Perl Script to Copile ASM-Sources with GPASM

audiocommander · January 3, 2007, 8:42pm

Hello everyone,

The information about this issue is a bit cluttered, so I start a new thread about it.

My aim and the reason for this posting is: I want to achive full *nix compatibility and be able to change and compile ASM based applications (like the SpeakJet PIC16F Firmware) without windows.

As you might know, one cannot compile MIOS sourcecodes (and also ASM based applications) with GPASM. Only MPASM can do the job without errors, but the cause of this is quite simple:

The problem is an instruction parameter passing in the Macros defined in “macros.h”

http://www.midibox.org/forum/index.php?topic=8175.msg57319#msg57319

As this bash script using sed is obviously not working on macs (don’t know the reason though, seems I’m too dumb to read regular Expressions…), I wrote a perl script, that you can download here:

http://www.audiocommander.de/downloads/midibox/mpasm2gpasm.tgz (4kb)

It’s my first perl script ever, so don’t expect poetry and elegance…

The thing is however:

I tested it with the bootloader src and it works fine, can compile it & the hex is the same the original hex-file.
I haven’t tested it with the MIOS src, because I don’t know how to setup a linker script to generate a Makefile for gputils

As some of you might have read in the speakJet thread, I have massive problems to compile the SJ firmware for the PIC16F. The generated .hex-file is different from .hex files compiled on Win-PCs. So I guess this is another incompatibility issue between MPLAB and GPUtils.

Because I checked the bootloader .hex and it is not different from the original bootloader .hex, I can roule out gpasm as the error… so the linker comes to my view.

I’m going to interspect that and report back, if someone likes to try the perl-script or has a nice link for me about GPUtils and Linkers, I’d be happy for feedbacks…

Cheers,

Michael

Related Links:

Wiki: How to compile MIOS on linux:

http://www.midibox.org/dokuwiki/doku.php?id=compiling_the_midibox_source_on_linux

Compile on Linux please:

http://www.midibox.org/forum/index.php?topic=6420.0

Assembler SID Help:

http://www.midibox.org/forum/index.php?topic=8175.30

Speakjet (hex file mismatch, unresolved):

http://www.midibox.org/forum/index.php?topic=2870.msg54723#msg54723

stryd_one · January 4, 2007, 9:00am

Man I wish I could help, but I’m a total *nix idiot… Just wanted to give props, cause I don’t intend upon being stuck on windoze forever.

audiocommander · January 4, 2007, 12:03pm

Hey stryd, your feedback is very welcome

What I need is no *nix guru, but some experienced know-how about ASM, Linker-Scripts and Makefiles. I remember your PIC18F620 tutorials and the talks about >256 byte arrays that require linker-adaptions… you know, when I’m looking at the linker script, my eyes pop out and all I see is numbers . And up to now, all readings I’ve done could not light up my personal brain darkness in this area ;D

Up to now I tried <updated>:

PIC18F-based apps:

bootloader: okay*
midimon_v2_0: okay*
mios_v1_9c_src: okay*
midibox_fm_v1_1: okay* (some warnings may occur)

PIC16F-based apps:

midimerger_v1.4: not okay*
mbhp_iic_speakjet_v1_0: not okay**
mbhp_iic_midi_v1_0: not okay**

^{* macro conversion with perl-script}

^{** no macro conversion needed}

I’m not yet totally clueless, but shortly before

see my last experiment with the mbhp_iic_midi below; I just generated the makefile with mkmk.pl and compiled the unaltered sources without errormessages with gpasm… the result is attached…

Cheers,

Michael

Edit: added test results for ASM based midimon & midimerger apps.

The good thing however is, that my perl-script seems to work sofar 8)

Edit2: added FM, changed MIOS_v1_9c results, split into PIC16 & 18 results… gets a lot clearer now…

749projecthexdiffgif6d52305279319a3f4879ccf4bce3c6fa

T.K · January 7, 2007, 4:37pm

Nice work!

I gave the MBFM source code a try; some gpasm errors are related to the reason, that the converter doesn’t ignore comments. E.g., if a comment contains “IFSET”, a new line will be generated with a propably wrong (or unwanted) instruction.

When you replace:

# search for macro occurences
if( $_ =~ m/$macros[$m]/ ) {
[/code]


by:
[code]  
 # search for macro occurences  
 if( $\_ =~ /^\s+$macros[$m]/ ) {

it will work better, as only valid macros will be replaced. Thereafter we have only a few number of errors:

cs_menu_print.inc:714:Error [103] syntax error
cs_menu_print.inc:715:Error [174] Unknown opcode "print"
cs_menu_print.inc:738:Error [103] syntax error
cs_menu_print.inc:739:Error [174] Unknown opcode "print"
[/code]


appr. replaced code:
[code]  
 btfss TMP1,&nbsp; 6,&nbsp; movlw '-'; (if value \<= 0x3f (6th bit cleared)  
 print "-"

means: we still have a problem with comments. A quick&dirty solution:

# search for macro occurences
if( $_ =~ /^\s+$macros[$m]/ ) {
        s/;.*//g;          # remove comments
[/code]



There is still one error thereafter:
[code]  
cs\_menu\_io\_tables.inc:75:Error [108] Illegal Character (,)

but this is a syntax error from my side which was ignored by MPASM - remove the “,” at the end of the line.

Code can be assembled with “gpasm main.asm”, there is no linker script required, w/o the “-c” option you will get a .hex file (main.hex) by default

Best Regards, Thorsten.

audiocommander · January 7, 2007, 9:05pm

cool, thanks a lot for the suggestions!

I updated the pl script.

There were some more warnings when I tried with the FM source, but at least a hex file was generated…

And I detected two more errors (in the script) while converting MIOS v1_9c, but I’ll look into that as soon as I have some more time, and maybe I’ll implement some error checking methods. When the script runs fine, I’ll put it to the Wiki.

There’s also some more good news regarding the differing hex files:

My testings revealed that – as far as I can say until now – only sources written for the PIC16F are wrong.

It seems all sources for PIC18F’s can be compiled well with gpasm!

Best regards,

Michael

chriss · January 7, 2007, 10:38pm

hello,

first i want to say that i am happy that you work on a solution for linux, mac users.

but in my case it does not work…

here my result in the terminal, maybe it helps. Application is the MbSEQ V3 build 60