xdt99: New TI 99 cross-development tools available

[Asmusr]

Thanks for looking into this. I think I would prefer option 2 or 3, but I'm not sure what the difference is. If 3 would mean that only macro calls had to be changed, that would be my preferred.

 

Another thing with requs is that

tmp0:  requ 0

doesn't work with a colon after the label.

 

And a question: is it possible to have separate name spaces within the same IDEA project, e.g. if I would like to be able to have identical labels for different game levels (assembled separately)?

Edited April 7, 2023 by Asmusr

[Asmusr]

Me again. How can you calculate the MSB and LSB of a label greater than >8000? I thought you could always use /256 and %256, but the second example below gives the wrong answer. I need this in order to generate low endian data.

 

0526               test1:
0527                      equ  >2090
0528 EA88 2090            byte test1/256,test1%256
0529 EA8A 2090            data test1
0530               test2:
0531                      equ  >a020
0532 EA8C A1E0            byte test2/256,test2%256
0533 EA8E A020            data test2

 

[Asmusr]

2 hours ago, Asmusr said:

How can you calculate the MSB and LSB of a label greater than >8000?

To answer my own question, it looks like this is working:

byte (test2&>ff00)/256,test2&>00ff

Edited April 7, 2023 by Asmusr

[ralphb]

5 hours ago, Asmusr said:

To answer my own question, it looks like this is working:

byte (test2&>ff00)/256,test2&>00ff

Are you sure about that?  I implemented signed multiplication and division, so that still shouldn't work.

 

But I guess we need unsigned operations as well, and while I'm at it, shift operations as well.  I hadn't added shift operators yet because >> conflicts with the hex prefix >.  But I'll figure something out.

 

EDIT: OK, that sentence made no sense before.

Edited April 7, 2023 by ralphb

[ralphb]

On 4/5/2023 at 6:59 PM, Asmusr said:

Another thing with requs is that

tmp0:  requ 0

doesn't work with a colon after the label.

 

And a question: is it possible to have separate name spaces within the same IDEA project, e.g. if I would like to be able to have identical labels for different game levels (assembled separately)?

That's odd -- I'm getting a different value if the colon is present.  I'll have to check this.

 

I'll have to do some research if a Namespace-like feature is possible, as IDEA treats all files as one project.  The duplicate label, unused label, etc. checks can be implemented per file, but I'm not sure about more basic things like code completion, go-to-definition and so on.  I'll check this, too.

[ralphb]

I've pushed new versions of xas99 and xga99 with support for unsigned division // and %% and logical shift operators >> and <<.  I also fixed the REQU bug.

 

I'm still working on the IDEA macro issue, which might take some time. But in the meantime, I could create a plugin that favors labels over register aliases, if you prefer.

[GDMike]

You are always all over the fixes. Ty

Edited April 9, 2023 by GDMike

[Asmusr]

20 hours ago, ralphb said:

I'm still working on the IDEA macro issue, which might take some time. But in the meantime, I could create a plugin that favors labels over register aliases, if you prefer.

Thank you, but it's not that urgent.

[ralphb]

I've just pushed a new IDEA plugin, which you can download here.  Macro call arguments are now parsed correctly, which wasn't actually that difficult after I learned how it's done.  The only downside of this change is that now even more code runs when editing a line.  Let me know If IDEA becomes too sluggish.

 

In addition to macro calls, I also fixed a bug that caused preprocessor commands without arguments to be shown as erroneous.

 

Edited April 15, 2023 by ralphb

[Asmusr]

2 hours ago, ralphb said:

I've just pushed a new IDEA plugin, which you can download here.  Macro call arguments are now parsed correctly, which wasn't actually that difficult after I learned how it's done.  The only downside of this change is that now even more code runs when editing a line.  Let me know If IDEA becomes too sluggish.

 

In addition to macro calls, I also fixed a bug that caused preprocessor commands without arguments to be shown as erroneous.

 

Very nice. The only thing now is that in

.ifeq start_at_level,3

the comma is marked as an error.

 

The CPU and memory usage in IDEA is a bit concerning. I don't know if it got worse with the last update, but the source code for Trap Door is using a lot of resources. You can try for yourself: https://github.com/Rasmus-M/trapdoor. This is probably difficult to do anything about, and so far my machine has been able to cope with it.

[ralphb]

I've pushed a new plugin version that fixes the non-macro parsing.  Later I'll have a look at your project, but I guess IDEA just really is a resource hog.  When debugging my plugin, the sheer amount of code that is running outside the plugin is staggering.

 

[Sergioz82]

Hi,

I have a problem with xas99.py in Linux Mint. When I run the script I  get the errors below, but if I compile the same code on Windows it works fine.

I also get this problem on another Windows PC, but in happens when I run the script in "Portable Python-3.10.5 x64" through PyScripter-Launcher.exe.

Basically Run command in the launcher gives me the same error, but if I select the option "run external" it works fine.

 

$ xas99.py -b -R banked.asm -L bit.lst

 

Any idea of what it could be?

If I comment COPY "BITMAP.asm" I get normal errors of missing references in BITMAP.asm but the errors below are gone. 

 

Python versions on Mint  (v21.1) are 3.10.6 and 3.11.4

XDT99 version is 3.6.0

 

EDIT: I copied the Python folder in portable python 3.10.5 on Linux Mint, then modified build.sh to execute the command "wine /Python/python.exe /usr/local/sbin/xas99.py ..." and it works.

 

Quote

 

media/sergio/Chiavetta/Emulatori/FatuoASM$ . build.sh
: xas99, version 3.6.0
> BITMAP.asm <1> 0059 -     COPY "BITMAP.asm"
***** Error: Syntax error
> VDPUTILS.asm <1> 0008 - INITBM    MOV R11,R12
***** Error: Duplicate symbol: INITBM
> VDPUTILS.asm <1> 0027 - INITST    MOV R11,R12
***** Error: Duplicate symbol: INITST
> VDPUTILS.asm <1> 0030 - L1         CLR R1

...

> VDPUTILS.asm <2> 0107 -        JNE VMBW1
***** Error: Out of range: VMBW1 +/- -0x82
> SETS.asm <2> **** - 
***** Warning: Unused constants: l1c:6, l1p:263
> LEVELS.asm <2> **** - 
***** Warning: Unused constants: strtl1:11, ragno:56, ppstrel:58, panel:64, panelp:90, panelc:124
> SPRPATT.asm <2> **** - 
***** Warning: Unused constants: fr2dx:72, fr1sx:89, fr2sx:105, fr1up:123, fr2up:139, fr1dw:157, fr2dw:173, bat1:191, bat2:207, spd1dx:224, spd2dx:240, spd1sx:256, spd2sx:272
58 Errors found.

 

 

Edited June 24, 2023 by Sergioz82

[ralphb]

13 hours ago, Sergioz82 said:

I have a problem with xas99.py in Linux Mint. When I run the script I  get the errors below, but if I compile the same code on Windows it works fine.

This is quite strange.  Could you please try something?  In file xas99.py, please comment out lines 859, 861 and 862 (should be try/except/raise) and try again on Linux.  What error do you get now?

 

[Sergioz82]

47 minutes ago, ralphb said:

This is quite strange.  Could you please try something?  In file xas99.py, please comment out lines 859, 861 and 862 (should be try/except/raise) and try again on Linux.  What error do you get now?

 

Now I get this, also using python 3.10.5

Both symbols are defined

 

Quote

. build.sh
: xas99, version 3.6.0
> banked.asm <2> 0055 -     CLR @BNKW0
***** Error: Unknown symbol: BNKW0
> banked.asm <2> 0056 -     B @INIT
***** Error: Unknown symbol: INIT
2 Errors found.

 

[ralphb]

Would it be possible for you to share the code in question?  Either post a link, or mail to r@ti99.pro.

[Sergioz82]

17 hours ago, ralphb said:

Would it be possible for you to share the code in question?  Either post a link, or mail to r@ti99.pro.

Ok, I'll send it in a email

[ralphb]

You COPY'ed file BITMAP.asm contains an invalid Unicode char in the word "più":

 

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf9 in position 5103: invalid start byte   (not shown without removing one except clause, though)

 

You probably used a different encoding.  So, either use the UTF-8 encoding or don't use non-ASCII chars.   I don't know why the behavior is different for earlier Python versions, but my guess is that they removed some ambiguity about the encoding for text mode open().

 

[Sergioz82]

17 hours ago, ralphb said:

You COPY'ed file BITMAP.asm contains an invalid Unicode char in the word "più":

 

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf9 in position 5103: invalid start byte   (not shown without removing one except clause, though)

 

You probably used a different encoding.  So, either use the UTF-8 encoding or don't use non-ASCII chars.   I don't know why the behavior is different for earlier Python versions, but my guess is that they removed some ambiguity about the encoding for text mode open().

 

Oh ok.

I'm using Emacs and I'm not very used to it, I saw there's an option to change the text encoding, I'll try to set it to UTF-8. Besides, sometimes I code on Crimson Editor on Windows 10, probably I also messed up something by moving the files back and forth.

I'll try UTF-8, if iy won't work I'll delete the word, it's not important.

 

Thanks for the support!

 

 

[JasonACT]

This is a great software package, so thank you for making it available.

 

I have a couple of things to report, none are enough to make me want to use anything else though...  Even though I'm useless with Python myself.

 

The disassembler creates DATA statements for "TB" (0x1F) but my usage has a negative offset (" TB -11") so maybe that's the cause (I really can't tell - Python???).

 

Thierry's TI DISK DSR has 4 instructions that don't assemble (no big deal, I just comment them out and stick in 3 DATA statements from the original bytes)...

 

A4D94 MOVB @>004D(9),@>004D(9)

near A5B38:

 MOV @>0050(9),@>004A(9) copy sector #

near A5D5A:

 MOVB @>0050(9),@>0050(9) test result

 SZCB @>0042(9),@>0042(9) clear current token

 

Anyway, enough from me... Thanks again for making this time saving software package available!

Edited October 13, 2023 by JasonACT

[+Lee Stewart]

1 hour ago, JasonACT said:

Thierry's TI DISK DSR has 4 instructions that don't assemble (no big deal, I just comment them out and stick in 3 DATA statements from the original bytes)...

 

A4D94 MOVB @>004D(9),@>004D(9)

near A5B38:

 MOV @>0050(9),@>004A(9) copy sector #

near A5D5A:

 MOVB @>0050(9),@>0050(9) test result

 SZCB @>0042(9),@>0042(9) clear current token

 

Those statements have only one space between the instruction and comments. There must be at least two spaces for xas99 to recognize the comments. Otherwise, xas99 tries to include the comments as part of the instruction, which will obviously fail. The reason for this is that xas99 allows single spacing of operands and operand expressions—something not allowed by the TI assembler.

 

...lee

[dasch]

@ralphb Let me add my voice to the praise of xdt99. xas99 has been enormously helpful in reconstituting the Munch Man source from the PDF. I must admit that I was initially caught out by the two-space-comment-field initially, though since I had to add a whole bunch of '@'s, my preservation of the original formatting had to give way at times. Barring that, I appreciate the strict mode. Plus, the superlative documentation turned out to be a highlight of the project. Thank you!

 

-- David

[JasonACT]

10 hours ago, Lee Stewart said:

Those statements have only one space between the instruction and comments. There must be at least two spaces for xas99 to recognize the comments. Otherwise, xas99 tries to include the comments as part of the instruction, which will obviously fail. The reason for this is that xas99 allows single spacing of operands and operand expressions—something not allowed by the TI assembler.

 

...lee

Thanks Lee, that has allowed me to clean up the code nicely.  I stuck a semi-colon in instead, to remind me.

 

On the disassembler side, I think the opcbitmask array should end with 8 instead of 12 to pick up negatives in "format 12" instructions (SBO, SBZ & TB).

[ralphb]

On 10/14/2023 at 12:44 AM, JasonACT said:

Thanks Lee, that has allowed me to clean up the code nicely.  I stuck a semi-colon in instead, to remind me.

 

On the disassembler side, I think the opcbitmask array should end with 8 instead of 12 to pick up negatives in "format 12" instructions (SBO, SBZ & TB).

For legacy sources that won't assemble as-is, you can use the strict syntax option -s to restore the original E/A syntax.

 

Regarding the disassembler, negative numbers are displayed as positive two's-complement, so -1 is displayed as >FF or >FFFF, depending on the size of the field.  For TB, -11 and >1F are not equal (-11 would display as >F5), though, so there might be a bug if those are actual numbers.  Note that the value field of SBO, SBZ, TB is 8 bit wide, so opcbitmask[12] == 8 should be correct, or did I misunderstand you?

[JasonACT]

12 hours ago, ralphb said:

Note that the value field of SBO, SBZ, TB is 8 bit wide, so opcbitmask[12] == 8 should be correct, or did I misunderstand you?

Not sure if you misunderstood. It's 12 in GitHub... (#special formats: 16, 12, 12, I think it should be 16, 12, 8 - at least, that worked for my disassemly.)

[ralphb]

9 hours ago, JasonACT said:

Not sure if you misunderstood. It's 12 in GitHub... (#special formats: 16, 12, 12, I think it should be 16, 12, 8 - at least, that worked for my disassemly.)

Oops, you're right, of course, I simply cannot count!    I'll make a fix that changes the value to 8.  Thanks for pointing it out.