80 TMS9918 driver
480-519 unused 520 Fonts
( ----- 005 )
( Z80 Assembler
006 Variables & consts
007 Utils 008 OP1
010 OP1r 012 OP1d
013 OP1rr 015 OP2
016 OP2i 017 OP2ri
018 OP2br 019 OProt
020 OP2r 021 OP2d
022 OP3di 023 OP3i
024 Specials 025 Flow
028 Macros )
1 23 LOADR+
( ----- 006 )
CREATE ORG 0 ,
CREATE BIN( 0 ,
VARIABLE L1 VARIABLE L2 VARIABLE L3 VARIABLE L4
: A 7 ; : B 0 ; : C 1 ; : D 2 ;
: E 3 ; : H 4 ; : L 5 ; : (HL) 6 ;
: BC 0 ; : DE 1 ; : HL 2 ; : AF 3 ; : SP AF ;
: CNZ 0 ; : CZ 1 ; : CNC 2 ; : CC 3 ;
: CPO 4 ; : CPE 5 ; : CP 6 ; : CM 7 ;
( ----- 007 )
: PC H@ ORG @ - BIN( @ + ;
: <<3 3 LSHIFT ; : <<4 4 LSHIFT ;
( As a general rule, IX and IY are equivalent to spitting an
extra 0xdd / 0xfd and then spit the equivalent of HL )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: IX 0xdd C, HL ; : IY 0xfd C, HL ;
: _ix+- 0xff AND 0xdd C, (HL) ;
: _iy+- 0xff AND 0xfd C, (HL) ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: IX+ _ix+- ; : IX- 0 -^ _ix+- ;
: IY+ _iy+- ; : IY- 0 -^ _iy+- ;
( ----- 008 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OP1 CREATE C, DOES> C@ C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xf3 OP1 DI, 0xfb OP1 EI,
0xeb OP1 EXDEHL, 0xd9 OP1 EXX,
0x08 OP1 EXAFAF', 0xe3 OP1 EX(SP)HL,
0x76 OP1 HALT, 0xe9 OP1 JP(HL),
0x12 OP1 LD(DE)A, 0x1a OP1 LDA(DE),
0x02 OP1 LD(BC)A, 0x0a OP1 LDA(BC),
0x00 OP1 NOP, 0xc9 OP1 RET,
0x17 OP1 RLA, 0x07 OP1 RLCA,
0x1f OP1 RRA, 0x0f OP1 RRCA,
0x37 OP1 SCF,
( ----- 009 )
( Relative jumps are a bit special. They're supposed to take
an argument, but they don't take it so they can work with
the label system. Therefore, relative jumps are an OP1 but
when you use them, you're expected to write the offset
afterwards yourself. )
0x18 OP1 JR, 0x10 OP1 DJNZ,
0x38 OP1 JRC, 0x30 OP1 JRNC,
0x28 OP1 JRZ, 0x20 OP1 JRNZ,
( ----- 010 )
( r -- )
: OP1r
CREATE C,
DOES>
C@ ( r op )
SWAP ( op r )
<<3 ( op r<<3 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0x04 OP1r INCr, 0x05 OP1r DECr,
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: INC(IXY+), INCr, C, ;
: DEC(IXY+), DECr, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( also works for c )
0xc0 OP1r RETc,
( ----- 011 )
: OP1r0 ( r -- )
CREATE C, DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@ ( r op ) OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x80 OP1r0 ADDr, 0x88 OP1r0 ADCr,
0xa0 OP1r0 ANDr, 0xb8 OP1r0 CPr,
0xb0 OP1r0 ORr, 0x90 OP1r0 SUBr,
0x98 OP1r0 SBCr, 0xa8 OP1r0 XORr,
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: CP(IXY+), CPr, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 012 )
: OP1d
CREATE C,
DOES>
C@ ( d op )
SWAP ( op d )
<<4 ( op d<<4 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0xc5 OP1d PUSH, 0xc1 OP1d POP,
0x03 OP1d INCd, 0x0b OP1d DECd,
0x09 OP1d ADDHLd,
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: ADDIXd, 0xdd C, ADDHLd, ; : ADDIXIX, HL ADDIXd, ;
: ADDIYd, 0xfd C, ADDHLd, ; : ADDIYIY, HL ADDIYd, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 013 )
: _1rr
C@ ( rd rr op )
ROT ( rr op rd )
<<3 ( rr op rd<<3 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
( rd rr )
: OP1rr
CREATE C,
DOES>
_1rr
;
0x40 OP1rr LDrr,
( ----- 014 )
( ixy+- HL rd )
: LDIXYr,
( dd/fd has already been spit )
LDrr, ( ixy+- )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
( rd ixy+- HL )
: LDrIXY,
ROT ( ixy+- HL rd )
SWAP ( ixy+- rd HL )
LDIXYr,
;
( ----- 015 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OP2 CREATE , DOES> @ |M C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xeda1 OP2 CPI, 0xedb1 OP2 CPIR,
0xeda9 OP2 CPD, 0xedb9 OP2 CPDR,
0xed46 OP2 IM0, 0xed56 OP2 IM1,
0xed5e OP2 IM2,
0xeda0 OP2 LDI, 0xedb0 OP2 LDIR,
0xeda8 OP2 LDD, 0xedb8 OP2 LDDR,
0xed44 OP2 NEG,
0xed4d OP2 RETI, 0xed45 OP2 RETN,
( ----- 016 )
: OP2i ( i -- )
CREATE C,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@ C, C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0xd3 OP2i OUTiA,
0xdb OP2i INAi,
0xc6 OP2i ADDi,
0xe6 OP2i ANDi,
0xf6 OP2i ORi,
0xd6 OP2i SUBi,
0xee OP2i XORi,
0xfe OP2i CPi,
( ----- 017 )
: OP2ri ( r i -- )
CREATE C,
DOES>
C@ ( r i op )
ROT ( i op r )
<<3 ( i op r<<3 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C, C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0x06 OP2ri LDri,
( ----- 018 )
( b r -- )
: OP2br
CREATE C,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0xcb C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
C@ ( b r op )
ROT ( r op b )
<<3 ( r op b<<3 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0xc0 OP2br SET,
0x80 OP2br RES,
0x40 OP2br BIT,
( ----- 019 )
( bitwise rotation ops have a similar sig )
: OProt ( r -- )
CREATE C,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0xcb C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
C@ ( r op )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0x10 OProt RL,
0x00 OProt RLC,
0x18 OProt RR,
0x08 OProt RRC,
0x20 OProt SLA,
0x38 OProt SRL,
( ----- 020 )
( cell contains both bytes. MSB is spit as-is, LSB is ORed
with r )
( r -- )
: OP2r
CREATE ,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
@ |M ( r lsb msb )
C, ( r lsb )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
SWAP <<3 ( lsb r<<3 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0xed41 OP2r OUT(C)r,
0xed40 OP2r INr(C),
( ----- 021 )
: OP2d ( d -- )
CREATE C,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0xed C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
C@ SWAP ( op d )
<<4 ( op d<< 4 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0x4a OP2d ADCHLd,
0x42 OP2d SBCHLd,
( ----- 022 )
( d i -- )
: OP3di
CREATE C,
DOES>
C@ ( d n op )
ROT ( n op d )
<<4 ( n op d<<4 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR C, ,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0x01 OP3di LDdi,
( ----- 023 )
( i -- )
: OP3i
CREATE C,
DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@ C, ,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
0xcd OP3i CALL,
0xc3 OP3i JP,
0x22 OP3i LD(i)HL, 0x2a OP3i LDHL(i),
0x32 OP3i LD(i)A, 0x3a OP3i LDA(i),
( ----- 024 )
: LDd(i), ( d i -- )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0xed C,
SWAP <<4 0x4b OR C, ,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
: LD(i)d, ( i d -- )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0xed C,
<<4 0x43 OR C, ,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: RST, 0xc7 OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: JP(IX), IX DROP JP(HL), ;
: JP(IY), IY DROP JP(HL), ;
( ----- 025 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: JPc, SWAP <<3 0xc2 OR C, , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: BCALL, BIN( @ + CALL, ;
: BJP, BIN( @ + JP, ;
: BJPc, BIN( @ + JPc, ;
CREATE lblchkPS 0 ,
: chkPS, lblchkPS @ CALL, ; ( chkPS, B305 )
CREATE lblnext 0 , ( stable ABI until set in B300 )
: JPNEXT, lblnext @ ?DUP IF JP, ELSE 0x1a BJP, THEN ;
: CODE ( same as CREATE, but with native word )
(entry) 0 C, ( 0 == native ) ;
: ;CODE JPNEXT, ;
( ----- 026 )
( Place BEGIN, where you want to jump back and AGAIN after
a relative jump operator. Just like BSET and BWR. )
z80 asm: document the 0x8000 limitation better When I get bitten by it myself, it's because it's clearly underdocumented!
2020-12-15 06:40:22 +11:00
: BEGIN,
PC DUP 0x8000 AND IF ABORT" PC must be < 0x8000" THEN ;
: BSET BEGIN, SWAP ! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( same as BSET, but we need to write a placeholder )
z80 asm: document the 0x8000 limitation better When I get bitten by it myself, it's because it's clearly underdocumented!
2020-12-15 06:40:22 +11:00
: FJR, BEGIN, 0 C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: IFZ, JRNZ, FJR, ;
: IFNZ, JRZ, FJR, ;
: IFC, JRNC, FJR, ;
: IFNC, JRC, FJR, ;
: THEN,
z80a: add ELSE,
2020-12-08 15:06:12 +11:00
DUP PC ( l l pc ) -^ 1- ( l off )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( warning: l is a PC offset, not a mem addr! )
z80a: add ELSE,
2020-12-08 15:06:12 +11:00
SWAP ORG @ + BIN( @ - ( off addr ) C! ;
: ELSE, JR, FJR, SWAP THEN, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 027 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: FWR BSET 0 C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: FSET @ THEN, ;
: BREAK, FJR, 0x8000 OR ;
: BREAK?, DUP 0x8000 AND IF
0x7fff AND 1 ALLOT THEN, -1 ALLOT
THEN ;
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: AGAIN, BREAK?, PC - 1- C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: BWR @ AGAIN, ;
( ----- 028 )
( Macros )
( clear carry + SBC )
: SUBHLd, A ORr, SBCHLd, ;
: PUSH0, DE 0 LDdi, DE PUSH, ;
: PUSH1, DE 1 LDdi, DE PUSH, ;
: PUSHZ, DE 0 LDdi, IFZ, DE INCd, THEN, DE PUSH, ;
: PUSHA, D 0 LDri, E A LDrr, DE PUSH, ;
: HLZ, A H LDrr, L ORr, ;
: DEZ, A D LDrr, E ORr, ;
: LDDE(HL), E (HL) LDrr, HL INCd, D (HL) LDrr, ;
: OUTHL, DUP A H LDrr, OUTiA, A L LDrr, OUTiA, ;
: OUTDE, DUP A D LDrr, OUTiA, A E LDrr, OUTiA, ;
( ----- 030 )
( 8086 assembler. See doc/asm.txt )
1 13 LOADR+
( ----- 031 )
VARIABLE ORG
CREATE BIN( 0 , : BIN(+ BIN( @ + ;
VARIABLE L1 VARIABLE L2 VARIABLE L3 VARIABLE L4
: AL 0 ; : CL 1 ; : DL 2 ; : BL 3 ;
: AH 4 ; : CH 5 ; : DH 6 ; : BH 7 ;
: AX 0 ; : CX 1 ; : DX 2 ; : BX 3 ;
: SP 4 ; : BP 5 ; : SI 6 ; : DI 7 ;
: ES 0 ; : CS 1 ; : SS 2 ; : DS 3 ;
: [BX+SI] 0 ; : [BX+DI] 1 ; : [BP+SI] 2 ; : [BP+DI] 3 ;
: [SI] 4 ; : [DI] 5 ; : [BP] 6 ; : [BX] 7 ;
: <<3 3 LSHIFT ;
( ----- 032 )
: PC H@ ORG @ - BIN( @ + ;
( ----- 033 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OP1 CREATE C, DOES> C@ C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xc3 OP1 RET, 0xfa OP1 CLI, 0xfb OP1 STI,
0xf4 OP1 HLT, 0xfc OP1 CLD, 0xfd OP1 STD,
0x90 OP1 NOP, 0x98 OP1 CBW,
0xf3 OP1 REPZ, 0xf2 OP1 REPNZ, 0xac OP1 LODSB,
0xad OP1 LODSW, 0xa6 OP1 CMPSB, 0xa7 OP1 CMPSW,
0xa4 OP1 MOVSB, 0xa5 OP1 MOVSW, 0xae OP1 SCASB,
0xaf OP1 SCASW, 0xaa OP1 STOSB, 0xab OP1 STOSW,
( no argument, jumps with relative addrs are special )
0xeb OP1 JMPs, 0xe9 OP1 JMPn, 0x74 OP1 JZ,
0x75 OP1 JNZ, 0x72 OP1 JC, 0x73 OP1 JNC,
0xe8 OP1 CALL,
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OP1r CREATE C, DOES> C@ + C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x40 OP1r INCx, 0x48 OP1r DECx,
0x58 OP1r POPx, 0x50 OP1r PUSHx,
( ----- 034 )
: OPr0 ( reg op ) CREATE C, C, DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@+ C, C@ <<3 OR 0xc0 OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0 0xd0 OPr0 ROLr1, 0 0xd1 OPr0 ROLx1, 4 0xf6 OPr0 MULr,
1 0xd0 OPr0 RORr1, 1 0xd1 OPr0 RORx1, 4 0xf7 OPr0 MULx,
4 0xd0 OPr0 SHLr1, 4 0xd1 OPr0 SHLx1, 6 0xf6 OPr0 DIVr,
5 0xd0 OPr0 SHRr1, 5 0xd1 OPr0 SHRx1, 6 0xf7 OPr0 DIVx,
0 0xd2 OPr0 ROLrCL, 0 0xd3 OPr0 ROLxCL, 1 0xfe OPr0 DECr,
1 0xd2 OPr0 RORrCL, 1 0xd3 OPr0 RORxCL, 0 0xfe OPr0 INCr,
4 0xd2 OPr0 SHLrCL, 4 0xd3 OPr0 SHLxCL,
5 0xd2 OPr0 SHRrCL, 5 0xd3 OPr0 SHRxCL,
( ----- 035 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPrr CREATE C, DOES> C@ C, <<3 OR 0xc0 OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x31 OPrr XORxx, 0x30 OPrr XORrr,
0x88 OPrr MOVrr, 0x89 OPrr MOVxx, 0x28 OPrr SUBrr,
0x29 OPrr SUBxx, 0x08 OPrr ORrr, 0x09 OPrr ORxx,
0x38 OPrr CMPrr, 0x39 OPrr CMPxx, 0x00 OPrr ADDrr,
0x01 OPrr ADDxx, 0x20 OPrr ANDrr, 0x21 OPrr ANDxx,
( ----- 036 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPm ( modrm op ) CREATE C, C, DOES> C@+ C, C@ OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0 0xff OPm INC[w], 0 0xfe OPm INC[b],
0x8 0xff OPm DEC[w], 0x8 0xfe OPm DEC[b],
0x30 0xff OPm PUSH[w], 0 0x8f OPm POP[w],
: OPm+ ( modrm op ) CREATE C, C, DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
( m off ) C@+ C, C@ ROT OR C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x40 0xff OPm+ INC[w]+, 0x40 0xfe OPm+ INC[b]+,
0x48 0xff OPm+ DEC[w]+, 0x48 0xfe OPm+ DEC[b]+,
0x70 0xff OPm+ PUSH[w]+, 0x40 0x8f OPm+ POP[w]+,
( ----- 037 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPrm CREATE C, DOES> C@ C, SWAP 3 LSHIFT OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x8a OPrm MOVr[], 0x8b OPrm MOVx[],
0x3a OPrm CMPr[], 0x3b OPrm CMPx[],
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPmr CREATE C, DOES> C@ C, 3 LSHIFT OR C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x88 OPmr MOV[]r, 0x89 OPmr MOV[]x,
: OPrm+ ( r m off ) CREATE C, DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@ C, ROT 3 LSHIFT ROT OR 0x40 OR C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x8a OPrm+ MOVr[]+, 0x8b OPrm+ MOVx[]+,
0x3a OPrm+ CMPr[]+, 0x3b OPrm+ CMPx[]+,
: OPm+r ( m off r ) CREATE C, DOES>
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C@ C, 3 LSHIFT ROT OR 0x40 OR C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x88 OPm+r MOV[]+r, 0x89 OPm+r MOV[]+x,
( ----- 038 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPi CREATE C, DOES> C@ C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x04 OPi ADDALi, 0x24 OPi ANDALi, 0x2c OPi SUBALi,
0xcd OPi INT,
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPI CREATE C, DOES> C@ C, , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x05 OPI ADDAXI, 0x25 OPI ANDAXI, 0x2d OPI SUBAXI,
( ----- 040 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: MOVri, SWAP 0xb0 OR C, C, ;
: MOVxI, SWAP 0xb8 OR C, , ;
: MOVsx, 0x8e C, SWAP <<3 OR 0xc0 OR C, ;
: MOVrm, 0x8a C, SWAP <<3 0x6 OR C, , ;
: MOVxm, 0x8b C, SWAP <<3 0x6 OR C, , ;
: MOVmr, 0x88 C, <<3 0x6 OR C, , ;
: MOVmx, 0x89 C, <<3 0x6 OR C, , ;
: PUSHs, <<3 0x06 OR C, ; : POPs, <<3 0x07 OR C, ;
: SUBxi, 0x83 C, SWAP 0xe8 OR C, C, ;
: ADDxi, 0x83 C, SWAP 0xc0 OR C, C, ;
: JMPr, 0xff C, 7 AND 0xe0 OR C, ;
: JMPf, ( seg off ) 0xea C, |L C, C, , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 041 )
( Place BEGIN, where you want to jump back and AGAIN after
a relative jump operator. Just like BSET and BWR. )
: BEGIN, PC ;
: BSET PC SWAP ! ;
( same as BSET, but we need to write a placeholder )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: FJR, PC 0 C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: IFZ, JNZ, FJR, ;
: IFNZ, JZ, FJR, ;
: IFC, JNC, FJR, ;
: IFNC, JC, FJR, ;
: THEN,
DUP PC ( l l pc )
-^ 1- ( l off )
( warning: l is a PC offset, not a mem addr! )
SWAP ORG @ + BIN( @ - ( off addr )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 042 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: FWRs BSET 0 C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: FSET @ THEN, ;
Use A! instead of C! in THEN, in z80 and 8086 assemblers
2020-12-08 04:32:32 +11:00
( TODO: add BREAK, )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: RPCs, PC - 1- DUP 128 + 0xff > IF ABORT" PC ovfl" THEN C, ;
: RPCn, PC - 2- , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: AGAIN, ( BREAK?, ) RPCs, ;
( Use RPCx with appropriate JMP/CALL op. Example:
JMPs, 0x42 RPCs, or CALL, 0x1234 RPCn, )
( ----- 043 )
: PUSHZ, CX 0 MOVxI, IFZ, CX INCx, THEN, CX PUSHx, ;
: CODE ( same as CREATE, but with native word )
(entry) 0 ( native ) C, ;
: ;CODE JMPn, 0x1a ( next ) RPCn, ;
VARIABLE lblchkPS
: chkPS, ( sz -- )
CX SWAP 2 * MOVxI, CALL, lblchkPS @ RPCn, ;
( ----- 050 )
1 12 LOADR+
( ----- 051 )
VARIABLE ORG
VARIABLE L1 VARIABLE L2 VARIABLE L3 VARIABLE L4
( We divide by 2 because each PC represents a word. )
: PC H@ ORG @ - 1 RSHIFT ;
( ----- 052 )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: _oor ." arg out of range: " .X SPC> ." PC: " PC .X NL> ABORT ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: _r8c DUP 7 > IF _oor THEN ;
: _r32c DUP 31 > IF _oor THEN ;
: _r16+c _r32c DUP 16 < IF _oor THEN ;
: _r64c DUP 63 > IF _oor THEN ;
: _r256c DUP 255 > IF _oor THEN ;
: _Rdp ( op rd -- op', place Rd ) 4 LSHIFT OR ;
( ----- 053 )
( 0000 000d dddd 0000 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPRd CREATE , DOES> @ SWAP _r32c _Rdp , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0b1001010000000101 OPRd ASR, 0b1001010000000000 OPRd COM,
0b1001010000001010 OPRd DEC, 0b1001010000000011 OPRd INC,
0b1001001000000110 OPRd LAC, 0b1001001000000101 OPRd LAS,
0b1001001000000111 OPRd LAT,
0b1001010000000110 OPRd LSR, 0b1001010000000001 OPRd NEG,
0b1001000000001111 OPRd POP, 0b1001001000001111 OPRd PUSH,
0b1001010000000111 OPRd ROR, 0b1001010000000010 OPRd SWAP,
0b1001001000000100 OPRd XCH,
( ----- 054 )
( 0000 00rd dddd rrrr )
: OPRdRr CREATE C, DOES> C@ ( rd rr op )
OVER _r32c 0x10 AND 3 RSHIFT OR ( rd rr op' )
8 LSHIFT OR 0xff0f AND ( rd op' )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
SWAP _r32c _Rdp , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x1c OPRdRr ADC, 0x0c OPRdRr ADD, 0x20 OPRdRr AND,
0x14 OPRdRr CP, 0x04 OPRdRr CPC, 0x10 OPRdRr CPSE,
0x24 OPRdRr EOR, 0x2c OPRdRr MOV, 0x9c OPRdRr MUL,
0x28 OPRdRr OR, 0x08 OPRdRr SBC, 0x18 OPRdRr SUB,
( 0000 0AAd dddd AAAA )
: OPRdA CREATE C, DOES> C@ ( rd A op )
OVER _r64c 0x30 AND 3 RSHIFT OR ( rd A op' )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
8 LSHIFT OR 0xff0f AND ( rd op' ) SWAP _r32c _Rdp , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xb0 OPRdA IN, 0xb8 OPRdA _ : OUT, SWAP _ ;
( ----- 055 )
( 0000 KKKK dddd KKKK )
: OPRdK CREATE C, DOES> C@ ( rd K op )
OVER _r256c 0xf0 AND 4 RSHIFT OR ( rd K op' )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
ROT _r16+c 4 LSHIFT ROT 0x0f AND OR ( op' rdK ) C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x70 OPRdK ANDI, 0x30 OPRdK CPI, 0xe0 OPRdK LDI,
0x60 OPRdK ORI, 0x40 OPRdK SBCI, 0x60 OPRdK SBR,
0x50 OPRdK SUBI,
( 0000 0000 AAAA Abbb )
: OPAb CREATE C, DOES> C@ ( A b op )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
ROT _r32c 3 LSHIFT ROT _r8c OR C, C, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x98 OPAb CBI, 0x9a OPAb SBI, 0x99 OPAb SBIC,
0x9b OPAb SBIS,
( ----- 056 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: OPNA CREATE , DOES> @ , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x9598 OPNA BREAK, 0x9488 OPNA CLC, 0x94d8 OPNA CLH,
0x94f8 OPNA CLI, 0x94a8 OPNA CLN, 0x94c8 OPNA CLS,
0x94e8 OPNA CLT, 0x94b8 OPNA CLV, 0x9498 OPNA CLZ,
0x9419 OPNA EIJMP, 0x9509 OPNA ICALL, 0x9519 OPNA EICALL,
0x9409 OPNA IJMP, 0x0000 OPNA NOP, 0x9508 OPNA RET,
0x9518 OPNA RETI, 0x9408 OPNA SEC, 0x9458 OPNA SEH,
0x9478 OPNA SEI, 0x9428 OPNA SEN, 0x9448 OPNA SES,
0x9468 OPNA SET, 0x9438 OPNA SEV, 0x9418 OPNA SEZ,
0x9588 OPNA SLEEP, 0x95a8 OPNA WDR,
( ----- 057 )
( 0000 0000 0sss 0000 )
: OPb CREATE , DOES> @ ( b op )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
SWAP _r8c _Rdp , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0b1001010010001000 OPb BCLR, 0b1001010000001000 OPb BSET,
( 0000 000d dddd 0bbb )
: OPRdb CREATE , DOES> @ ( rd b op )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
ROT _r32c _Rdp SWAP _r8c OR , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0b1111100000000000 OPRdb BLD, 0b1111101000000000 OPRdb BST,
0b1111110000000000 OPRdb SBRC, 0b1111111000000000 OPRdb SBRS,
( special cases )
: CLR, DUP EOR, ; : TST, DUP AND, ; : LSL, DUP ADD, ;
( ----- 058 )
( a -- k12, absolute addr a, relative to PC in a k12 addr )
: _r7ffc DUP 0x7ff > IF _oor THEN ;
: _raddr12
PC - DUP 0< IF 0x800 + _r7ffc 0x800 OR ELSE _r7ffc THEN ;
: RJMP _raddr12 0xc000 OR ;
: RCALL _raddr12 0xd000 OR ;
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: RJMP, RJMP , ; : RCALL, RCALL , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 059 )
( a -- k7, absolute addr a, relative to PC in a k7 addr )
: _r3fc DUP 0x3f > IF _oor THEN ;
: _raddr7
PC - DUP 0< IF 0x40 + _r3fc 0x40 OR ELSE _r3fc THEN ;
: _brbx ( a b op -- a ) OR SWAP _raddr7 3 LSHIFT OR ;
: BRBC 0xf400 _brbx ; : BRBS 0xf000 _brbx ; : BRCC 0 BRBC ;
: BRCS 0 BRBS ; : BREQ 1 BRBS ; : BRNE 1 BRBC ; : BRGE 4 BRBC ;
: BRHC 5 BRBC ; : BRHS 5 BRBS ; : BRID 7 BRBC ; : BRIE 7 BRBS ;
: BRLO BRCS ; : BRLT 4 BRBS ; : BRMI 2 BRBS ; : BRPL 2 BRBC ;
: BRSH BRCC ; : BRTC 6 BRBC ; : BRTS 6 BRBS ; : BRVC 3 BRBC ;
: BRVS 3 BRBS ;
( ----- 060 )
0b11100 CONSTANT X 0b01000 CONSTANT Y 0b00000 CONSTANT Z
0b11101 CONSTANT X+ 0b11001 CONSTANT Y+ 0b10001 CONSTANT Z+
0b11110 CONSTANT -X 0b11010 CONSTANT -Y 0b10010 CONSTANT -Z
: _ldst ( Rd XYZ op ) SWAP DUP 0x10 AND 8 LSHIFT SWAP 0xf AND
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OR OR ( Rd op' ) SWAP _Rdp , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: LD, 0x8000 _ldst ; : ST, SWAP 0x8200 _ldst ;
( ----- 061 )
( L1 LBL! .. L1 ' RJMP LBL, )
: LBL! ( l -- ) PC SWAP ! ;
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: LBL, ( l op -- ) SWAP @ 1- SWAP EXECUTE , ;
: SKIP, PC 0 , ;
Rename A!, A@ and A, to C!*, C@* and C,* Also, rename "Addressed devices" to "Indirect memory access". I do this because I need to add indirect versions of !, @ and , to allow boostrapping directly to EEPROM and that A,, thing I've added to assemblers felt like really bad names. With this change, I'd like to generalize the use of the * suffix for aliases.
2020-12-08 11:06:58 +11:00
: TO, ( opw pc ) ( TODO: use !* instead of ! )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( warning: pc is a PC offset, not a mem addr! )
2 * ORG @ + PC 1- H@ ( opw addr tgt hbkp )
ROT HERE ! ( opw tgt hbkp ) SWAP ROT EXECUTE H@ ! ( hbkp )
HERE ! ;
( L1 FLBL, .. L1 ' RJMP FLBL! )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: FLBL, ( l -- ) LBL! 0 , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: FLBL! ( l opw -- ) SWAP @ TO, ;
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: BEGIN, PC ; : AGAIN?, ( op ) SWAP 1- SWAP EXECUTE , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: AGAIN, ['] RJMP AGAIN?, ;
: IF, ['] BREQ SKIP, ; : THEN, TO, ;
( ----- 062 )
( Constant common to all AVR models )
: R0 0 ; : R1 1 ; : R2 2 ; : R3 3 ; : R4 4 ; : R5 5 ; : R6 6 ;
: R7 7 ; : R8 8 ; : R9 9 ; : R10 10 ; : R11 11 ; : R12 12 ;
: R13 13 ; : R14 14 ; : R15 15 ; : R16 16 ; : R17 17 ;
: R18 18 ; : R19 19 ; : R20 20 ; : R21 21 ; : R22 22 ;
: R24 24 ; : R25 25 ; : R26 26 ; : R27 27 ; : R28 28 ;
: R29 29 ; : R30 30 ; : R31 31 ; : XL R26 ; : XH R27 ;
: YL R28 ; : YH R29 ; : ZL R30 ; : ZH R31 ;
( ----- 065 )
( ATmega328P definitions ) : > CONSTANT ;
0xc6 > UDR0 0xc4 > UBRR0L 0xc5 > UBRR0H 0xc2 > UCSR0C
0xc1 > UCSR0B 0xc0 > UCSR0A 0xbd > TWAMR 0xbc > TWCR
0xbb > TWDR 0xba > TWAR 0xb9 > TWSR 0xb8 > TWBR 0xb6 > ASSR
0xb4 > OCR2B 0xb3 > OCR2A 0xb2 > TCNT2 0xb1 > TCCR2B
0xb0 > TCCR2A 0x8a > OCR1BL 0x8b > OCR1BH 0x88 > OCR1AL
0x89 > OCR1AH 0x86 > ICR1L 0x87 > ICR1H 0x84 > TCNT1L
0x85 > TCNT1H 0x82 > TCCR1C 0x81 > TCCR1B 0x80 > TCCR1A
0x7f > DIDR1 0x7e > DIDR0 0x7c > ADMUX 0x7b > ADCSRB
0x7a > ADCSRA 0x79 > ADCH 0x78 > ADCL 0x70 > TIMSK2
0x6f > TIMSK1 0x6e > TIMSK0 0x6c > PCMSK1 0x6d > PCMSK2
0x6b > PCMSK0 0x69 > EICRA 0x68 > PCICR 0x66 > OSCCAL
0x64 > PRR 0x61 > CLKPR 0x60 > WDTCSR 0x3f > SREG 0x3d > SPL
0x3e > SPH 0x37 > SPMCSR 0x35 > MCUCR 0x34 > MCUSR 0x33 > SMCR
0x30 > ACSR 0x2e > SPDR 0x2d > SPSR 0x2c > SPCR 0x2b > GPIOR2
0x2a > GPIOR1 0x28 > OCR0B 0x27 > OCR0A 0x26 > TCNT0 ( cont. )
( ----- 066 )
( cont. ) 0x25 > TCCR0B 0x24 > TCCR0A 0x23 > GTCCR
0x22 > EEARH 0x21 > EEARL 0x20 > EEDR 0x1f > EECR
0x1e > GPIOR0 0x1d > EIMSK 0x1c > EIFR 0x1b > PCIFR
0x17 > TIFR2 0x16 > TIFR1 0x15 > TIFR0 0x0b > PORTD 0x0a > DDRD
0x09 > PIND 0x08 > PORTC 0x07 > DDRC 0x06 > PINC 0x05 > PORTB
0x04 > DDRB 0x03 > PINB
( ----- 100 )
Block editor
This is an application to conveniently browse the contents of
the disk blocks and edit them. You can load it with "105 LOAD".
See doc/ed.txt
( ----- 105 )
1 7 LOADR+
( ----- 106 )
CREATE ACC 0 ,
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
: _LIST ." Block " DUP . NL> LIST ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: L BLK> @ _LIST ;
: B BLK> @ 1- BLK@ L ;
: N BLK> @ 1+ BLK@ L ;
( ----- 107 )
( Cursor position in buffer. EDPOS/64 is line number )
CREATE EDPOS 0 ,
CREATE IBUF 64 ALLOT0
CREATE FBUF 64 ALLOT0
: _cpos BLK( + ;
: _lpos 64 * _cpos ;
: _pln ( lineno -- )
DUP _lpos DUP 64 + SWAP DO ( lno )
I EDPOS @ _cpos = IF '^' EMIT THEN
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
I C@ DUP SPC < IF DROP SPC THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
EMIT
LOOP ( lno ) 1+ . ;
: _zbuf 64 0 FILL ; ( buf -- )
( ----- 108 )
: _type ( buf -- )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
C< DUP CR = IF 2DROP EXIT THEN SWAP DUP _zbuf ( c a )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
BEGIN ( c a ) C!+ C< TUCK 0x0d = UNTIL ( c a ) C! ;
( user-facing lines are 1-based )
: T 1- DUP 64 * EDPOS ! _pln ;
: P IBUF _type IBUF EDPOS @ _cpos 64 MOVE BLK!! ;
: _mvln+ ( ln -- move ln 1 line down )
DUP 14 > IF DROP EXIT THEN
_lpos DUP 64 + 64 MOVE ;
: _mvln- ( ln -- move ln 1 line up )
DUP 14 > IF DROP 15 _lpos _zbuf
ELSE 1+ _lpos DUP 64 - 64 MOVE THEN ;
( ----- 109 )
: _U ( U without P, used in VE )
15 EDPOS @ 64 / - ?DUP IF
0 DO
14 I - _mvln+
LOOP THEN ;
: U _U P ;
( ----- 110 )
: _F ( F without _type and _pln. used in VE )
FBUF EDPOS @ _cpos 1+ ( a1 a2 )
BEGIN
C@+ ROT ( a2+1 c2 a1 ) C@+ ROT ( a2+1 a1+1 c1 c2 )
= NOT IF DROP FBUF THEN ( a2 a1 )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
TUCK C@ CR = ( a1 a2 f1 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
OVER BLK) = OR ( a1 a2 f1|f2 )
UNTIL ( a1 a2 )
DUP BLK) < IF BLK( - FBUF + -^ EDPOS ! ELSE DROP THEN ;
: F FBUF _type _F EDPOS @ 64 / _pln ;
( ----- 111 )
: _blen ( buf -- length of str in buf )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
DUP BEGIN C@+ SPC < UNTIL -^ 1- ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: _rbufsz ( size of linebuf to the right of curpos )
EDPOS @ 64 MOD 63 -^ ;
VE: ensure absence of ctrl chars before insertion point Previously, inserting at a position that had null chars before it would result in a block that wouldn't properly display with LIST.
2020-12-25 04:18:53 +11:00
: _lnfix ( --, ensure no ctl chars in line before EDPOS )
EDPOS @ DUP 0xffc0 AND 2DUP = IF 2DROP EXIT THEN DO
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
I _cpos DUP C@ SPC < IF SPC SWAP C! ELSE DROP THEN LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: _i ( i without _pln and _type. used in VE )
_rbufsz IBUF _blen 2DUP > IF
VE: ensure absence of ctrl chars before insertion point Previously, inserting at a position that had null chars before it would result in a block that wouldn't properly display with LIST.
2020-12-25 04:18:53 +11:00
_lnfix TUCK - ( ilen chars-to-move )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
SWAP EDPOS @ _cpos 2DUP + ( ctm ilen a a+ilen )
3 PICK MOVE- ( ctm ilen ) NIP ( ilen )
ELSE DROP 1+ ( ilen becomes rbuffsize+1 ) THEN
DUP IBUF EDPOS @ _cpos ROT MOVE ( ilen ) EDPOS +! BLK!! ;
: i IBUF _type _i EDPOS @ 64 / _pln ;
( ----- 112 )
: icpy ( n -- copy n chars from cursor to IBUF )
IBUF _zbuf EDPOS @ _cpos IBUF ( n a buf ) ROT MOVE ;
: _X ( n -- )
DUP icpy EDPOS @ _cpos 2DUP + ( n a1 a1+n )
SWAP _rbufsz MOVE ( n )
( get to next line - n )
DUP EDPOS @ 0xffc0 AND 0x40 + -^ _cpos ( n a )
SWAP 0 FILL BLK!! ;
: X _X EDPOS @ 64 / _pln ;
: _E FBUF _blen _X ;
: E FBUF _blen X ;
: Y FBUF _blen icpy ;
( ----- 120 )
Visual Editor
This editor, unlike the Block Editor (B100), is grid-based
instead of being command-based. It requires the AT-XY, COLS
and LINES words to be implemented. If you don't have those,
use the Block Editor.
It is loaded with "125 LOAD" and invoked with "VE". Note that
this also fully loads the Block Editor.
This editor uses 19 lines. The top line is the status line and
it's followed by 2 lines showing the contents of IBUF and
FBUF (see B100). There are then 16 contents lines. The contents
shown is that of the currently selected block.
(cont.)
( ----- 121 )
The status line displays the active block number, then the
"modifier" and then the cursor position. When the block is dir-
ty, an "*" is displayed next. At the right corner, a mode letter
can appear. 'R' for replace, 'I' for insert, 'F' for find.
(cont.)
( ----- 122 )
All keystrokes are directly interpreted by VE and have the
effect described below.
Pressing a 0-9 digit accumulates that digit into what is named
the "modifier". That modifier affects the behavior of many
keystrokes described below. The modifier starts at zero, but
most commands interpret a zero as a 1 so that they can have an
effect.
'G' selects the block specified by the modifier as the current
block. Any change made to the previously selected block is
saved beforehand.
'[' and ']' advances the selected block by "modifier". 't' opens
cvm: use Grid in /cvm/forth Also, fixed broken AT-XY in Grid...
2020-11-15 12:08:18 +11:00
the previously opened block.
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
(cont.)
( ----- 123 )
'h' and 'l' move the cursor by "modifier" characters. 'j' and
'k', by lines. 'g' moves to "modifier" line.
'H' goes to the beginning of the line, 'L' to the end.
'w' moves forward by "modifier" words. 'b' moves backward.
'W' moves to end-of-word. 'B', backwards.
'I', 'F', 'Y', 'X' and 'E' invoke the corresponding command
'o' inserts a blank line after the cursor. 'O', before.
'D' deletes "modifier" lines at the cursor. The first of those
lines is copied to IBUF.
(cont.)
( ----- 124 )
'f' puts the contents of your previous cursor movement into
FBUF. If that movement was a forward movement, it brings the
cursor back where it was. This allows for an efficient combi-
nation of movements and 'E'. For example, if you want to delete
the next word, you type 'w', then 'f', then check your FBUF to
be sure, then press 'E'.
'R' goes into replace mode at current cursor position.
Following keystrokes replace current character and advance
cursor. Press return to return to normal mode.
cvm: use Grid in /cvm/forth Also, fixed broken AT-XY in Grid...
2020-11-15 12:08:18 +11:00
'@' re-reads current block even if it's dirty, thus undoing
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
recent changes.
( ----- 125 )
-20 LOAD+ ( B105, block editor )
1 7 LOADR+
( ----- 126 )
CREATE CMD 2 C, '$' C, 0 C,
CREATE PREVPOS 0 , CREATE PREVBLK 0 , CREATE xoff 0 ,
: MIN ( n n - n ) 2DUP > IF SWAP THEN DROP ;
: MAX ( n n - n ) 2DUP < IF SWAP THEN DROP ;
: large? COLS 67 > ; : col- 67 COLS MIN -^ ;
: width large? IF 64 ELSE COLS THEN ;
: acc@ ACC @ 1 MAX ; : pos@ ( x y -- ) EDPOS @ 64 /MOD ;
: num ACC @ SWAP _pdacc IF ACC ! ELSE DROP THEN ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: nspcs ( n -- , spit n space ) 0 DO SPC> LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: aty 0 SWAP AT-XY ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: clrscr COLS LINES * 0 DO SPC I CELL! LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: gutter ( ln n ) OVER + SWAP DO 67 I AT-XY '|' EMIT LOOP ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: status 0 aty ." BLK" SPC> BLK> ? SPC> ACC ?
SPC> pos@ . ',' EMIT . xoff @ IF '>' EMIT THEN SPC>
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
BLKDTY @ IF '*' EMIT THEN 4 nspcs ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: nums 17 1 DO 2 I + aty I . SPC> SPC> LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 127 )
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
: mode! ( c -- ) 4 col- CELL! ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: @emit C@ SPC MAX 0x7f MIN EMIT ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: contents
16 0 DO
large? IF 3 ELSE 0 THEN I 3 + AT-XY
64 I * BLK( + ( lineaddr ) xoff @ + DUP width + SWAP
VE: avoid spitting control chars in "bufp" On a TRS-80 4P, spitting control chars, in addition to being visually distasteful, generated a bug where the blinking cursor would disappear, making using VE much, much harder. I had a hard time reproducing the bug: you don't trigger it by simply spitting control chars, it had to be a specific number of them in a specific order. I didn't identify the sequence, but I know that "bufp" triggered it. This commit was written on a TRS-80 4P and downloaded from my work floppies!
2020-12-22 08:57:39 +11:00
DO I @emit LOOP LOOP
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
large? IF 3 16 gutter THEN ;
: selblk BLK> @ PREVBLK ! BLK@ contents ;
: pos! ( newpos -- ) EDPOS @ PREVPOS !
DUP 0< IF DROP 0 THEN 1023 MIN EDPOS ! ;
: xoff? pos@ DROP ( x )
xoff @ ?DUP IF < IF 0 xoff ! contents THEN ELSE
width >= IF 64 COLS - xoff ! contents THEN THEN ;
: setpos ( -- ) pos@ 3 + ( header ) SWAP ( y x ) xoff @ -
large? IF 3 + ( gutter ) THEN SWAP AT-XY ;
( ----- 128 )
VE: don't emit chars higher than 0x7f Uninitialized blocks often contains garbage, and emitting it on the TRS-80 4P results in serious visual glitches.
2020-12-23 10:34:03 +11:00
: cmv ( n -- , char movement ) acc@ * EDPOS @ + pos! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: buftype ( buf ln -- )
3 OVER AT-XY KEY DUP EMIT
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
DUP SPC < IF 2DROP DROP EXIT THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( buf ln c ) 4 col- nspcs SWAP 4 SWAP AT-XY ( buf c )
Revisit RDLN words It's been a long while since I visited this part of the code and it has become a bit messy after having gone through all evolutions of the core code. It is now simpler, more compact.
2021-01-03 05:30:32 +11:00
SWAP C!+ IN( _zbuf RDLN IN( SWAP 63 MOVE ;
VE: refresh IBUF and FBUF only when needed In emulation, emitting is near-immediate. On a real machine, it's often not-so-near-immediate. These little things help usability. Again, this commit was meta-developed!
2020-12-24 03:37:42 +11:00
: bufp ( buf -- )
DUP 3 col- + SWAP DO I @emit LOOP ;
: bufs
1 aty ." I: " IBUF bufp
2 aty ." F: " FBUF bufp
large? IF 0 3 gutter THEN ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 129 )
: $G ACC @ selblk ;
: $[ BLK> @ acc@ - selblk ;
: $] BLK> @ acc@ + selblk ;
cvm: use Grid in /cvm/forth Also, fixed broken AT-XY in Grid...
2020-11-15 12:08:18 +11:00
: $t PREVBLK @ selblk ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: $I 'I' mode! IBUF 1 buftype _i bufs contents SPC mode! ;
: $F 'F' mode! FBUF 2 buftype _F bufs setpos SPC mode! ;
VE: refresh IBUF and FBUF only when needed In emulation, emitting is near-immediate. On a real machine, it's often not-so-near-immediate. These little things help usability. Again, this commit was meta-developed!
2020-12-24 03:37:42 +11:00
: $Y Y bufs ;
: $E _E bufs contents ;
: $X acc@ _X bufs contents ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: $h -1 cmv ; : $l 1 cmv ; : $k -64 cmv ; : $j 64 cmv ;
: $H EDPOS @ 0x3c0 AND pos! ;
: $L EDPOS @ 0x3f OR pos! ;
: $g ACC @ 1 MAX 1- 64 * pos! ;
Make BLK@* and BLK!* into ialiases
2021-01-03 07:51:12 +11:00
: $@ BLK> @ BLK@* 0 BLKDTY ! contents ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 130 )
: $w EDPOS @ BLK( + acc@ 0 DO
BEGIN C@+ WS? UNTIL BEGIN C@+ WS? NOT UNTIL LOOP
1- BLK( - pos! ;
: $W EDPOS @ BLK( + acc@ 0 DO
1+ BEGIN C@+ WS? NOT UNTIL BEGIN C@+ WS? UNTIL LOOP
2- BLK( - pos! ;
: $b EDPOS @ BLK( + acc@ 0 DO
1- BEGIN C@- WS? NOT UNTIL BEGIN C@- WS? UNTIL LOOP
2+ BLK( - pos! ;
: $B EDPOS @ BLK( + acc@ 0 DO
BEGIN C@- WS? UNTIL BEGIN C@- WS? NOT UNTIL LOOP
1+ BLK( - pos! ;
( ----- 131 )
: $f EDPOS @ PREVPOS @ 2DUP = IF 2DROP EXIT THEN
2DUP > IF DUP pos! SWAP THEN
( p1 p2, p1 < p2 ) OVER - 64 MIN ( pos len ) FBUF _zbuf
VE: refresh IBUF and FBUF only when needed In emulation, emitting is near-immediate. On a real machine, it's often not-so-near-immediate. These little things help usability. Again, this commit was meta-developed!
2020-12-24 03:37:42 +11:00
SWAP _cpos FBUF ( len src dst ) ROT MOVE bufs ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: $R ( replace mode )
VE: try to use CELL! more At-XY+EMIT is slow compared to direct CELL! usage. It is seen very clearly on a SMS when VE is doing its initial clrscr (yes, I could manage to load VE from a SD card and run it on a real SMS!). The grid protocol wasn't there when I wrote VE initially. I only had AT-XY. Now that we have a solid protocol to build on, let's do it.
2020-11-18 04:07:22 +11:00
'R' mode!
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
BEGIN setpos KEY DUP BS? IF -1 EDPOS +! DROP 0 THEN
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
DUP SPC >= IF
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
DUP EMIT EDPOS @ _cpos C! 1 EDPOS +! BLK!! 0
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
THEN UNTIL SPC mode! contents ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: $O _U EDPOS @ 0x3c0 AND DUP pos! _cpos _zbuf BLK!! contents ;
: $o EDPOS @ 0x3c0 < IF EDPOS @ 64 + EDPOS ! $O THEN ;
: $D $H 64 icpy
acc@ 0 DO 16 EDPOS @ 64 / DO I _mvln- LOOP LOOP
VE: refresh IBUF and FBUF only when needed In emulation, emitting is near-immediate. On a real machine, it's often not-so-near-immediate. These little things help usability. Again, this commit was meta-developed!
2020-12-24 03:37:42 +11:00
BLK!! bufs contents ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 132 )
: UPPER DUP 'a' 'z' =><= IF 32 - THEN ;
: handle ( c -- f )
DUP '0' '9' =><= IF num 0 EXIT THEN
DUP CMD 2+ C! CMD FIND IF EXECUTE ELSE DROP THEN
0 ACC ! UPPER 'Q' = ;
: VE
VE: refresh IBUF and FBUF only when needed In emulation, emitting is near-immediate. On a real machine, it's often not-so-near-immediate. These little things help usability. Again, this commit was meta-developed!
2020-12-24 03:37:42 +11:00
1 XYMODE C! clrscr 0 ACC ! 0 PREVPOS ! nums bufs contents
BEGIN xoff? status setpos KEY handle UNTIL
Revisit RDLN words It's been a long while since I visited this part of the code and it has become a bit messy after having gone through all evolutions of the core code. It is now simpler, more compact.
2021-01-03 05:30:32 +11:00
0 XYMODE C! 19 aty IN$ ;
Add a new "Remote Shell" application This allows a COS shell to remotely control another shell from a serial line. It's very crude now, but the plan is to add tools much like the POSIX tools, but for COS.
2021-01-04 06:01:34 +11:00
( ----- 150 )
( Remote Shell )
0 :* rsh<? 0 :* rsh>
: rsh BEGIN
rsh<? IF
DUP 4 ( EOT ) = IF DROP EXIT THEN EMIT THEN
KEY? IF DUP 0x80 < IF rsh> ELSE DROP EXIT THEN THEN
AGAIN ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 160 )
( AVR Programmer, load range 160-163. doc/avr.txt )
( page size in words, 64 is default on atmega328P )
CREATE aspfpgsz 64 ,
VARIABLE aspprevx
: _x ( a -- b ) DUP aspprevx ! (spix) ;
: _xc ( a -- b ) DUP (spix) ( a b )
DUP aspprevx @ = NOT IF ABORT" AVR err" THEN ( a b )
SWAP aspprevx ! ( b ) ;
: _cmd ( b4 b3 b2 b1 -- r4 ) _xc DROP _xc DROP _xc DROP _x ;
: asprdy ( -- ) BEGIN 0 0 0 0xf0 _cmd 1 AND NOT UNTIL ;
: asp$ ( spidevid -- )
( RESET pulse ) DUP (spie) 0 (spie) (spie)
( wait >20ms ) 220 TICKS
( enable prog ) 0xac (spix) DROP
0x53 _x DROP 0 _xc DROP 0 _x DROP ;
: asperase 0 0 0x80 0xac _cmd asprdy ;
( ----- 161 )
( fuse access. read/write one byte at a time )
: aspfl@ ( -- lfuse ) 0 0 0 0x50 _cmd ;
: aspfh@ ( -- hfuse ) 0 0 0x08 0x58 _cmd ;
: aspfe@ ( -- efuse ) 0 0 0x00 0x58 _cmd ;
: aspfl! ( lfuse -- ) 0 0xa0 0xac _cmd ;
: aspfh! ( hfuse -- ) 0 0xa8 0xac _cmd ;
: aspfe! ( efuse -- ) 0 0xa4 0xac _cmd ;
( ----- 162 )
: aspfb! ( n a --, write word n to flash buffer addr a )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
SWAP |L ( a hi lo ) ROT ( hi lo a )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
DUP ROT ( hi a a lo ) SWAP ( hi a lo a )
0 0x40 ( hi a lo a 0 0x40 ) _cmd DROP ( hi a )
0 0x48 _cmd DROP ;
: aspfp! ( page --, write buffer to page )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
0 SWAP aspfpgsz @ * |M ( 0 lsb msb )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x4c _cmd DROP asprdy ;
: aspf@ ( page a -- n, read word from flash )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
SWAP aspfpgsz @ * OR ( addr ) |M ( lsb msb )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
2DUP 0 ROT> ( lsb msb 0 lsb msb )
0x20 _cmd ( lsb msb low )
ROT> 0 ROT> ( low 0 lsb msb ) 0x28 _cmd 8 LSHIFT OR ;
( ----- 163 )
: aspe@ ( addr -- byte, read from EEPROM )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
0 SWAP |L ( 0 msb lsb )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xa0 ( 0 msb lsb 0xa0 ) _cmd ;
: aspe! ( byte addr --, write to EEPROM )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
|L ( b msb lsb )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0xc0 ( b msb lsb 0xc0 ) _cmd DROP asprdy ;
sms: generate TMR SEGA signature in Collapse OS itself Having the signature generation code in /tools prevents self-hosting on the SMS.
2020-11-26 13:42:07 +11:00
( ----- 165 )
( Sega ROM signer. See doc/sega.txt )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
: C!+^ ( a c -- a+1 ) OVER C! 1+ ;
sms: generate TMR SEGA signature in Collapse OS itself Having the signature generation code in /tools prevents self-hosting on the SMS.
2020-11-26 13:42:07 +11:00
: segasig ( addr size -- )
0x2000 OVER LSHIFT ( a sz bytesz )
ROT TUCK + 0x10 - ( sz a end )
TUCK SWAP 0 ROT> ( sz end sum end a ) DO ( sz end sum )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
I C@ + LOOP ( sz end sum ) SWAP ( sz sum end )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
'T' C!+^ 'M' C!+^ 'R' C!+^ SPC C!+^ 'S' C!+^
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
'E' C!+^ 'G' C!+^ 'A' C!+^ 0 C!+^ 0 C!+^
( sum's LSB ) OVER C!+^ ( MSB ) SWAP 8 RSHIFT OVER C! 1+
( sz end ) 0 C!+^ 0 C!+^ 0 C!+^ SWAP 0x4a + SWAP C! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 260 )
Cross compilation program
This programs allows cross compilation of boot binary and
core. Run "262 LOAD" right before your cross compilation and
then "270 LOAD" to apply xcomp overrides.
This unit depends on a properly initialized z80a with ORG and
BIN( set. That is how we determine compilation offsets.
This redefines defining words to achieve cross compilation.
The goal is two-fold:
1. Add an offset to all word references in definitions.
2. Don't shadow important words we need right now.
(cont.)
( ----- 261 )
Words overrides like ":", "IMMEDIATE" and "CODE" are not
automatically shadowed to allow the harmless inclusion of
this unit. This shadowing has to take place in your xcomp
configuration.
See /doc/cross.txt for details.
( ----- 262 )
1 3 LOADR+
( ----- 263 )
CREATE XCURRENT 0 ,
: XCON XCURRENT CURRENT* ! ; : XCOFF 0x02 RAM+ CURRENT* ! ;
: (xentry) XCON (entry) XCOFF ; : XCREATE (xentry) 2 C, ;
: X:** (xentry) 5 C, , ;
: XCODE XCON CODE XCOFF ; : XIMM XCON IMMEDIATE XCOFF ;
: _xapply ( a -- a-off )
DUP ORG @ > IF ORG @ - BIN( @ + THEN ;
: XFIND XCURRENT @ SWAP _find DROP _xapply ;
: XLITN LIT" (n)" XFIND , , ;
: X' XCON ' XCOFF ; : X'? XCON '? XCOFF ;
: X['] XCON ' _xapply XLITN XCOFF ;
: XCOMPILE XCON ' _xapply XLITN
LIT" ," FIND DROP _xapply , XCOFF ;
: X[COMPILE] XCON ' _xapply , XCOFF ;
( ----- 264 )
: XDO LIT" 2>R" XFIND , H@ ;
: XLOOP LIT" (loop)" XFIND , H@ - C, ;
: XIF LIT" (?br)" XFIND , H@ 1 ALLOT ;
: XELSE LIT" (br)" XFIND , 1 ALLOT [COMPILE] THEN H@ 1- ;
: XAGAIN LIT" (br)" XFIND , H@ - C, ;
: XUNTIL LIT" (?br)" XFIND , H@ - C, ;
: XLIT"
LIT" (s)" XFIND , H@ 0 C, ,"
DUP H@ -^ 1- SWAP C!
;
( ----- 265 )
: X:
(xentry) 1 ( compiled ) C,
BEGIN
WORD DUP LIT" ;" S= IF
DROP LIT" EXIT" XFIND , EXIT THEN
XCURRENT @ SWAP ( xcur w ) _find ( a f )
IF ( a )
DUP IMMED? IF ABORT THEN
_xapply ,
ELSE ( w )
0x02 RAM+ @ SWAP ( cur w ) _find ( a f )
IF DUP IMMED? NOT IF ABORT THEN EXECUTE
ELSE (parse) XLITN THEN
THEN
AGAIN ;
( ----- 270 )
: CODE XCODE ;
: '? X'? ;
: ['] X['] ; IMMEDIATE
: COMPILE XCOMPILE ; IMMEDIATE
: [COMPILE] X[COMPILE] ; IMMEDIATE
: DO XDO ; IMMEDIATE : LOOP XLOOP ; IMMEDIATE
: IF XIF ; IMMEDIATE : ELSE XELSE ; IMMEDIATE
: AGAIN XAGAIN ; IMMEDIATE : UNTIL XUNTIL ; IMMEDIATE
: LIT" XLIT" ; IMMEDIATE : LITN XLITN ;
: IMMEDIATE XIMM ;
: (entry) (xentry) ; : CREATE XCREATE ; : :** X:** ;
: : [ ' X: , ] ;
CURRENT @ XCURRENT !
( ----- 280 )
Z80 boot code
This assembles the boot binary. It requires the Z80 assembler
(B5) and cross compilation setup (B260). It requires some
constants to be set. See doc/bootstrap.txt for details.
RESERVED REGISTERS: At all times, IX points to RSP TOS and BC
is IP. SP points to PSP TOS, but you can still use the stack
in native code. you just have to make sure you've restored it
before "next".
The boot binary is loaded in 2 parts. The first part, "decla-
rations", are loaded after xcomp, before xcomp overrides, with
"282 LOAD". The rest, after xcomp overrides, with "283 335
LOADR".
( ----- 282 )
VARIABLE lbluflw VARIABLE lblexec
( see comment at TICKS' definition )
( 7.373MHz target: 737t. outer: 37t inner: 16t )
( tickfactor = (737 - 37) / 16 )
CREATE tickfactor 44 ,
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
( Perform a byte write by taking into account the SYSVARS+3e
override. )
: LD(HL)E*, SYSVARS 0x3e + LDA(i), A ORr,
IFZ, (HL) E LDrr, ELSE, SYSVARS 0x3e + CALL, THEN, ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 283 )
H@ ORG ! ( STABLE ABI )
0 JP, ( 00, main ) NOP, ( unused ) NOP, NOP, ( 04, BOOT )
NOP, NOP, ( 06, uflw ) NOP, NOP, ( 08, LATEST )
NOP, NOP, NOP, NOP, NOP, NOP, ( unused )
0 JP, ( RST 10 ) NOP, NOP, ( 13, oflw )
NOP, NOP, NOP, NOP, NOP, ( unused )
0 JP, ( 1a, next ) NOP, NOP, NOP, ( unused )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
0 JP, ( RST 20 ) 5 ALLOT0
0 JP, ( RST 28 ) 5 ALLOT0
0 JP, ( RST 30 ) 5 ALLOT0
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0 JP, ( RST 38 )
( ----- 284 )
PC ORG @ 1 + ! ( main )
SP PS_ADDR LDdi, IX RS_ADDR LDdi,
( LATEST is a label to the latest entry of the dict. It is
written at offset 0x08 by the process or person building
Forth. )
BIN( @ 0x08 + LDHL(i),
SYSVARS 0x02 ( CURRENT ) + LD(i)HL,
HERESTART [IF]
HL HERESTART LDdi,
[THEN]
SYSVARS 0x04 + LD(i)HL, ( RAM+04 == HERE )
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
A XORr, SYSVARS 0x3e + LD(i)A, ( 3e == ~C! )
Add ~C!ERR A ~C! override can, if it wants, go put an error code in there, which ~AT28 does. This way, after a copy or xcomp process directly to EEPROM, one can verify whether all bytes were successfully written by checking whether "~C!ERR C@" is zero.
2020-12-10 11:44:02 +11:00
SYSVARS 0x41 + LD(i)A, ( 41 == ~C!ERR )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
DE BIN( @ 0x04 ( BOOT ) + LDd(i),
JR, L1 FWR ( execute, B287 )
( ----- 286 )
lblnext BSET PC ORG @ 0x1b + ! ( next )
( This routine is jumped to at the end of every word. In it,
we jump to current IP, but we also take care of increasing
it by 2 before jumping. )
( Before we continue: are we overflowing? )
IX PUSH, EX(SP)HL, ( do EX to count the IX push in SP )
SP SUBHLd, HL POP,
IFNC, ( SP <= IX? overflow )
SP PS_ADDR LDdi, IX RS_ADDR LDdi,
DE BIN( @ 0x13 ( oflw ) + LDd(i),
JR, L2 FWR ( execute, B287 )
THEN,
LDA(BC), E A LDrr, BC INCd,
LDA(BC), D A LDrr, BC INCd,
( continue to execute )
( ----- 287 )
lblexec BSET L1 FSET ( B284 ) L2 FSET ( B286 )
( DE -> wordref )
LDA(DE), DE INCd, EXDEHL, ( HL points to PFA )
A ORr, IFZ, JP(HL), THEN,
A DECr, ( compiled? ) IFNZ, ( no )
3 CPi, IFZ, ( alias ) LDDE(HL), JR, lblexec BWR THEN,
Rename switch to ialias The switch name was confusing.
2020-11-29 02:58:16 +11:00
IFNC, ( ialias )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
LDDE(HL), EXDEHL, LDDE(HL), JR, lblexec BWR THEN,
( cell or does. push PFA ) HL PUSH,
A DECr, JRZ, lblnext BWR ( cell )
HL INCd, HL INCd, LDDE(HL), EXDEHL, ( does )
THEN, ( continue to compiledWord )
( ----- 289 )
Improve impl's word execution documentation Add an example, which I think helps a lot to grasp the idea. Also, improve comments in Z80 boot code.
2020-11-29 03:59:15 +11:00
( compiled word. HL points to its first wordref, which we'll
execute now.
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
1. Push current IP to RS
Improve impl's word execution documentation Add an example, which I think helps a lot to grasp the idea. Also, improve comments in Z80 boot code.
2020-11-29 03:59:15 +11:00
2. Set new IP to PFA+2
3. Execute wordref )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
IX INCd, IX INCd,
0 IX+ C LDIXYr,
1 IX+ B LDIXYr,
( While we inc, dereference into DE for execute call later. )
Improve impl's word execution documentation Add an example, which I think helps a lot to grasp the idea. Also, improve comments in Z80 boot code.
2020-11-29 03:59:15 +11:00
LDDE(HL), ( DE is new wordref )
HL INCd, ( HL is new PFA+2 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
B H LDrr, C L LDrr, ( --> IP )
JR, lblexec BWR ( execute-B287 )
( ----- 290 )
lblchkPS BSET ( chkPS )
( thread carefully in there: sometimes, we're in the
middle of a EXX to protect BC. BC must never be touched
here. )
EXX,
( We have the return address for this very call on the stack
and protected registers. 2- is to compensate that. )
HL PS_ADDR 2- LDdi,
SP SUBHLd,
EXX,
CNC RETc, ( PS_ADDR >= SP? good )
( continue to uflw )
lbluflw BSET ( abortUnderflow )
DE BIN( @ 0x06 ( uflw ) + LDd(i),
JR, lblexec BWR
( ----- 291 )
( Native words )
H@ 5 + XCURRENT ! ( make next CODE have 0 prev field )
CODE _find ( cur w -- a f )
HL POP, ( w ) DE POP, ( cur ) chkPS,
HL PUSH, ( --> lvl 1 )
( First, figure out string len )
A (HL) LDrr, A ORr,
( special case. zero len? we never find anything. )
IFZ, PUSH0, JPNEXT, THEN,
BC PUSH, ( --> lvl 2, protect )
( Let's do something weird: We'll hold HL by the *tail*.
Because of our dict structure and because we know our
lengths, it's easier to compare starting from the end. )
C A LDrr, B 0 LDri, ( C holds our length )
BC ADDHLd, HL INCd, ( HL points to after-last-char )
( cont . )
( ----- 292 )
BEGIN, ( loop )
( DE is a wordref, first step, do our len correspond? )
HL PUSH, ( --> lvl 3 )
DE PUSH, ( --> lvl 4 )
DE DECd,
LDA(DE),
0x7f ANDi, ( remove IMMEDIATE flag )
C CPr, ( cont. )
( ----- 293 )
IFZ,
( match, let's compare the string then )
DE DECd, ( Skip prev field. One less because we )
DE DECd, ( pre-decrement )
B C LDrr, ( loop C times )
BEGIN,
( pre-decrement for easier Z matching )
DE DECd,
HL DECd,
LDA(DE),
(HL) CPr,
JRNZ, BREAK,
DJNZ, AGAIN,
THEN,
( cont. )
( ----- 294 )
( At this point, Z is set if we have a match. In all cases,
we want to pop HL and DE )
DE POP, ( <-- lvl 4 )
IFZ, ( match, we're done! )
HL POP, BC POP, HL POP, ( <-- lvl 1-3 ) DE PUSH,
PUSH1, JPNEXT,
THEN,
( no match, go to prev and continue )
DE DECd, DE DECd, DE DECd, ( prev field )
DE PUSH, ( --> lvl 4 )
EXDEHL,
LDDE(HL),
( cont. )
( ----- 295 )
( DE contains prev offset )
HL POP, ( <-- lvl 4, prev field )
DEZ, IFNZ, ( offset not zero )
( get absolute addr from offset )
( carry cleared from "or e" )
DE SBCHLd,
EXDEHL, ( result in DE )
THEN,
HL POP, ( <-- lvl 3 )
JRNZ, AGAIN, ( loop-B292, try to match again )
BC POP, ( <-- lvl 2 )
( Z set? end of dict, not found. "w" already on PSP TOS )
PUSH0,
;CODE
( ----- 297 )
CODE (br)
L1 BSET ( used in ?br and loop )
LDA(BC), H 0 LDri, L A LDrr,
RLA, IFC, H DECr, THEN,
BC ADDHLd, B H LDrr, C L LDrr,
;CODE
CODE (?br)
HL POP,
HLZ,
JRZ, L1 BWR ( br + 1. False, branch )
( True, skip next byte and don't branch )
BC INCd,
;CODE
( ----- 298 )
CODE (loop)
0 IX+ INC(IXY+), IFZ, 1 IX+ INC(IXY+), THEN, ( I++ )
( Jump if I <> I' )
A 0 IX+ LDrIXY, 2 IX- CP(IXY+), JRNZ, L1 BWR ( branch )
A 1 IX+ LDrIXY, 1 IX- CP(IXY+), JRNZ, L1 BWR ( branch )
( don't branch )
IX DECd, IX DECd, IX DECd, IX DECd,
BC INCd,
;CODE
( ----- 305 )
CODE EXECUTE
DE POP,
chkPS,
lblexec @ JP,
CODE EXIT
C 0 IX+ LDrIXY,
B 1 IX+ LDrIXY,
IX DECd, IX DECd,
JPNEXT,
( ----- 306 )
CODE (n) ( number literal )
( Literal value to push to stack is next to (n) reference
in the atom list. That is where IP is currently pointing.
Read, push, then advance IP. )
LDA(BC), L A LDrr, BC INCd,
LDA(BC), H A LDrr, BC INCd,
HL PUSH,
;CODE
( ----- 307 )
CODE (s) ( string literal )
( Like (n) but instead of being followed by a 2 bytes
number, it's followed by a string. When called, puts the
string's address on PS )
BC PUSH,
LDA(BC), C ADDr,
IFC, B INCr, THEN,
C A LDrr,
BC INCd,
;CODE
( ----- 308 )
CODE ROT ( a b c -- b c a )
HL POP, ( C ) DE POP, ( B ) IY POP, ( A ) chkPS,
DE PUSH, ( B ) HL PUSH, ( C ) IY PUSH, ( A )
;CODE
CODE ROT> ( a b c -- c a b )
HL POP, ( C ) DE POP, ( B ) IY POP, ( A ) chkPS,
HL PUSH, ( C ) IY PUSH, ( A ) DE PUSH, ( B )
;CODE
CODE DUP ( a -- a a )
HL POP, chkPS,
HL PUSH, HL PUSH,
;CODE
CODE ?DUP
HL POP, chkPS, HL PUSH,
HLZ, IFNZ, HL PUSH, THEN,
;CODE
( ----- 309 )
CODE DROP ( a -- )
HL POP, chkPS,
;CODE
CODE SWAP ( a b -- b a )
HL POP, ( B ) DE POP, ( A )
chkPS,
HL PUSH, ( B ) DE PUSH, ( A )
;CODE
CODE OVER ( a b -- a b a )
HL POP, ( B ) DE POP, ( A )
chkPS,
DE PUSH, ( A ) HL PUSH, ( B ) DE PUSH, ( A )
;CODE
( ----- 310 )
CODE PICK EXX, ( protect BC )
HL POP,
( x2 )
L SLA, H RL,
SP ADDHLd,
C (HL) LDrr,
HL INCd,
B (HL) LDrr,
( check PS range before returning )
EXDEHL,
HL PS_ADDR LDdi,
DE SUBHLd,
IFC, EXX, lbluflw @ JP, THEN,
BC PUSH,
EXX, ( unprotect BC ) ;CODE
( ----- 311 )
( Low-level part of ROLL. Example:
"1 2 3 4 4 (roll)" --> "1 3 4 4". No sanity checks, never
call with 0. )
CODE (roll)
HL POP,
B H LDrr,
C L LDrr,
SP ADDHLd,
HL INCd,
D H LDrr,
E L LDrr,
HL DECd,
HL DECd,
LDDR,
;CODE
( ----- 312 )
CODE 2DROP ( a b -- )
HL POP, HL POP, chkPS,
;CODE
CODE 2DUP ( a b -- a b a b )
HL POP, ( b ) DE POP, ( a )
chkPS,
DE PUSH, HL PUSH,
DE PUSH, HL PUSH,
;CODE
( ----- 313 )
CODE S0
HL PS_ADDR LDdi,
HL PUSH,
;CODE
CODE 'S
HL 0 LDdi,
SP ADDHLd,
HL PUSH,
;CODE
( ----- 314 )
CODE AND
HL POP,
DE POP,
chkPS,
A E LDrr,
L ANDr,
L A LDrr,
A D LDrr,
H ANDr,
H A LDrr,
HL PUSH,
;CODE
( ----- 315 )
CODE OR
HL POP,
DE POP,
chkPS,
A E LDrr,
L ORr,
L A LDrr,
A D LDrr,
H ORr,
H A LDrr,
HL PUSH,
;CODE
( ----- 316 )
CODE XOR
HL POP,
DE POP,
chkPS,
A E LDrr,
L XORr,
L A LDrr,
A D LDrr,
H XORr,
H A LDrr,
HL PUSH,
;CODE
( ----- 317 )
CODE NOT
HL POP,
chkPS,
HLZ,
PUSHZ,
;CODE
( ----- 318 )
CODE +
HL POP,
DE POP,
chkPS,
DE ADDHLd,
HL PUSH,
;CODE
CODE -
DE POP,
HL POP,
chkPS,
DE SUBHLd,
HL PUSH,
;CODE
( ----- 319 )
CODE * EXX, ( protect BC )
( DE * BC -> DE (high) and HL (low) )
DE POP, BC POP, chkPS,
HL 0 LDdi,
A 0x10 LDri,
BEGIN,
HL ADDHLd,
E RL, D RL,
IFC,
BC ADDHLd,
IFC, DE INCd, THEN,
THEN,
A DECr,
JRNZ, AGAIN,
HL PUSH,
EXX, ( unprotect BC ) ;CODE
( ----- 320 )
( Borrowed from http://wikiti.brandonw.net/ )
( Divides AC by DE and places the quotient in AC and the
remainder in HL )
CODE /MOD EXX, ( protect BC )
DE POP, BC POP, chkPS,
A B LDrr, B 16 LDri,
HL 0 LDdi,
BEGIN,
SCF, C RL, RLA,
HL ADCHLd, DE SBCHLd,
IFC, DE ADDHLd, C DECr, THEN,
DJNZ, AGAIN,
B A LDrr,
HL PUSH, BC PUSH,
EXX, ( unprotect BC ) ;CODE
( ----- 321 )
( The word below is designed to wait the proper 100us per tick
at 500kHz when tickfactor is 1. If the CPU runs faster,
tickfactor has to be adjusted accordingly. "t" in comments
below means "T-cycle", which at 500kHz is worth 2us. )
CODE TICKS
HL POP, chkPS,
( we pre-dec to compensate for initialization )
BEGIN,
HL DECd, ( 6t )
IFZ, ( 12t ) JPNEXT, THEN,
A tickfactor @ LDri, ( 7t )
BEGIN, A DECr, ( 4t ) JRNZ, ( 12t ) AGAIN,
JR, ( 12t ) AGAIN, ( outer: 37t inner: 16t )
( ----- 322 )
CODE !
HL POP, DE POP, chkPS,
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
LD(HL)E*, HL INCd,
E D LDrr, LD(HL)E*,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;CODE
CODE @
HL POP, chkPS,
E (HL) LDrr,
HL INCd,
D (HL) LDrr,
DE PUSH,
;CODE
( ----- 323 )
CODE C!
HL POP, DE POP, chkPS,
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
LD(HL)E*, ;CODE
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
CODE C@
HL POP, chkPS,
L (HL) LDrr,
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
H 0 LDri, HL PUSH, ;CODE
CODE ~C!
HL POP, chkPS,
SYSVARS 0x3f + LD(i)HL,
HLZ, ( makes A zero if Z is set ) IFNZ,
A 0xc3 ( JP ) LDri, THEN,
( A is either 0 or c3 ) SYSVARS 0x3e + LD(i)A,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
;CODE
( ----- 324 )
CODE PC! EXX, ( protect BC )
BC POP, HL POP, chkPS,
L OUT(C)r,
EXX, ( unprotect BC ) ;CODE
CODE PC@ EXX, ( protect BC )
BC POP, chkPS,
H 0 LDri,
L INr(C),
HL PUSH,
EXX, ( unprotect BC ) ;CODE
( ----- 325 )
CODE I
L 0 IX+ LDrIXY, H 1 IX+ LDrIXY,
HL PUSH,
;CODE
CODE I'
L 2 IX- LDrIXY, H 1 IX- LDrIXY,
HL PUSH,
;CODE
CODE J
L 4 IX- LDrIXY, H 3 IX- LDrIXY,
HL PUSH,
;CODE
CODE >R
HL POP, chkPS,
IX INCd, IX INCd, 0 IX+ L LDIXYr, 1 IX+ H LDIXYr,
;CODE
( ----- 326 )
CODE R>
L 0 IX+ LDrIXY, H 1 IX+ LDrIXY, IX DECd, IX DECd, HL PUSH,
;CODE
CODE 2>R
DE POP, HL POP, chkPS,
IX INCd, IX INCd, 0 IX+ L LDIXYr, 1 IX+ H LDIXYr,
IX INCd, IX INCd, 0 IX+ E LDIXYr, 1 IX+ D LDIXYr,
;CODE
CODE 2R>
L 0 IX+ LDrIXY, H 1 IX+ LDrIXY, IX DECd, IX DECd,
E 0 IX+ LDrIXY, D 1 IX+ LDrIXY, IX DECd, IX DECd,
DE PUSH, HL PUSH,
;CODE
( ----- 327 )
CODE BYE
HALT,
;CODE
CODE (resSP)
SP PS_ADDR LDdi,
;CODE
CODE (resRS)
IX RS_ADDR LDdi,
;CODE
( ----- 328 )
CODE S= EXX, ( protect BC )
DE POP, HL POP, chkPS,
LDA(DE),
(HL) CPr,
IFZ, ( same size? )
B A LDrr, ( loop A times )
BEGIN,
HL INCd, DE INCd,
LDA(DE),
(HL) CPr,
JRNZ, BREAK, ( not equal? break early. NZ is set. )
DJNZ, AGAIN,
THEN,
PUSHZ,
EXX, ( unprotect BC ) ;CODE
( ----- 329 )
CODE CMP
HL POP,
DE POP,
chkPS,
DE SUBHLd,
DE 0 LDdi,
IFNZ, ( < or > )
DE INCd,
IFNC, ( < )
DE DECd,
DE DECd,
THEN,
THEN,
DE PUSH,
;CODE
( ----- 331 )
CODE (im1)
IM1,
EI,
;CODE
CODE 0 PUSH0, ;CODE
CODE 1 PUSH1, ;CODE
CODE -1
HL -1 LDdi,
HL PUSH,
;CODE
( ----- 332 )
CODE 1+
HL POP,
chkPS,
HL INCd,
HL PUSH,
;CODE
CODE 1-
HL POP,
chkPS,
HL DECd,
HL PUSH,
;CODE
( ----- 333 )
CODE 2+
HL POP,
chkPS,
HL INCd,
HL INCd,
HL PUSH,
;CODE
CODE 2-
HL POP,
chkPS,
HL DECd,
HL DECd,
HL PUSH,
;CODE
( ----- 334 )
CODE RSHIFT ( n u -- n )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
DE POP, ( u ) HL POP, ( n ) chkPS,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
A E LDrr,
A ORr, IFNZ,
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
BEGIN, H SRL, L RR, A DECr, JRNZ, AGAIN,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
THEN,
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
HL PUSH, ;CODE
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
CODE LSHIFT ( n u -- n )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
DE POP, ( u ) HL POP, ( n ) chkPS,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
A E LDrr,
A ORr, IFNZ,
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
BEGIN, L SLA, H RL, A DECr, JRNZ, AGAIN,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
THEN,
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
HL PUSH, ;CODE
( ----- 335 )
CODE |L ( n -- msb lsb )
HL POP, chkPS,
D 0 LDri, E H LDrr, DE PUSH,
E L LDrr, DE PUSH, ;CODE
CODE |M ( n -- lsb msb )
HL POP, chkPS,
D 0 LDri, E L LDrr, DE PUSH,
E H LDrr, DE PUSH, ;CODE
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 350 )
Core words
This section contains arch-independent core words of Collapse
OS. Those words are written in a way that make them entirely
cross-compilable (see B260). When building Collapse OS, these
words come right after the boot binary (B280).
Because this unit is designed to be cross-compiled, things are
a little weird. It is compiling in the context of a full
Forth interpreter with all bells and whistles (and z80
assembler), but it has to obey strict rules:
1. Although it cannot compile a word that isn't defined yet,
it can still execute an immediate from the host system.
(cont.)
( ----- 351 )
2. Immediate words that have been cross compiled *cannot* be
used. Only immediates from the host system can be used.
3. If an immediate word compiles words, it can only be words
that are part of the stable ABI.
All of this is because when cross compiling, all atom ref-
erences are offsetted to the target system and are thus
unusable directly. For the same reason, any reference to a word
in the host system will obviously be wrong in the target
system. More details in B260.
(cont.)
( ----- 352 )
This unit is loaded in two "low" and "high" parts. The low part
is the biggest chunk and has the most definitions. The high
part is the "sensitive" chunk and contains "LITN", ":" and ";"
definitions which, once defined, kind of make any more defs
impossible.
The gap between these 2 parts is the ideal place to put device
driver code. Load the low part with "353 LOAD", the high part
with "390 LOAD"
( ----- 353 )
: RAM+ [ SYSVARS LITN ] + ; : BIN+ [ BIN( @ LITN ] + ;
Add ~C!ERR A ~C! override can, if it wants, go put an error code in there, which ~AT28 does. This way, after a copy or xcomp process directly to EEPROM, one can verify whether all bytes were successfully written by checking whether "~C!ERR C@" is zero.
2020-12-10 11:44:02 +11:00
: HERE 0x04 RAM+ ; : ~C!ERR 0x41 RAM+ ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: CURRENT* 0x51 RAM+ ; : CURRENT CURRENT* @ ;
: H@ HERE @ ;
: FIND ( w -- a f ) CURRENT @ SWAP _find ;
: IN> 0x30 RAM+ ; ( current position in INBUF )
Integrate RDLN input buffer into sysvars In the beginning of Collapse OS' Forth version, the readline sub- system was optional. This is why we had this separate RDLN$ routine and that the input buffer was allocated at boot time. It's been a while since the RDLN system has been made mandatory, but we still paid the complexity tax of this separation. Not anymore.
2021-01-03 06:20:50 +11:00
: IN( 0x60 RAM+ ; ( points to INBUF )
Revisit RDLN words It's been a long while since I visited this part of the code and it has become a bit messy after having gone through all evolutions of the core code. It is now simpler, more compact.
2021-01-03 05:30:32 +11:00
: IN$ 0 IN( DUP IN> ! ! ; ( flush input buffer )
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
: C<* 0x0c RAM+ ;
: QUIT (resRS) 0 C<* ! IN$ LIT" (main)" FIND DROP EXECUTE ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
1 25 LOADR+
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 354 )
: ABORT (resSP) QUIT ;
: = CMP NOT ; : < CMP -1 = ; : > CMP 1 = ;
: 0< 32767 > ; : >= < NOT ; : <= > NOT ; : 0>= 0< NOT ;
: >< ( n l h -- f ) 2 PICK > ( n l f ) ROT> > AND ;
: =><= 2 PICK >= ( n l f ) ROT> >= AND ;
: NIP SWAP DROP ; : TUCK SWAP OVER ;
: -^ SWAP - ;
: C@+ ( a -- a+1 c ) DUP C@ SWAP 1+ SWAP ;
: C!+ ( c a -- a+1 ) TUCK C! 1+ ;
: C@- ( a -- a-1 c ) DUP C@ SWAP 1- SWAP ;
: C!- ( c a -- a-1 ) TUCK C! 1- ;
: LEAVE R> R> DROP I 1- >R >R ; : UNLOOP R> 2R> 2DROP >R ;
( ----- 355 )
: +! TUCK @ + SWAP ! ;
: *! ( addr alias -- ) 1+ ! ;
Rename switch to ialias The switch name was confusing.
2020-11-29 02:58:16 +11:00
: **! ( addr ialias -- ) 1+ @ ! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: / /MOD NIP ;
: MOD /MOD DROP ;
: ALLOT HERE +! ;
Have FILL use indirect memory access This allows SMS xcomp to use ALLOT0 instead of ZFILL,
2020-12-08 11:17:51 +11:00
: FILL ( a n b -- )
ROT> OVER ( b a n a ) + SWAP ( b a+n a ) DO ( b )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
DUP I C! LOOP DROP ;
Have FILL use indirect memory access This allows SMS xcomp to use ALLOT0 instead of ZFILL,
2020-12-08 11:17:51 +11:00
: ALLOT0 ( n -- ) H@ OVER 0 FILL ALLOT ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 356 )
SYSVARS 0x53 + :** EMIT
Rename (print) to STYPE This brings us a bit closer to TYPE from Forth 2012. I don't think I'll add TYPE (sig "addr len") anytime soon because there is few use cases for it, but having "STYPE" instead of "(print)" feels cleaner.
2020-11-29 03:28:28 +11:00
: STYPE C@+ ( a len ) 0 DO C@+ EMIT LOOP DROP ;
Add EOT ASCII const
2021-01-03 05:55:20 +11:00
: EOT 0x4 ; : BS 0x8 ; : LF 0xa ; : CR 0xd ; : SPC 0x20 ;
: SPC> SPC EMIT ;
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
: NL> 0x50 RAM+ C@ ?DUP IF EMIT ELSE 13 EMIT 10 EMIT THEN ;
Rename (print) to STYPE This brings us a bit closer to TYPE from Forth 2012. I don't think I'll add TYPE (sig "addr len") anytime soon because there is few use cases for it, but having "STYPE" instead of "(print)" feels cleaner.
2020-11-29 03:28:28 +11:00
: ERR STYPE ABORT ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: (uflw) LIT" stack underflow" ERR ;
XCURRENT @ _xapply ORG @ 0x06 ( stable ABI uflw ) + !
: (oflw) LIT" stack overflow" ERR ;
XCURRENT @ _xapply ORG @ 0x13 ( stable ABI oflw ) + !
Rename (print) to STYPE This brings us a bit closer to TYPE from Forth 2012. I don't think I'll add TYPE (sig "addr len") anytime soon because there is few use cases for it, but having "STYPE" instead of "(print)" feels cleaner.
2020-11-29 03:28:28 +11:00
: (wnf) STYPE LIT" word not found" ERR ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 357 )
( r c -- r f )
( Parse digit c and accumulate into result r.
Flag f is true when c was a valid digit )
: _pdacc
'0' - DUP 10 < IF ( good, add to running result )
SWAP 10 * + 1 ( r*10+n f )
ELSE ( bad ) DROP 0 THEN ;
( ----- 358 )
: _pd ( a -- n f, parse decimal )
C@+ OVER C@ 0 ( a len firstchar startat )
( if we have '-', we only advance. more processing later. )
SWAP '-' = IF 1+ THEN ( a len startat )
( if we can do the whole string, success. if _pdacc returns
false before, failure. )
0 ROT> ( len ) ( startat ) DO ( a r )
OVER I + C@ ( a r c ) _pdacc ( a r f )
NOT IF DROP 1- 0 UNLOOP EXIT THEN LOOP ( a r )
( if we had '-', we need to invert result. )
SWAP C@ '-' = IF 0 -^ THEN 1 ( r 1 ) ;
( ----- 359 )
( strings being sent to parse routines are always null
terminated )
: _pc ( a -- n f, parse character )
( apostrophe is ASCII 39 )
DUP 1+ C@ 39 = OVER 3 + C@ 39 = AND ( a f )
NOT IF 0 EXIT THEN ( a 0 )
( surrounded by apos, good, return )
2+ C@ 1 ( n 1 )
;
( ----- 360 )
( returns negative value on error )
: _ ( c -- n )
DUP '0' '9' =><= IF '0' - EXIT THEN
DUP 'a' 'f' =><= IF 0x57 ( 'a' - 10 ) - EXIT THEN
DROP -1 ( bad )
;
: _ph ( a -- n f, parse hex )
( '0': ASCII 0x30 'x': 0x78 0x7830 )
DUP 1+ @ 0x7830 = NOT IF 0 EXIT THEN ( a 0 )
( We have "0x" prefix )
DUP C@ ( a len )
0 SWAP 1+ ( len+1 ) 3 DO ( a r )
OVER I + C@ ( a r c ) _ ( a r n )
DUP 0< IF 2DROP 0 UNLOOP EXIT THEN
SWAP 4 LSHIFT + ( a r*16+n ) LOOP
NIP 1 ;
( ----- 361 )
: _pb ( a -- n f, parse binary )
( '0': ASCII 0x30 'b': 0x62 0x6230 )
DUP 1+ @ 0x6230 = NOT IF 0 EXIT THEN ( a 0 )
( We have "0b" prefix )
DUP C@ ( a len )
0 SWAP 1+ ( len+1 ) 3 DO ( a r )
OVER I + C@ ( a r c )
DUP '0' '1' =><= NOT IF 2DROP 0 UNLOOP EXIT THEN
'0' - SWAP 1 LSHIFT + ( a r*2+n ) LOOP
NIP 1 ;
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
: (parse) ( a -- n )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
_pc IF EXIT THEN
_ph IF EXIT THEN
_pb IF EXIT THEN
_pd IF EXIT THEN
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
( nothing works ) (wnf) ;
( ----- 362 )
: EOT? EOT = ;
SYSVARS 0x55 + :** KEY?
: KEY BEGIN KEY? UNTIL ;
( del is same as backspace )
: BS? DUP 0x7f = SWAP BS = OR ;
: RDLN ( Read 1 line in input buff and make IN> point to it )
IN$ BEGIN
( buffer overflow? same as if we typed a newline )
IN> @ IN( - 0x3e = IF CR ELSE KEY THEN ( c )
DUP BS? IF
IN> @ IN( > IF -1 IN> +! BS EMIT THEN SPC> BS EMIT
ELSE DUP LF = IF DROP CR THEN ( same as CR )
Have RDLN spit newlines more consistently On CRLF interfaces, newlines were mostly broken after previous RDLN consolidation.
2021-01-03 08:46:57 +11:00
DUP SPC >= IF DUP EMIT ( echo back ) THEN
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
DUP IN> @ ! 1 IN> +! THEN ( c )
Have RDLN spit newlines more consistently On CRLF interfaces, newlines were mostly broken after previous RDLN consolidation.
2021-01-03 08:46:57 +11:00
DUP CR = SWAP EOT? OR UNTIL NL> IN( IN> ! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 363 )
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
: RDLN<
IN> @ C@ ( c )
DUP IF ( not EOL? good, inc and return )
1 IN> +!
ELSE ( EOL ? readline. we still return null though )
RDLN
THEN ( c )
( update C<? flag )
IN> @ C@ 0 > 0x06 RAM+ ! ( 06 == C<? ) ;
( ----- 364 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: C<? 0x06 RAM+ @ ;
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
: C< C<* @ ?DUP NOT IF RDLN< ELSE EXECUTE THEN ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: , H@ ! H@ 2+ HERE ! ;
: C, H@ C!+ HERE ! ;
: ,"
BEGIN
C< DUP 34 ( ASCII " ) = IF DROP EXIT THEN C,
AGAIN ;
Add EOT ASCII const
2021-01-03 05:55:20 +11:00
: EOT, EOT C, ;
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
: WS? SPC <= ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
Add EOT ASCII const
2021-01-03 05:55:20 +11:00
: TOWORD ( -- c, c being the first letter of the word )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0 ( dummy ) BEGIN
Add EOT ASCII const
2021-01-03 05:55:20 +11:00
DROP C< DUP WS? NOT OVER EOT? OR UNTIL ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 365 )
( Read word from C<, copy to WORDBUF, null-terminate, and
return WORDBUF. )
: _wb 0x0e RAM+ ;
: _eot 0x0401 _wb ! _wb ;
: WORD
_wb 1+ TOWORD ( a c )
DUP EOT? IF 2DROP _eot EXIT THEN
BEGIN
OVER C! 1+ C< ( a c )
OVER 0x2e RAM+ = OVER WS? OR
UNTIL ( a c )
SWAP _wb - 1- ( ws len ) _wb C!
EOT? IF _eot ELSE _wb THEN ;
( ----- 366 )
: IMMEDIATE
CURRENT @ 1-
DUP C@ 128 OR SWAP C! ;
: IMMED? 1- C@ 0x80 AND ;
: '? WORD FIND ;
: ' '? NOT IF (wnf) THEN ;
: ROLL
?DUP NOT IF EXIT THEN
1+ DUP PICK ( n val )
SWAP 2 * (roll) ( val )
NIP ;
( ----- 367 )
: MOVE ( a1 a2 u -- )
?DUP IF ( u ) 0 DO ( a1 a2 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OVER I + C@ ( src dst x )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
OVER I + ( src dst x dst )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C! ( src dst )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
LOOP THEN 2DROP ;
: MOVE- ( a1 a2 u -- )
?DUP IF ( u ) 0 DO ( a1 a2 )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
OVER I' + I - 1- C@ ( src dst x )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
OVER I' + I - 1- ( src dst x dst )
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
C! ( src dst )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
LOOP THEN 2DROP ;
: MOVE, ( a u -- ) H@ OVER ALLOT SWAP MOVE ;
( ----- 368 )
: [entry] ( w -- )
C@+ ( w+1 len ) TUCK MOVE, ( len )
( write prev value )
H@ CURRENT @ - ,
C, ( write size )
H@ CURRENT ! ;
: (entry) WORD [entry] ;
: CREATE (entry) 2 ( cellWord ) C, ;
: VARIABLE CREATE 2 ALLOT ;
( ----- 369 )
: FORGET
Remove WORD( and PREV They're only used once.
2020-12-10 09:17:53 +11:00
' DUP ( w w )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( HERE must be at the end of prev's word, that is, at the
beginning of w. )
Remove WORD( and PREV They're only used once.
2020-12-10 09:17:53 +11:00
DUP 1- C@ ( name len field )
0x7f AND ( remove IMMEDIATE flag )
3 + ( fixed header len )
- HERE ! ( w )
( get prev addr ) 3 - DUP @ - CURRENT ! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: EMPTY LIT" _sys" FIND IF DUP HERE ! CURRENT ! THEN ;
( ----- 370 )
: DOES>
( Overwrite cellWord in CURRENT )
3 ( does ) CURRENT @ C!
( When we have a DOES>, we forcefully place HERE to 4
bytes after CURRENT. This allows a DOES word to use ","
and "C," without messing everything up. )
CURRENT @ 3 + HERE !
( HERE points to where we should write R> )
R> ,
( We're done. Because we've popped RS, we'll exit parent
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
definition ) ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: CONSTANT CREATE , DOES> @ ;
: [IF]
IF EXIT THEN
LIT" [THEN]" BEGIN DUP WORD S= UNTIL DROP ;
: [THEN] ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 371 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( n -- Fetches block n and write it to BLK( )
Make BLK@* and BLK!* into ialiases
2021-01-03 07:51:12 +11:00
SYSVARS 0x34 + :** BLK@*
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( n -- Write back BLK( to storage at block n )
Make BLK@* and BLK!* into ialiases
2021-01-03 07:51:12 +11:00
SYSVARS 0x36 + :** BLK!*
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( Current blk pointer in ( )
: BLK> 0x38 RAM+ ;
( Whether buffer is dirty )
: BLKDTY 0x3a RAM+ ;
: BLK( 0x3c RAM+ @ ;
: BLK) BLK( 1024 + ;
: BLK$
H@ 0x3c ( BLK(* ) RAM+ !
1024 ALLOT
( LOAD detects end of block with ASCII EOT. This is why
we write it there. )
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
EOT, 0 BLKDTY ! -1 BLK> ! ;
( ----- 372 )
Make BLK@* and BLK!* into ialiases
2021-01-03 07:51:12 +11:00
: BLK! ( -- ) BLK> @ BLK!* 0 BLKDTY ! ;
Make FLUSH also invalidate the blk cache
2020-12-26 08:43:20 +11:00
: FLUSH BLKDTY @ IF BLK! THEN -1 BLK> ! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: BLK@ ( n -- )
DUP BLK> @ = IF DROP EXIT THEN
Make BLK@* and BLK!* into ialiases
2021-01-03 07:51:12 +11:00
FLUSH DUP BLK> ! BLK@* ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: BLK!! 1 BLKDTY ! ;
: WIPE BLK( 1024 0 FILL BLK!! ;
: WIPED? ( -- f )
1 ( f ) BLK) BLK( DO
I C@ IF DROP 0 ( f ) LEAVE THEN LOOP ;
: COPY ( src dst -- )
FLUSH SWAP BLK@ BLK> ! BLK! ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 373 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: . ( n -- )
?DUP NOT IF '0' EMIT EXIT THEN ( 0 is a special case )
( handle negative )
DUP 0< IF '-' EMIT -1 * THEN
Make word "." a bit more compact
2021-01-03 08:04:36 +11:00
999 SWAP ( stop indicator )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
BEGIN
Make word "." a bit more compact
2021-01-03 08:04:36 +11:00
10 /MOD ( r q )
SWAP '0' + SWAP ( d q )
?DUP NOT UNTIL
BEGIN EMIT DUP '9' > UNTIL DROP ( drop stop ) ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 374 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: ? @ . ;
: _
DUP 9 > IF 10 - 'a' +
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
ELSE '0' + THEN ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( For hex display, there are no negatives )
: .x
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
0xff AND 16 /MOD ( l h )
_ EMIT _ EMIT ;
: .X |M .x .x ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 375 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: _ ( a -- a+8 )
DUP ( a a )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
':' EMIT DUP .x SPC>
4 0 DO DUP @ |L .x .x SPC> 2+ LOOP
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
DROP ( a )
8 0 DO
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
C@+ DUP SPC 0x7e =><= NOT IF DROP '.' THEN EMIT
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
LOOP NL> ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: DUMP ( n a -- )
SWAP 8 /MOD SWAP IF 1+ THEN
Remove LF word Also, remove extraneous uses of it. In the rdln routine, we would spit a new line after each line input, but that NL was spurious: we already spit the newline we receive as it is typed.
2021-01-03 01:45:08 +11:00
0 DO _ LOOP ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 376 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: LIST
BLK@
16 0 DO
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
I 1+ DUP 10 < IF SPC> THEN . SPC>
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
64 I * BLK( + DUP 64 + SWAP DO
I C@ DUP 0x1f > IF EMIT ELSE DROP LEAVE THEN
LOOP
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
NL>
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
LOOP ;
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
( ----- 377 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: INTERPRET
BEGIN
WORD DUP @ 0x0401 = ( EOT ) IF DROP EXIT THEN
FIND NOT IF (parse) ELSE EXECUTE THEN
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
C<? NOT IF SPC> LIT" ok" STYPE NL> THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
AGAIN ;
( Read from BOOT C< PTR and inc it. )
: (boot<)
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
0x2e ( BOOT C< PTR ) RAM+ @ C@+ ( a+1 c )
SWAP 0x2e RAM+ ! ( c ) ;
( ----- 378 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: LOAD
BLK> @ >R ( save restorable variables to RSP )
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
C<* @ >R
Small optimization and block compaction
2021-01-03 08:20:38 +11:00
0x06 RAM+ ( C<? ) @ >R 0x2e RAM+ ( boot ptr ) @ >R
BLK@ BLK( 0x2e RAM+ ! ( Point to beginning of BLK )
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
['] (boot<) 0x0c RAM+ !
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
1 0x06 RAM+ ! ( 06 == C<? )
INTERPRET
R> 0x2e RAM+ ! R> 0x06 RAM+ !
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
R> C<* ! R> BLK@ ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: LOAD+ BLK> @ + LOAD ;
( b1 b2 -- )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: LOADR 1+ SWAP DO I DUP . SPC> LOAD LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: LOADR+ BLK> @ + SWAP BLK> @ + SWAP LOADR ;
( ----- 390 )
( xcomp core high )
: (main) INTERPRET BYE ;
: BOOT
0x02 RAM+ CURRENT* !
CURRENT @ 0x2e RAM+ ! ( 2e == BOOT C< PTR )
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
0 0x50 RAM+ C! ( NL> )
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
['] (emit) ['] EMIT **! ['] (key?) ['] KEY? **!
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
['] (boot<) C<* !
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( boot< always has a char waiting. 06 == C<?* )
1 0x06 RAM+ ! INTERPRET
Remove one level of C< override Two was too many.
2021-01-03 07:32:08 +11:00
0 C<* ! IN$
Integrate RDLN input buffer into sysvars In the beginning of Collapse OS' Forth version, the readline sub- system was optional. This is why we had this separate RDLN$ routine and that the input buffer was allocated at boot time. It's been a while since the RDLN system has been made mandatory, but we still paid the complexity tax of this separation. Not anymore.
2021-01-03 06:20:50 +11:00
LIT" _sys" [entry]
Make NL, renamed to NL> a bit simpler It doesn't require emitting words anymore. The rename to NL> is linked to the upcoming commit.
2021-01-03 02:03:27 +11:00
LIT" Collapse OS" STYPE NL> (main) ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
XCURRENT @ _xapply ORG @ 0x04 ( stable ABI BOOT ) + !
1 4 LOADR+
( ----- 391 )
( Now we have "as late as possible" stuff. See bootstrap doc. )
: :* ( addr -- ) (entry) 4 ( alias ) C, , ;
Rename switch to ialias The switch name was confusing.
2020-11-29 02:58:16 +11:00
: :** ( addr -- ) (entry) 5 ( ialias ) C, , ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 392 )
Rename (print) to STYPE This brings us a bit closer to TYPE from Forth 2012. I don't think I'll add TYPE (sig "addr len") anytime soon because there is few use cases for it, but having "STYPE" instead of "(print)" feels cleaner.
2020-11-29 03:28:28 +11:00
: _bchk DUP 0x7f + 0xff > IF LIT" br ovfl" STYPE ABORT THEN ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: DO COMPILE 2>R H@ ; IMMEDIATE
: LOOP COMPILE (loop) H@ - _bchk C, ; IMMEDIATE
( LEAVE is implemented in low xcomp )
: LITN COMPILE (n) , ;
( gets its name at the very end. can't comment afterwards )
: _ BEGIN LIT" )" WORD S= UNTIL ; IMMEDIATE
: _ ( : will get its name almost at the very end )
(entry) 1 ( compiled ) C,
BEGIN
WORD DUP LIT" ;" S= IF DROP COMPILE EXIT EXIT THEN
FIND IF ( is word ) DUP IMMED? IF EXECUTE ELSE , THEN
ELSE ( maybe number ) (parse) LITN THEN
AGAIN ;
( ----- 393 )
: IF ( -- a | a: br cell addr )
COMPILE (?br) H@ 1 ALLOT ( br cell allot )
; IMMEDIATE
: THEN ( a -- | a: br cell addr )
DUP H@ -^ _bchk SWAP ( a-H a ) C!
; IMMEDIATE
: ELSE ( a1 -- a2 | a1: IF cell a2: ELSE cell )
COMPILE (br)
1 ALLOT
[COMPILE] THEN
H@ 1- ( push a. 1- for allot offset )
; IMMEDIATE
: LIT"
COMPILE (s) H@ 0 C, ,"
DUP H@ -^ 1- ( a len ) SWAP C!
; IMMEDIATE
( ----- 394 )
( We don't use ." and ABORT in core, they're not xcomp-ed )
Rename (print) to STYPE This brings us a bit closer to TYPE from Forth 2012. I don't think I'll add TYPE (sig "addr len") anytime soon because there is few use cases for it, but having "STYPE" instead of "(print)" feels cleaner.
2020-11-29 03:28:28 +11:00
: ." [COMPILE] LIT" COMPILE STYPE ; IMMEDIATE
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: ABORT" [COMPILE] ." COMPILE ABORT ; IMMEDIATE
: BEGIN H@ ; IMMEDIATE
: AGAIN COMPILE (br) H@ - _bchk C, ; IMMEDIATE
: UNTIL COMPILE (?br) H@ - _bchk C, ; IMMEDIATE
: [ INTERPRET ; IMMEDIATE
: ] R> DROP ;
: COMPILE ' LITN ['] , , ; IMMEDIATE
: [COMPILE] ' , ; IMMEDIATE
: ['] ' LITN ; IMMEDIATE
':' X' _ 4 - C! ( give : its name )
'(' X' _ 4 - C!
( ----- 400 )
Add ~C!ERR A ~C! override can, if it wants, go put an error code in there, which ~AT28 does. This way, after a copy or xcomp process directly to EEPROM, one can verify whether all bytes were successfully written by checking whether "~C!ERR C@" is zero.
2020-12-10 11:44:02 +11:00
( Write byte E at addr HL, assumed to be an AT28 EEPROM. After
that, poll repeatedly that address until writing is complete.
If last polled value is different than orig, set ~C!ERR )
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
(entry) ~AT28 ( warning: don't touch D register )
Add ~C!ERR A ~C! override can, if it wants, go put an error code in there, which ~AT28 does. This way, after a copy or xcomp process directly to EEPROM, one can verify whether all bytes were successfully written by checking whether "~C!ERR C@" is zero.
2020-12-10 11:44:02 +11:00
(HL) E LDrr, A E LDrr, ( orig ) EXAFAF', ( save )
E (HL) LDrr, ( poll ) BEGIN,
Add ~C! to override memory low level write routines Also, turn AT28! and AT28, into ~AT28, pluggable into ~C!. ~AT28 doesn't check for mismatches. It was too complicated to turn a mismatch into a compiled word we would jump to next. Data integrity has to be checked through another path. Also, remove MOVEW. Without indirect memory access, this word doesn't make sense. Some AVR-specific words will have to be defined.
2020-12-09 13:51:05 +11:00
A (HL) LDrr, ( poll ) E CPr, ( same as old? )
E A LDrr, ( save old poll, Z preserved )
Add ~C!ERR A ~C! override can, if it wants, go put an error code in there, which ~AT28 does. This way, after a copy or xcomp process directly to EEPROM, one can verify whether all bytes were successfully written by checking whether "~C!ERR C@" is zero.
2020-12-10 11:44:02 +11:00
JRNZ, AGAIN,
EXAFAF', ( orig ) E SUBr, ( equal? )
IFNZ, SYSVARS 0x41 + ( ~C!ERR ) LD(i)A, THEN,
RET,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 401 )
Grid subsystem
grid: add XYMODE In VE on the SMS, the first contents line would always be cleared because of NEWLN being called when the FBUF would spit its last char. Inconvenient... I've added a "graphical" mode to the grid subsystem to inhibit this behavior in a graphical situation such as in VE. Also, write a more complete Grid documentation.
2020-11-15 14:35:08 +11:00
See doc/grid.txt.
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
Load range: B402-B403
( ----- 402 )
grid: add XYMODE In VE on the SMS, the first contents line would always be cleared because of NEWLN being called when the FBUF would spit its last char. Inconvenient... I've added a "graphical" mode to the grid subsystem to inhibit this behavior in a graphical situation such as in VE. Also, write a more complete Grid documentation.
2020-11-15 14:35:08 +11:00
: XYPOS [ GRID_MEM LITN ] ; : XYMODE [ GRID_MEM LITN ] 2+ ;
sms: add cursor visual indicator in VDP mode 4 This required the change of CURSOR! signature: we now supply it with both old and new cursor positions. Also, fix broken _bs in Grid subsystem.
2020-11-17 02:05:48 +11:00
'? CURSOR! NIP NOT [IF] : CURSOR! 2DROP ; [THEN]
: XYPOS! COLS LINES * MOD DUP XYPOS @ CURSOR! XYPOS ! ;
grid: Add CURSOR! to the grid protocol With the move of CVM's forth to the grid protocol, we've lost the cursor's visual indication. Now, we have it back. The challenge now is in implementing it in SMS' text mode. In mode 4, it's easy to mark a cell as inverted, but in text mode, that's not possible.
2020-11-17 00:41:09 +11:00
: AT-XY ( x y -- ) COLS * + XYPOS! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
'? NEWLN NIP NOT [IF]
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: NEWLN ( ln -- ) COLS * DUP COLS + SWAP DO SPC I CELL! LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
[THEN]
grid: add XYMODE In VE on the SMS, the first contents line would always be cleared because of NEWLN being called when the FBUF would spit its last char. Inconvenient... I've added a "graphical" mode to the grid subsystem to inhibit this behavior in a graphical situation such as in VE. Also, write a more complete Grid documentation.
2020-11-15 14:35:08 +11:00
: _lf XYMODE C@ IF EXIT THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
XYPOS @ COLS / 1+ LINES MOD DUP NEWLN
grid: Add CURSOR! to the grid protocol With the move of CVM's forth to the grid protocol, we've lost the cursor's visual indication. Now, we have it back. The challenge now is in implementing it in SMS' text mode. In mode 4, it's easy to mark a cell as inverted, but in text mode, that's not possible.
2020-11-17 00:41:09 +11:00
COLS * XYPOS! ;
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
: _bs SPC XYPOS @ TUCK CELL! ( pos ) 1- XYPOS! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 403 )
: (emit)
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
DUP BS? IF DROP _bs EXIT THEN
DUP CR = IF DROP _lf EXIT THEN
DUP SPC < IF DROP EXIT THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
XYPOS @ CELL!
grid: fix PSP leak on newline
2020-11-18 03:19:37 +11:00
XYPOS @ 1+ DUP COLS MOD IF XYPOS! ELSE DROP _lf THEN ;
sms: add cursor visual indicator in VDP mode 4 This required the change of CURSOR! signature: we now supply it with both old and new cursor positions. Also, fix broken _bs in Grid subsystem.
2020-11-17 02:05:48 +11:00
: GRID$ 0 XYPOS ! 0 XYMODE C! ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 410 )
PS/2 keyboard subsystem
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
Provides (key?) from a driver providing the PS/2 protocol. That
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
is, for a driver taking care of providing all key codes emanat-
ing from a PS/2 keyboard, this subsystem takes care of mapping
those keystrokes to ASCII characters. This code is designed to
be cross-compiled and loaded with drivers.
Requires PS2_MEM to be defined.
Load range: 411-414
( ----- 411 )
: PS2_SHIFT [ PS2_MEM LITN ] ;
: PS2$ 0 PS2_SHIFT C! ;
( A list of the values associated with the 0x80 possible scan
codes of the set 2 of the PS/2 keyboard specs. 0 means no
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
value. That value is a character that can be read in (key?)
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
No make code in the PS/2 set 2 reaches 0x80. )
CREATE PS2_CODES
( 00 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C,
( 08 ) 0 C, 0 C, 0 C, 0 C, 0 C, 9 C, '`' C, 0 C,
( 10 ) 0 C, 0 C, 0 C, 0 C, 0 C, 'q' C, '1' C, 0 C,
( I don't know why, but the key 2 is sent as 0x1f by 2 of my
keyboards. Is it a timing problem on the ATtiny? TODO )
( 18 ) 0 C, 0 C, 'z' C, 's' C, 'a' C, 'w' C, '2' C, '2' C,
( 20 ) 0 C, 'c' C, 'x' C, 'd' C, 'e' C, '4' C, '3' C, 0 C,
( 28 ) 0 C, 32 C, 'v' C, 'f' C, 't' C, 'r' C, '5' C, 0 C,
( ----- 412 )
( 30 ) 0 C, 'n' C, 'b' C, 'h' C, 'g' C, 'y' C, '6' C, 0 C,
( 38 ) 0 C, 0 C, 'm' C, 'j' C, 'u' C, '7' C, '8' C, 0 C,
( 40 ) 0 C, ',' C, 'k' C, 'i' C, 'o' C, '0' C, '9' C, 0 C,
( 48 ) 0 C, '.' C, '/' C, 'l' C, ';' C, 'p' C, '-' C, 0 C,
( 50 ) 0 C, 0 C, ''' C, 0 C, '[' C, '=' C, 0 C, 0 C,
( 58 ) 0 C, 0 C, 13 C, ']' C, 0 C, '\' C, 0 C, 0 C,
( 60 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 8 C, 0 C,
( 68 ) 0 C, '1' C, 0 C, '4' C, '7' C, 0 C, 0 C, 0 C,
( 70 ) '0' C, '.' C, '2' C, '5' C, '6' C, '8' C, 27 C, 0 C,
( 78 ) 0 C, 0 C, '3' C, 0 C, 0 C, '9' C, 0 C, 0 C,
( Same values, but shifted )
( 00 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C,
( 08 ) 0 C, 0 C, 0 C, 0 C, 0 C, 9 C, '~' C, 0 C,
( 10 ) 0 C, 0 C, 0 C, 0 C, 0 C, 'Q' C, '!' C, 0 C,
( 18 ) 0 C, 0 C, 'Z' C, 'S' C, 'A' C, 'W' C, '@' C, '@' C,
( 20 ) 0 C, 'C' C, 'X' C, 'D' C, 'E' C, '$' C, '#' C, 0 C,
( ----- 413 )
( 28 ) 0 C, 32 C, 'V' C, 'F' C, 'T' C, 'R' C, '%' C, 0 C,
( 30 ) 0 C, 'N' C, 'B' C, 'H' C, 'G' C, 'Y' C, '^' C, 0 C,
( 38 ) 0 C, 0 C, 'M' C, 'J' C, 'U' C, '&' C, '*' C, 0 C,
( 40 ) 0 C, '<' C, 'K' C, 'I' C, 'O' C, ')' C, '(' C, 0 C,
( 48 ) 0 C, '>' C, '?' C, 'L' C, ':' C, 'P' C, '_' C, 0 C,
( 50 ) 0 C, 0 C, '"' C, 0 C, '{' C, '+' C, 0 C, 0 C,
( 58 ) 0 C, 0 C, 13 C, '}' C, 0 C, '|' C, 0 C, 0 C,
( 60 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 8 C, 0 C,
( 68 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C,
( 70 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 27 C, 0 C,
( 78 ) 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C, 0 C,
( ----- 414 )
: _shift? ( kc -- f ) DUP 0x12 = SWAP 0x59 = OR ;
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
: (key?) ( -- c? f )
(ps2kc) DUP NOT IF EXIT THEN ( kc )
DUP 0xe0 ( extended ) = IF ( ignore ) DROP 0 EXIT THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
DUP 0xf0 ( break ) = IF DROP ( )
( get next kc and see if it's a shift )
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
BEGIN (ps2kc) ?DUP UNTIL ( kc )
_shift? IF ( drop shift ) 0 PS2_SHIFT C! THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( whether we had a shift or not, we return the next )
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
0 EXIT THEN
DUP 0x7f > IF DROP 0 EXIT THEN
DUP _shift? IF DROP 1 PS2_SHIFT C! 0 EXIT THEN
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ah, finally, we have a gentle run-of-the-mill KC )
Make KEY non-blocking ... and rename it to KEY?. Then, add KEY from KEY? for its blocking version. I need this for an upcoming Remote Shell feature. If a Collapse OS system remotely controls another shell, it needs to be able to poll both the remote system and the local keyboard at the same time. A blocking KEY is incompatible with this. In some places, the polling mechanism doesn't make sense, so this new KEY? always returns a character. In some places, I just haven't implemented the mechanism yet, so I kept the old blocking code and added a "always 1" flag as a temporary shim. I have probably broken something, but in emulators, Collapse OS runs fine. It's an important reminder of what will be lost with the new "dogfooding" approach (see recent mailing list message): without emulators, it's much harder to to sweeping changes like this without breaking stuff. It's fine, I don't expect many more of these core changes to the system. It's nearly feature-complete.
2021-01-02 00:05:29 +11:00
PS2_CODES PS2_SHIFT C@ IF 0x80 + THEN + C@ ( c, maybe 0 )
?DUP ( c? f ) ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 418 )
SPI relay driver
This driver is designed for a ad-hoc adapter card that acts as a
SPI relay between the z80 bus and the SPI device. When writing
to SPI_CTL, we expect a bitmask of the device to select, with
0 meaning that everything is de-selected. Reading SPI_CTL
returns 0 if the device is ready or 1 if it's still running an
exchange. Writing to SPI_DATA initiates an exchange.
Provides the SPI relay protocol. Load driver with "419 LOAD".
( ----- 419 )
CODE (spix) ( n -- n )
HL POP, chkPS, A L LDrr,
SPI_DATA OUTiA,
( wait until xchg is done )
BEGIN, SPI_CTL INAi, 1 ANDi, JRNZ, AGAIN,
SPI_DATA INAi,
L A LDrr,
HL PUSH,
;CODE
CODE (spie) ( n -- )
HL POP, chkPS, A L LDrr,
SPI_CTL OUTiA,
;CODE
( ----- 420 )
SD Card subsystem
Load range: B423-B436
This subsystem is designed for a ad-hoc adapter card that acts
as a SPI relay between the z80 bus and the SD card. It requires
a driver providing the SPI Relay protocol. You need to define
SDC_DEVID to specify which ID will be supplied to (spie).
Through that layer, this driver implements the SDC protocol
allowing it to provide BLK@ and BLK!.
( ----- 423 )
( Computes n into crc c with polynomial 0x1021 )
CODE _crc16 ( c n -- c ) EXX, ( protect BC )
HL POP, ( n ) DE POP, ( c )
A L LDrr, D XORr, D A LDrr,
B 8 LDri,
BEGIN,
E SLA, D RL,
IFC, ( msb is set, apply polynomial )
A D LDrr, 0x10 XORi, D A LDrr,
A E LDrr, 0x21 XORi, E A LDrr,
THEN,
DJNZ, AGAIN,
DE PUSH,
EXX, ( unprotect BC ) ;CODE
( ----- 424 )
( -- n )
: _idle 0xff (spix) ;
( -- n )
( spix 0xff until the response is something else than 0xff
for a maximum of 20 times. Returns 0xff if no response. )
: _wait
0 ( dummy ) 20 0 DO
DROP _idle DUP 0xff = NOT IF LEAVE THEN
LOOP ;
( ----- 425 )
( -- )
( The opposite of sdcWaitResp: we wait until response is 0xff.
After a successful read or write operation, the card will be
busy for a while. We need to give it time before interacting
with it again. Technically, we could continue processing on
our side while the card it busy, and maybe we will one day,
but at the moment, I'm having random write errors if I don't
do this right after a write, so I prefer to stay cautious
for now. )
: _ready BEGIN _idle 0xff = UNTIL ;
( ----- 426 )
( c n -- c )
( Computes n into crc c with polynomial 0x09
Note that the result is "left aligned", that is, that 8th
bit to the "right" is insignificant (will be stop bit). )
: _crc7
XOR ( c )
8 0 DO
2 * ( <<1 )
DUP 255 > IF
( MSB was set, apply polynomial )
0xff AND
0x12 XOR ( 0x09 << 1, we apply CRC on high bits )
THEN
LOOP
;
( ----- 427 )
( send-and-crc7 )
( n c -- c )
: _s+crc SWAP DUP (spix) DROP _crc7 ;
( ----- 428 )
( cmd arg1 arg2 -- resp )
( Sends a command to the SD card, along with arguments and
specified CRC fields. (CRC is only needed in initial commands
though). This does *not* handle CS. You have to
select/deselect the card outside this routine. )
: _cmd
_wait DROP ROT ( a1 a2 cmd )
0 _s+crc ( a1 a2 crc )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
ROT |M ROT ( a2 h l crc )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
_s+crc _s+crc ( a2 crc )
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
SWAP |M ROT ( h l crc )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
_s+crc _s+crc ( crc )
1 OR ( ensure stop bit )
(spix) DROP ( send CRC )
_wait ( wait for a valid response... )
;
( ----- 429 )
( cmd arg1 arg2 -- r )
( Send a command that expects a R1 response, handling CS. )
: SDCMDR1 [ SDC_DEVID LITN ] (spie) _cmd 0 (spie) ;
( cmd arg1 arg2 -- r arg1 arg2 )
( Send a command that expects a R7 response, handling CS. A R7
is a R1 followed by 4 bytes. arg1 contains bytes 0:1, arg2
has 2:3 )
: SDCMDR7
[ SDC_DEVID LITN ] (spie)
_cmd ( r )
_idle 8 LSHIFT _idle + ( r arg1 )
_idle 8 LSHIFT _idle + ( r arg1 arg2 )
0 (spie)
;
( ----- 430 )
: _err 0 (spie) LIT" SDerr" ERR ;
( Tight definition ahead, pre-comment.
Initialize a SD card. This should be called at least 1ms
after the powering up of the card. We begin by waking up the
SD card. After power up, a SD card has to receive at least
74 dummy clocks with CS and DI high. We send 80.
Then send cmd0 for a maximum of 10 times, success is when
we get 0x01. Then comes the CMD8. We send it with a 0x01aa
argument and expect a 0x01aa argument back, along with a
0x01 R1 response. After that, we need to repeatedly run
CMD55+CMD41 (0x40000000) until the card goes out of idle
mode, that is, when it stops sending us 0x01 response and
send us 0x00 instead. Any other response means that
initialization failed. )
( ----- 431 )
: SDC$
10 0 DO _idle DROP LOOP
0 ( dummy ) 10 0 DO ( r )
DROP 0x40 0 0 SDCMDR1 ( CMD0 )
1 = DUP IF LEAVE THEN
LOOP NOT IF _err THEN
0x48 0 0x1aa ( CMD8 ) SDCMDR7 ( r arg1 arg2 )
( expected 1 0 0x1aa )
0x1aa = ROT ( arg1 f r ) 1 = AND SWAP ( f&f arg1 )
NOT ( 0 expected ) AND ( f&f&f ) NOT IF _err THEN
BEGIN
0x77 0 0 SDCMDR1 ( CMD55 )
1 = NOT IF _err THEN
0x69 0x4000 0 SDCMDR1 ( CMD41 )
DUP 1 > IF _err THEN
NOT UNTIL ; ( out of idle mode, success! )
( ----- 432 )
: _ ( dstaddr blkno -- )
[ SDC_DEVID LITN ] (spie)
0x51 ( CMD17 ) 0 ROT ( a cmd 0 blkno ) _cmd
IF _err THEN
_wait 0xfe = NOT IF _err THEN
0 SWAP ( crc a )
512 0 DO ( crc a )
_idle ( crc a n )
DUP ROT C!+ ( crc n a+1 )
ROT> _crc16 ( a+1 crc )
SWAP ( crc a+1 )
LOOP
DROP ( crc1 )
_idle 8 LSHIFT _idle + ( crc2 )
_wait DROP 0 (spie)
= NOT IF _err THEN ;
( ----- 433 )
: SDC@
2 * DUP BLK( SWAP ( b a b ) _
1+ BLK( 512 + SWAP _
;
( ----- 434 )
: _ ( srcaddr blkno -- )
[ SDC_DEVID LITN ] (spie)
0x58 ( CMD24 ) 0 ROT ( a cmd 0 blkno ) _cmd
IF _err THEN
_idle DROP 0xfe (spix) DROP 0 SWAP ( crc a )
512 0 DO ( crc a )
C@+ ( crc a+1 n )
ROT OVER ( a n crc n )
_crc16 ( a n crc )
SWAP ( a crc n )
(spix) DROP ( a crc )
SWAP ( crc a )
LOOP
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
DROP ( crc ) |M ( lsb msb )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
(spix) DROP (spix) DROP
_wait DROP 0 (spie) ;
( ----- 435 )
: SDC!
2 * DUP BLK( SWAP ( b a b ) _
1+ BLK( 512 + SWAP _
;
( ----- 440 )
8086 boot code
Code in the following blocks assemble into a binary that is
suitable to plug into Core words (B350) to achieve a fully
functional Collapse OS. It is structured in a way that is
very similar to Z80 boot code (B280) and requires the same
constants to be pre-declared.
RESERVED REGISTERS: SP is reserved for PSP, BP is for RSP and
DX is for IP. Whenever you use these registers for another
purpose, be sure to protect their initial value. Like with
Z80, you can use SP freely in native code, but you have to make
sure it goes back to its previous level before next is called.
(cont.)
( ----- 441 )
Fix mis-documentation in 8086 boot code comments
2020-11-15 12:28:44 +11:00
PS CHECKS: chkPS, is a bit different than in z80: it is para-
metrizable. The idea is that we always call chkPS, before pop-
ping, telling the expected size of stack. This allows for some
interesting optimization. For example, in SWAP, no need to pop,
chkPS, then push, we can chkPS and then proceed to optimized
swapping in PS.
Load range: B445-B461
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 445 )
VARIABLE lblexec VARIABLE lblnext
H@ ORG !
Remove indirect memory access I got bitten again, I've over-designed my solution. The last time it happened, it was that memory mapping thing I was wanting to add. The indirect memory access feature I was adding was to solve a specific problem: Allow Collapse OS to cross-compile directly on a AT28 EEPROM. It began well. As long as we were staying in the assembler realm, things were looking good. However, when we got into the xcomp realm (B260), things became ugly, and I had to creep up indirection where I didn't want to. All of this because I wanted to solve my initial problem in a slightly more generalized way. The broad idea was that these indirect memory access could allow xcomp into a broad kind of memory-like devices. This idea broke on the "@" part of the equation. If I want indirections to be two-way and allow xcomp to work properly, I have to add this indirection to FIND (and possibly others) and this just isn't practical or elegant. So, I'm taking a step back and accepting that the solution I design for now is exclusively for the AT28. What I'm thinking is to add a low-level hook for memory writing, at the assembly level.
2020-12-08 14:18:10 +11:00
JMPn, 0 , ( 00, main ) 0 C, ( 03, boot driveno )
0 , ( 04, BOOT )
0 , ( 06, uflw ) 0 , ( 08, LATEST ) 0 , ( unused )
0 C, 0 , ( 0b, EXIT )
0 , 0 , ( unused ) 0 , ( 13, oflw )
0 , 0 , 0 C, ( unused )
JMPn, 0 , ( 1a, next )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 446 )
( TODO: move these words with other native words. )
H@ 4 + XCURRENT ! ( make next CODE have 0 prev field )
CODE (br) L1 BSET ( used in ?br )
DI DX MOVxx, AL [DI] MOVr[], AH AH XORrr, CBW,
DX AX ADDxx,
;CODE
CODE (?br)
AX POPx, AX AX ORxx, JZ, L1 @ RPCs, ( False, branch )
( True, skip next byte and don't branch )
DX INCx,
;CODE
( ----- 447 )
CODE (loop)
[BP] 0 INC[w]+, ( I++ )
( Jump if I <> I' )
AX [BP] 0 MOVx[]+, AX [BP] -2 CMPx[]+,
JNZ, L1 @ RPCs, ( branch )
( don't branch )
BP 4 SUBxi, DX INCx,
;CODE
( ----- 448 )
lblnext BSET PC 0x1d - ORG @ 0x1b + ! ( next )
( ovfl check )
BP SP CMPxx,
IFNC, ( BP >= SP )
SP PS_ADDR MOVxI, BP RS_ADDR MOVxI,
DI 0x13 ( oflw ) MOVxm, JMPs, L1 FWRs ( execute )
THEN,
DI DX MOVxx, ( <-- IP ) DX INCx, DX INCx,
DI [DI] MOVx[], ( wordref )
( continue to execute ) L1 FSET
( ----- 449 )
lblexec BSET ( DI -> wordref )
AL [DI] MOVr[], DI INCx, ( PFA )
AL AL ORrr, IFZ, DI JMPr, THEN, ( native )
AL DECr, IFNZ, ( not compiled )
AL DECr, IFZ, ( cell )
DI PUSHx, JMPs, lblnext @ RPCs, THEN,
AL DECr, IFZ, ( does )
DI PUSHx, DI INCx, DI INCx, DI [DI] MOVx[], THEN,
Rename switch to ialias The switch name was confusing.
2020-11-29 02:58:16 +11:00
( alias or ialias ) DI [DI] MOVx[],
AL DECr, IFNZ, ( ialias ) DI [DI] MOVx[], THEN,
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
JMPs, lblexec @ RPCs,
THEN, ( continue to compiled )
BP INCx, BP INCx, [BP] 0 DX MOV[]+x, ( pushRS )
DX DI MOVxx, DX INCx, DX INCx, ( --> IP )
DI [DI] MOVx[], JMPs, lblexec @ RPCs,
( ----- 450 )
lblchkPS BSET ( CX -> expected size )
AX PS_ADDR MOVxI, AX SP SUBxx, 2 SUBAXI, ( CALL adjust )
AX CX CMPxx,
IFNC, ( we're good ) RET, THEN,
( underflow ) DI 0x06 MOVxm, JMPs, lblexec @ RPCs,
PC 3 - ORG @ 1+ ! ( main )
DX POPx, ( boot drive no ) 0x03 DL MOVmr,
SP PS_ADDR MOVxI, BP RS_ADDR MOVxI,
DI 0x08 MOVxm, ( LATEST )
( HERE begins at CURRENT )
SYSVARS 0x4 ( HERE ) + DI MOVmx,
SYSVARS 0x2 ( CURRENT ) + DI MOVmx,
DI 0x04 ( BOOT ) MOVxm,
JMPn, lblexec @ RPCn, ( execute )
( ----- 451 )
( native words )
CODE EXECUTE 1 chkPS,
DI POPx, JMPn, lblexec @ RPCn,
CODE EXIT
DX [BP] 0 MOVx[]+, BP DECx, BP DECx, ( popRS )
;CODE
( ----- 452 )
CODE (n) ( number literal )
DI DX MOVxx, DI [DI] MOVx[], DI PUSHx,
DX INCx, DX INCx,
;CODE
CODE (s) ( string literal, see B287 )
DI DX MOVxx, ( IP )
AH AH XORrr, AL [DI] MOVr[], ( slen )
DX PUSHx, DX INCx, DX AX ADDxx,
;CODE
( ----- 453 )
CODE >R 1 chkPS,
BP INCx, BP INCx, [BP] 0 POP[w]+,
;CODE NOP, NOP, NOP,
CODE R>
[BP] 0 PUSH[w]+, BP DECx, BP DECx,
;CODE
CODE 2>R
[BP] 4 POP[w]+, [BP] 2 POP[w]+, BP 4 ADDxi,
;CODE
CODE 2R> 2 chkPS,
[BP] -2 PUSH[w]+, [BP] 0 PUSH[w]+, BP 4 SUBxi,
;CODE
( ----- 454 )
CODE ROT ( a b c -- b c a ) 3 chkPS,
CX POPx, BX POPx, AX POPx,
BX PUSHx, CX PUSHx, AX PUSHx, ;CODE
CODE ROT> ( a b c -- c a b ) 3 chkPS,
CX POPx, BX POPx, AX POPx,
CX PUSHx, AX PUSHx, BX PUSHx, ;CODE
CODE DUP 1 chkPS, AX POPx, AX PUSHx, AX PUSHx, ;CODE
CODE ?DUP 1 chkPS, AX POPx, AX AX ORxx, AX PUSHx,
IFNZ, AX PUSHx, THEN, ;CODE
CODE OVER ( a b -- a b a ) 2 chkPS,
DI SP MOVxx, AX [DI] 2 MOVx[]+, AX PUSHx, ;CODE
CODE PICK
DI POPx, DI SHLx1, ( x2 )
CX DI MOVxx, CX 2 ADDxi, CALL, lblchkPS @ RPCn,
DI SP ADDxx, DI [DI] MOVx[], DI PUSHx,
;CODE
( ----- 455 )
CODE (roll) ( "2 3 4 5 4 --> 2 4 5 5". See B311 )
CX POPx, CX 2 ADDxi, CALL, lblchkPS @ RPCn, CX 2 SUBxi,
SI SP MOVxx, SI CX ADDxx,
DI SI MOVxx, DI 2 ADDxi, STD, REPZ, MOVSB,
;CODE
CODE SWAP AX POPx, BX POPx, AX PUSHx, BX PUSHx, ;CODE
CODE DROP 1 chkPS, AX POPx, ;CODE
CODE 2DROP 2 chkPS, SP 4 ADDxi, ;CODE
CODE 2DUP 2 chkPS,
AX POPx, BX POPx,
BX PUSHx, AX PUSHx, BX PUSHx, AX PUSHx,
;CODE
CODE S0 AX PS_ADDR MOVxI, AX PUSHx, ;CODE
CODE 'S SP PUSHx, ;CODE
CODE AND 2 chkPS,
AX POPx, BX POPx, AX BX ANDxx, AX PUSHx, ;CODE
( ----- 456 )
CODE OR 2 chkPS,
AX POPx, BX POPx, AX BX ORxx, AX PUSHx, ;CODE
CODE XOR 2 chkPS,
AX POPx, BX POPx, AX BX XORxx, AX PUSHx, ;CODE
CODE NOT 1 chkPS,
AX POPx, AX AX ORxx,
IFNZ, AX -1 MOVxI, THEN, AX INCx, AX PUSHx, ;CODE
CODE + 2 chkPS,
AX POPx, BX POPx, AX BX ADDxx, AX PUSHx, ;CODE
CODE - 2 chkPS,
BX POPx, AX POPx, AX BX SUBxx, AX PUSHx, ;CODE
CODE * 2 chkPS,
AX POPx, BX POPx,
DX PUSHx, ( protect from MUL ) BX MULx, DX POPx,
AX PUSHx, ;CODE
( ----- 457 )
CODE /MOD 2 chkPS,
BX POPx, AX POPx, DX PUSHx, ( protect )
DX DX XORxx, BX DIVx,
BX DX MOVxx, DX POPx, ( unprotect )
BX PUSHx, ( modulo ) AX PUSHx, ( division )
;CODE
CODE ! 2 chkPS, DI POPx, AX POPx, [DI] AX MOV[]x, ;CODE
CODE @ 1 chkPS, DI POPx, AX [DI] MOVx[], AX PUSHx, ;CODE
CODE C! 2 chkPS, DI POPx, AX POPx, [DI] AX MOV[]r, ;CODE
CODE C@ 1 chkPS,
DI POPx, AH AH XORrr, AL [DI] MOVr[], AX PUSHx, ;CODE
CODE I [BP] 0 PUSH[w]+, ;CODE
CODE I' [BP] -2 PUSH[w]+, ;CODE
CODE J [BP] -4 PUSH[w]+, ;CODE
CODE (resSP) SP PS_ADDR MOVxI, ;CODE
CODE (resRS) BP RS_ADDR MOVxI, ;CODE
( ----- 458 )
CODE BYE HLT, BEGIN, JMPs, AGAIN, ;CODE
CODE S= 2 chkPS,
SI POPx, DI POPx, CH CH XORrr, CL [SI] MOVr[],
CL [DI] CMPr[],
IFZ, ( same size? )
SI INCx, DI INCx, CLD, REPZ, CMPSB,
THEN,
PUSHZ,
;CODE
CODE CMP 2 chkPS,
BX POPx, AX POPx, CX CX XORxx, AX BX CMPxx,
IFNZ, ( < or > )
CX INCx, IFC, ( < ) CX DECx, CX DECx, THEN,
THEN,
CX PUSHx,
;CODE
( ----- 459 )
CODE _find ( cur w -- a f ) 2 chkPS,
SI POPx, ( w ) DI POPx, ( cur )
CH CH XORrr, CL [SI] MOVr[], ( CX -> strlen )
SI INCx, ( first char ) AX AX XORxx, ( initial prev )
BEGIN, ( loop )
DI AX SUBxx, ( jump to prev wordref )
AL [DI] -1 MOVr[]+, 0x7f ANDALi, ( strlen )
CL AL CMPrr, IFZ, ( same len )
SI PUSHx, DI PUSHx, CX PUSHx, ( --> lvl 3 )
3 ADDALi, ( header ) AH AH XORrr, DI AX SUBxx,
CLD, REPZ, CMPSB,
CX POPx, DI POPx, SI POPx, ( <-- lvl 3 )
IFZ, DI PUSHx, AX 1 MOVxI, AX PUSHx,
JMPn, lblnext @ RPCn, THEN,
THEN,
DI 3 SUBxi, AX [DI] MOVx[], ( prev ) AX AX ORxx, ( cont. )
( ----- 460 )
( cont. find ) JNZ, AGAIN, ( loop )
SI DECx, SI PUSHx, AX AX XORrr, AX PUSHx,
;CODE
CODE 0 AX AX XORxx, AX PUSHx, ;CODE
CODE 1 AX 1 MOVxI, AX PUSHx, ;CODE
CODE -1 AX -1 MOVxI, AX PUSHx, ;CODE
CODE 1+ 1 chkPS, DI SP MOVxx, [DI] INC[w], ;CODE
CODE 1- 1 chkPS, DI SP MOVxx, [DI] DEC[w], ;CODE
CODE 2+ 1 chkPS, DI SP MOVxx, [DI] INC[w], [DI] INC[w], ;CODE
CODE 2- 1 chkPS, DI SP MOVxx, [DI] DEC[w], [DI] DEC[w], ;CODE
CODE RSHIFT ( n u -- n ) 2 chkPS,
CX POPx, AX POPx, AX SHRxCL, AX PUSHx, ;CODE
CODE LSHIFT ( n u -- n ) 2 chkPS,
CX POPx, AX POPx, AX SHLxCL, AX PUSHx, ;CODE
( ----- 461 )
( See comment in B321. TODO: test on real hardware. in qemu,
the resulting delay is more than 10x too long. )
CODE TICKS 1 chkPS, ( n=100us )
SI DX MOVxx, ( protect IP )
AX POPx, BX 100 MOVxI, BX MULx,
CX DX MOVxx, ( high ) DX AX MOVxx, ( low )
AX 0x8600 MOVxI, ( 86h, WAIT ) 0x15 INT,
DX SI MOVxx, ( restore IP )
;CODE
Add words |M and |L Splitting a word into MSB/LSB pairs happens often and is worth, I think, native words. Also, I'm going to need it in the upcoming commits.
2020-12-08 12:08:45 +11:00
CODE |M ( n -- lsb msb ) 1 chkPS,
CX POPx, AH 0 MOVri,
AL CL MOVrr, AX PUSHx, AL CH MOVrr, AX PUSHx, ;CODE
CODE |L ( n -- msb lsb ) 1 chkPS,
CX POPx, AH 0 MOVri,
AL CH MOVrr, AX PUSHx, AL CL MOVrr, AX PUSHx, ;CODE
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( ----- 470 )
( Z80 driver for TMS9918. Implements grid protocol. Requires
TMS_CTLPORT, TMS_DATAPORT and ~FNT from the Font compiler at
sms: fix cursor mis-display in text mode In CURSOR!, I was using a write commande to read from VRAM and the emulator didn't properly behave and did as if everything was fine. The result on a real SMS was that the cursor would contain the inverted glyph of the contents of the *old* cursor position.
2020-11-17 05:10:04 +11:00
B520. Patterns are at addr 0x0000, Names are at 0x3800.
Load range B470-472 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
CODE _ctl ( a -- sends LSB then MSB )
HL POP, chkPS,
A L LDrr, TMS_CTLPORT OUTiA,
A H LDrr, TMS_CTLPORT OUTiA,
;CODE
CODE _data
HL POP, chkPS,
A L LDrr, TMS_DATAPORT OUTiA,
;CODE
( ----- 471 )
sms: add cursor indicator to text mode I do this by adding an inverted version of all glyphs in the upper range of the pattern memory. This is a big waste of the pattern space (only one inverted character is needed at once), but it's the simplest way to proceed. I'll change this if I ever need more pattern space. Also, remove _blank. It was wastefully blanking the whole memory. Only the name table needs to be blanked on initialization.
2020-11-17 03:11:30 +11:00
: _zero ( x -- send 0 _data x times )
( x ) 0 DO 0 _data LOOP ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
( Each row in ~FNT is a row of the glyph and there is 7 of
them. We insert a blank one at the end of those 7. )
: _sfont ( a -- Send font to TMS )
7 0 DO C@+ _data LOOP DROP
( blank row ) 0 _data ;
sms: add cursor indicator to text mode I do this by adding an inverted version of all glyphs in the upper range of the pattern memory. This is a big waste of the pattern space (only one inverted character is needed at once), but it's the simplest way to proceed. I'll change this if I ever need more pattern space. Also, remove _blank. It was wastefully blanking the whole memory. Only the name table needs to be blanked on initialization.
2020-11-17 03:11:30 +11:00
: _sfont^ ( a -- Send inverted font to TMS )
7 0 DO C@+ 0xff XOR _data LOOP DROP
( blank row ) 0xff _data ;
grid: change the meaning of CELL! Replace the "g" arg (glyph) with "c" (character). The reason why "g" was used was to save a "0x20 -" operation at all CELL! implementations, but this came with too big a drawback: it made CELL! hardly usable outside of the Grid subsystem, mostly because the user of CELL! would often have to do "0x20 -". For example, I want the SMS's Pad driver to use CELL! directly instead of having to do EMIT+XYPOS-messing-around. I would have had to do a "0x20 -" there.
2020-11-17 01:11:47 +11:00
: CELL! ( c pos )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x7800 OR _ctl ( tilenum )
Add BS, CR, LF, SPC ASCII consts Previously, these words would be ascii emitters, but seldom used except for the SPC emitter. However, I would often end up hardcoding these constants. With useless emitters removed and ASCII constants added, we have a more usable system. Also, fix broken test harness.
2021-01-03 02:19:42 +11:00
SPC - ( glyph ) 0x5f MOD _data ;
sms: add cursor indicator to text mode I do this by adding an inverted version of all glyphs in the upper range of the pattern memory. This is a big waste of the pattern space (only one inverted character is needed at once), but it's the simplest way to proceed. I'll change this if I ever need more pattern space. Also, remove _blank. It was wastefully blanking the whole memory. Only the name table needs to be blanked on initialization.
2020-11-17 03:11:30 +11:00
( ----- 472 )
: CURSOR! ( new old -- )
sms: fix cursor mis-display in text mode In CURSOR!, I was using a write commande to read from VRAM and the emulator didn't properly behave and did as if everything was fine. The result on a real SMS was that the cursor would contain the inverted glyph of the contents of the *old* cursor position.
2020-11-17 05:10:04 +11:00
DUP 0x3800 OR _ctl [ TMS_DATAPORT LITN ] PC@
0x7f AND ( new old glyph ) SWAP 0x7800 OR _ctl _data
DUP 0x3800 OR _ctl [ TMS_DATAPORT LITN ] PC@
0x80 OR ( new glyph ) SWAP 0x7800 OR _ctl _data ;
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
: COLS 40 ; : LINES 24 ;
: TMS$
sms: add cursor indicator to text mode I do this by adding an inverted version of all glyphs in the upper range of the pattern memory. This is a big waste of the pattern space (only one inverted character is needed at once), but it's the simplest way to proceed. I'll change this if I ever need more pattern space. Also, remove _blank. It was wastefully blanking the whole memory. Only the name table needs to be blanked on initialization.
2020-11-17 03:11:30 +11:00
0x8100 _ctl ( blank screen )
0x7800 _ctl COLS LINES * _zero
tms9918: fix character range which was too narrow We couldn't get a tilde!
2020-12-13 09:28:32 +11:00
0x4000 _ctl 0x5f 0 DO ~FNT I 7 * + _sfont LOOP
0x4400 _ctl 0x5f 0 DO ~FNT I 7 * + _sfont^ LOOP
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x820e _ctl ( name table 0x3800 )
grid: change the meaning of CELL! Replace the "g" arg (glyph) with "c" (character). The reason why "g" was used was to save a "0x20 -" operation at all CELL! implementations, but this came with too big a drawback: it made CELL! hardly usable outside of the Grid subsystem, mostly because the user of CELL! would often have to do "0x20 -". For example, I want the SMS's Pad driver to use CELL! directly instead of having to do EMIT+XYPOS-messing-around. I would have had to do a "0x20 -" there.
2020-11-17 01:11:47 +11:00
0x8400 _ctl ( pattern table 0x0000 )
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
0x87f0 _ctl ( colors 0 and 1 )
0x8000 _ctl 0x81d0 _ctl ( text mode, display on ) ;
( ----- 520 )
Fonts
Fonts are kept in "source" form in the following blocks and
then compiled to binary bitmasks by the following code. In
source form, fonts are a simple sequence of '.' and 'X'. '.'
means empty, 'X' means filled. Glyphs are entered one after the
other, starting at 0x21 and ending at 0x7e. To be space
efficient in blocks, we align glyphs horizontally in the blocks
to fit as many character as we can. For example, a 5x7 font
would mean that we would have 12x2 glyphs per block.
521 Font compiler 530 3x5 font
532 5x7 font 536 7x7 font
( ----- 521 )
( Converts "dot-X" fonts to binary "glyph rows". One byte for
each row. In a 5x7 font, each glyph thus use 7 bytes.
Resulting bytes are aligned to the left of the byte.
Therefore, for a 5-bit wide char, "X.X.X" translates to
0b10101000. Left-aligned bytes are easier to work with when
compositing glyphs. )
( ----- 522 )
: _g ( given a top-left of dot-X in BLK(, spit 5 bin lines )
5 0 DO
0 3 0 DO ( a r )
1 LSHIFT
OVER J 64 * I + + C@ 'X' = IF 1+ THEN
LOOP 5 LSHIFT C, LOOP DROP ;
: _l ( a u -- a, spit a line of u glyphs )
( u ) 0 DO ( a )
DUP I 3 * + _g
LOOP ;
: CPFNT3x5
0 , 0 , 0 C, ( space char )
530 BLK@ BLK( 21 _l 320 + 21 _l 320 + 21 _l DROP ( 63 )
531 BLK@ BLK( 21 _l 320 + 10 _l DROP ( 94! )
;
( ----- 523 )
: _g ( given a top-left of dot-X in BLK(, spit 7 bin lines )
7 0 DO
0 5 0 DO ( a r )
1 LSHIFT
OVER J 64 * I + + C@ 'X' = IF 1+ THEN
LOOP 3 LSHIFT C, LOOP DROP ;
: _l ( a u -- a, spit a line of u glyphs )
( u ) 0 DO ( a )
DUP I 5 * + _g
LOOP ;
: CPFNT5x7
0 , 0 , 0 , 0 C, ( space char )
535 532 DO I BLK@ BLK( 12 _l 448 + 12 _l DROP LOOP ( 72 )
535 BLK@ BLK( 12 _l 448 + 10 _l DROP ( 94! )
;
( ----- 524 )
: _g ( given a top-left of dot-X in BLK(, spit 7 bin lines )
7 0 DO
0 7 0 DO ( a r )
1 LSHIFT
OVER J 64 * I + + C@ 'X' = IF 1+ THEN
LOOP 1 LSHIFT C, LOOP DROP ;
: _l ( a u -- a, spit a line of u glyphs )
( u ) 0 DO ( a )
DUP I 7 * + _g
LOOP ;
: CPFNT7x7
0 , 0 , 0 , 0 C, ( space char )
541 536 DO I BLK@ BLK( 9 _l 448 + 9 _l DROP LOOP ( 90 )
542 BLK@ BLK( 4 _l DROP ( 94! )
;
( ----- 530 )
.X.X.XX.X.XXX...X..X...XX...X...............X.X..X.XX.XX.X.XXXX
.X.X.XXXXXX...XX.X.X..X..X.XXX.X............XX.XXX...X..XX.XX..
.X........XX.X..X.....X..X..X.XXX...XXX....X.X.X.X..X.XX.XXXXX.
......XXXXX.X..X.X....X..X.X.X.X..X.......X..X.X.X.X....X..X..X
.X....X.X.X...X.XX.....XX........X......X.X...X.XXXXXXXX...XXX.
.XXXXXXXXXXX........X...X..XX..X..X.XX..XXXX.XXXXXX.XXX.XXXXXXX
X....XX.XX.X.X..X..X.XXX.X...XXXXX.XX.XX..X.XX..X..X..X.X.X...X
XXX.X.XXXXXX......X.......X.X.XXXXXXXX.X..X.XXX.XX.X.XXXX.X...X
X.XX..X.X..X.X..X..X.XXX.X....X..X.XX.XX..X.XX..X..X.XX.X.X...X
XXXX..XXXXX....X....X...X...X..XXX.XXX..XXXX.XXXX...XXX.XXXXXX.
X.XX..X.XXX.XXXXX.XXXXX..XXXXXX.XX.XX.XX.XX.XXXXXXXX..XXX.X....
XX.X..XXXX.XX.XX.XX.XX.XX...X.X.XX.XX.XX.XX.X..XX..X....XX.X...
X..X..XXXX.XX.XXX.X.XXX..X..X.X.XX.XXXX.X..X..X.X...X...X......
XX.X..X.XX.XX.XX..XXXX.X..X.X.X.XX.XXXXX.X.X.X..X....X..X......
X.XXXXX.XX.XXXXX...XXX.XXX..X.XXX.X.X.XX.X.X.XXXXXX..XXXX...XXX
!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_
( ----- 531 )
X.....X.......X....XX...X...X...XX..XX.......................X.
.X.XX.X...XX..X.X.X...X.X........X.X.X.X.XXX..X.XX..XX.XX.XXXXX
.....XXX.X...XXX.XXX.X.XXX..X...XXX..X.XXXX.XX.XX.XX.XX..XX..X.
...XXXX.XX..X.XXX.X...XXX.X.X...XX.X.X.X.XX.XX.XXX..XXX....X.X.
...XXXXX..XX.XX.XXX..XX.X.X.X.XX.X.X.XXX.XX.X.X.X....XX..XX..XX
...................XX.X.XX.....................................
X.XX.XX.XX.XX.XXXX.X..X..X..XX
X.XX.XX.X.X..X..XXX...X...XXX.
X.XX.XXXX.X..X.XX..X..X..X....
XXX.X.X.XX.X.X.XXX.XX.X.XX....
`abcdefghijklmnopqrstuvwxyz{|}~
( ----- 532 )
Fix "," char in 5x7 font
2020-11-18 03:24:51 +11:00
..X...X.X........X..............X....X....X.................
..X...X.X..X.X..XXXXX...X.XX....X...X......X.X.X.X..X.......
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
..X.......XXXXXX.......X.X..X......X........X.XXX...X.......
..X........X.X..XXX...X...XX.......X........XXXXXXXXXXX.....
..........XXXXX....X.X....XX.X.....X........X.XXX...X.......
Fix "," char in 5x7 font
2020-11-18 03:24:51 +11:00
..X........X.X.XXXX.X...XX..X.......X......X.X.X.X..X.....X.
..X..............X.......XXX.X.......X....X..............X..
Transform "blk/" folders into "blk.fs" text files Working in "blk/" folder from a modern system is harder than it should be. Moving blocks around is a bit awkward, grepping is a bit less convenient than it could be, git blame has troubles following, etc. In this commit, we modify blkpack and blkunpack to work with single text files with blocks being separated by a special markup. I think this will make the code significantly more convenient to work into.
2020-11-15 10:00:31 +11:00
................XXX...XX..XXX..XXX...XX.XXXXX.XXX.XXXXX.XXX.
..............XX...X.X.X.X...XX...X.X.X.X....X........XX...X
.............X.X..XX...X.....X....XX..X.XXXX.X........XX...X
XXXXX.......X..X.X.X...X....X...XX.XXXXX....XXXXX....X..XXX.
...........X...XX..X...X...X......X...X.....XX...X..X..X...X
......XX..X....X...X...X..X...X...X...X.X...XX...X.X...X...X
......XX........XXX..XXXXXXXXX.XXX....X..XXX..XXX.X.....XXX.
!"#$%&'()*+,-./012345678
( ----- 533 )
.XXX...............X.....X.....XXX..XXX..XXX.XXXX..XXX.XXXX.
X...X..X....X....XX.......XX..X...XX...XX...XX...XX...XX...X
X...X..X....X...XX..XXXXX..XX.....XX..XXX...XX...XX....X...X
.XXX...........X.............X...X.X..XXXXXXXXXXX.X....X...X
....X..X....X...XX..XXXXX..XX...X..X....X...XX...XX....X...X
....X..X...X.....XX.......XX.......X...XX...XX...XX...XX...X
.XXX...............X.....X......X...XXX.X...XXXXX..XXX.XXXX.
XXXXXXXXXX.XXX.X...X.XXX....XXX..X.X....X...XX...X.XXX.XXXX.
X....X....X...XX...X..X......XX.X..X....XX.XXXX..XX...XX...X
X....X....X....X...X..X......XXX...X....X.X.XXX..XX...XX...X
XXXX.XXXX.X..XXXXXXX..X......XX....X....X...XX.X.XX...XXXXX.
X....X....X...XX...X..X......XXX...X....X...XX..XXX...XX....
X....X....X...XX...X..X..X...XX.X..X....X...XX..XXX...XX....
XXXXXX.....XXX.X...X.XXX..XXX.X..X.XXXXXX...XX...X.XXX.X....
9:;<=>?@ABCDEFGHIJKLMNOP
( ----- 534 )
.XXX.XXXX..XXX.XXXXXX...XX...XX...XX...XX...XXXXXXXXX.......
X...XX...XX...X..X..X...XX...XX...XX...XX...XX...XX....X....
X...XX...XX......X..X...XX...XX...X.X.X..X.X....X.X.....X...
X...XXXXX..XXX...X..X...XX...XX...X..X....X....X..X......X..
X.X.XX.X......X..X..X...XX...XX.X.X.X.X...X...X...X.......X.
X..XXX..X.X...X..X..X...X.X.X.X.X.XX...X..X..X...XX........X
.XXXXX...X.XXX...X...XXX...X...X.X.X...X..X..XXXXXXXX.......
..XXX..X.........X..........................................
....X.X.X.........X.........................................
....XX...X...........XXX.X.....XXX.....X.XXX..XX....XXXX....
....X...................XX....X...X....XX...XX..X..X..XX....
....X................XXXXXXX..X......XXXXXXXXX......XXXXXX..
....X...............X...XX..X.X...X.X..XX....XXX......XX..X.
..XXX.....XXXXX......XXXXXXX...XXX...XXX.XXXXX......XX.X..X.
QRSTUVWXYZ[\]^_`abcdefgh
( ----- 535 )
............................................................
............................................................
..X......XX..X..XX...X.X.XXX...XXX.XXX....XXXX.XX..XXX..X...
..........X.X....X..X.X.XX..X.X...XX..X..X..XXX...X....XXX..
..X......XXX.....X..X...XX...XX...XXXX....XXXX.....XXX..X...
..X...X..XX.X....X..X...XX...XX...XX........XX........X.X...
..X....XX.X..X...XX.X...XX...X.XXX.X........XX.....XXX...XX.
................................XX...X...XX.......
...............................X.....X.....X......
X...XX...XX...XX...XX...XXXXXX.X.....X.....X..X.X.
X...XX...XX...X.X.X..X.X....X.X......X......XX.X..
X...XX...XX...X..X....X....X...X.....X.....X......
X...X.X.X.X.X.X.X.X..X....X....X.....X.....X......
.XXX...X...X.X.X...XX....XXXXX..XX...X...XX.......
ijklmnopqrstuvwxyz{|}~
( ----- 536 )
..XX....XX.XX..XX.XX....XX..XX......XXX......XX.....XX...XX....
..XX....XX.XX..XX.XX..XXXXXXXX..XX.XX.XX....XX.....XX.....XX...
..XX....XX.XX.XXXXXXXXX.X......XX..XX.XX...XX.....XX.......XX..
..XX...........XX.XX..XXXXX...XX....XXX...........XX.......XX..
..XX..........XXXXXXX...X.XX.XX....XX.XX.X........XX.......XX..
...............XX.XX.XXXXXX.XX..XX.XX..XX..........XX.....XX...
..XX...........XX.XX...XX.......XX..XXX.XX..........XX...XX....
...........................................XXXX....XX....XXXX..
..XX.....XX............................XX.XX..XX..XXX...XX..XX.
XXXXXX...XX...........................XX..XX.XXX...XX.......XX.
.XXXX..XXXXXX........XXXXXX..........XX...XXXXXX...XX......XX..
XXXXXX...XX.........................XX....XXX.XX...XX.....XX...
..XX.....XX.....XX............XX...XX.....XX..XX...XX....XX....
...............XX.............XX...........XXXX..XXXXXX.XXXXXX.
!"#$%&'()*+,-./012
( ----- 537 )
.XXXX.....XX..XXXXXX...XXX..XXXXXX..XXXX...XXXX................
XX..XX...XXX..XX......XX........XX.XX..XX.XX..XX...............
....XX..XXXX..XXXXX..XX........XX..XX..XX.XX..XX...XX.....XX...
..XXX..XX.XX......XX.XXXXX....XX....XXXX...XXXXX...XX.....XX...
....XX.XXXXXX.....XX.XX..XX..XX....XX..XX.....XX...............
XX..XX....XX..XX..XX.XX..XX..XX....XX..XX....XX....XX.....XX...
.XXXX.....XX...XXXX...XXXX...XX.....XXXX...XXX.....XX....XX....
...XX.........XX......XXXX...XXXX...XXXX..XXXXX...XXXX..XXXX...
..XX...........XX....XX..XX.XX..XX.XX..XX.XX..XX.XX..XX.XX.XX..
.XX....XXXXXX...XX......XX..XX.XXX.XX..XX.XX..XX.XX.....XX..XX.
XX...............XX....XX...XX.X.X.XXXXXX.XXXXX..XX.....XX..XX.
.XX....XXXXXX...XX.....XX...XX.XXX.XX..XX.XX..XX.XX.....XX..XX.
..XX...........XX...........XX.....XX..XX.XX..XX.XX..XX.XX.XX..
...XX.........XX.......XX....XXXX..XX..XX.XXXXX...XXXX..XXXX...
3456789:;<=>?@ABCD
( ----- 538 )
XXXXXX.XXXXXX..XXXX..XX..XX.XXXXXX..XXXXX.XX..XX.XX.....XX...XX
XX.....XX.....XX..XX.XX..XX...XX......XX..XX.XX..XX.....XXX.XXX
XX.....XX.....XX.....XX..XX...XX......XX..XXXX...XX.....XXXXXXX
XXXXX..XXXXX..XX.XXX.XXXXXX...XX......XX..XXX....XX.....XX.X.XX
XX.....XX.....XX..XX.XX..XX...XX......XX..XXXX...XX.....XX.X.XX
XX.....XX.....XX..XX.XX..XX...XX...XX.XX..XX.XX..XX.....XX...XX
XXXXXX.XX......XXXX..XX..XX.XXXXXX..XXX...XX..XX.XXXXXX.XX...XX
XX..XX..XXXX..XXXXX...XXXX..XXXXX...XXXX..XXXXXX.XX..XX.XX..XX.
XX..XX.XX..XX.XX..XX.XX..XX.XX..XX.XX..XX...XX...XX..XX.XX..XX.
XXX.XX.XX..XX.XX..XX.XX..XX.XX..XX.XX.......XX...XX..XX.XX..XX.
XXXXXX.XX..XX.XXXXX..XX..XX.XXXXX...XXXX....XX...XX..XX.XX..XX.
XX.XXX.XX..XX.XX.....XX.X.X.XX.XX......XX...XX...XX..XX.XX..XX.
XX..XX.XX..XX.XX.....XX.XX..XX..XX.XX..XX...XX...XX..XX..XXXX..
XX..XX..XXXX..XX......XX.XX.XX..XX..XXXX....XX....XXXX....XX...
EFGHIJKLMNOPQRSTUVWXYZ[\]^_
( ----- 539 )
XX...XXXX..XX.XX..XX.XXXXXX.XXXXX.........XXXXX....XX..........
XX...XXXX..XX.XX..XX.....XX.XX.....XX........XX...XXXX.........
XX.X.XX.XXXX..XX..XX....XX..XX......XX.......XX..XX..XX........
XX.X.XX..XX....XXXX....XX...XX.......XX......XX..X....X........
XXXXXXX.XXXX....XX....XX....XX........XX.....XX................
XXX.XXXXX..XX...XX...XX.....XX.........XX....XX................
XX...XXXX..XX...XX...XXXXXX.XXXXX.........XXXXX.........XXXXXXX
.XX...........XX................XX..........XXX.........XX.....
..XX..........XX................XX.........XX.....XXXX..XX.....
...XX...XXXX..XXXXX...XXXX...XXXXX..XXXX...XX....XX..XX.XXXXX..
...........XX.XX..XX.XX..XX.XX..XX.XX..XX.XXXXX..XX..XX.XX..XX.
........XXXXX.XX..XX.XX.....XX..XX.XXXXXX..XX.....XXXXX.XX..XX.
.......XX..XX.XX..XX.XX..XX.XX..XX.XX......XX........XX.XX..XX.
........XXXXX.XXXXX...XXXX...XXXXX..XXXX...XX.....XXX...XX..XX.
WXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~
( ----- 540 )
..XX.....XX...XX......XXX......................................
..............XX.......XX......................................
.XXX....XXX...XX..XX...XX....XX.XX.XXXXX...XXXX..XXXXX...XXXXX.
..XX.....XX...XX.XX....XX...XXXXXXXXX..XX.XX..XX.XX..XX.XX..XX.
..XX.....XX...XXXX.....XX...XX.X.XXXX..XX.XX..XX.XX..XX.XX..XX.
..XX.....XX...XX.XX....XX...XX.X.XXXX..XX.XX..XX.XXXXX...XXXXX.
.XXXX..XX.....XX..XX..XXXX..XX...XXXX..XX..XXXX..XX.........XX.
...............XX..............................................
...............XX..............................................
XX.XX...XXXXX.XXXXX..XX..XX.XX..XX.XX...XXXX..XX.XX..XX.XXXXXX.
XXX.XX.XX......XX....XX..XX.XX..XX.XX.X.XX.XXXX..XX..XX....XX..
XX......XXXX...XX....XX..XX.XX..XX.XX.X.XX..XX...XX..XX...XX...
XX.........XX..XX....XX..XX..XXXX..XXXXXXX.XXXX...XXXXX..XX....
XX.....XXXXX....XXX...XXXXX...XX....XX.XX.XX..XX.....XX.XXXXXX.
ijklmnopqrstuvwxyz{|}~
( ----- 541 )
...XX....XX...XX......XX...X
..XX.....XX....XX....XX.X.XX
..XX.....XX....XX....X...XX.
XXX......XX.....XXX.........
..XX.....XX....XX...........
..XX.....XX....XX...........
...XX....XX...XX............
{|}~
Reference in New Issue Copy Permalink