asm-pro 1.19

[a/b]

Here's a little something I've been working on the last year or so. Started out of necessity because asm-one has too many issues and is definitely out of development, and since asm-pro is in a much better shape I started using it more, but every once in a while I ran iinto a problem. Fortunately, it comes with the source code, so here we are...
I've spent too much time today typing the readme and release notes, so I'll copypasta the readme file here. It's available on aminet in dev/asm.

====
Update for the ASM-Pro package, includes latest executable.
Some of the older versions use ENVARC:Asm-Pro_beta.Pref instead of
ENVARC:Asm-Pro.Pref. If your preferences don't look OK, please ensure
that you have the correct (not beta) .Pref file.

This release fixes a number of KS2.x, 68000/010, and FPU compatibility
issues, as well as a large number of other issues (over 60 bugfixes in
total, see the history file for details). You can run it on KS2.04+
and 68000 based systems again.
Code has been cleaned up and optimized (over 2600 micro optimizations,
and a number of subroutines rewritten), resulting in 15KB reduction of
the executable size and a smaller memory footprint overall.

Additions (detailed description available in the history file):
1. WX command to enable short code and data hunks (code+bss, data+boss)
2. ADDBP directive to add a breakpoint from the source code
3. Command line history jump by 10 lines, don't store repeated lines
4. Macro symbol parameter \<symbol> and \<$symbol> (vasm style)
5. Assembly elapsed time

IF YOU FIND ANY ADDITIONAL KS2.X OR 68000/010 ISSUES, OR IF SOMETHING
DOES NOT WORK PROPERLY, COMPARED TO v1.18 (read: if I messed up :P),
PLEASE LET ME KNOW. I'll do my best to fix them.

[admiral]

Restored 68k support is very welcome.

Got a git repository up somewhere?

[a/b]

Source is not available yet, the plan is to release it with v1.20.

[Edders]

That is great news, thanks for sharing. It is my favourite asm IDE for the Amiga - especially the colour coding of the source. I'm looking forward to trying this new version out.

[a/b]

Latest version (1.19b) is in the zone, make sure to use that one.

[kamelito]

A/b did you knew about that one? https://github.com/bsdf/TRASHM-PRO

[a/b]

Originally Posted by kamelito

A/b did you knew about that one? https://github.com/bsdf/TRASHM-PRO

Yeah, it was on the radar (there's a thread on this forum).
To keep it simple and politics free: I don't like the idea of forking and rebranding it, there's no need for that. Also, the way he approached it made finding useful changes difficult (diff from original source? "everything" is diff-erent).

[pintcat]

@ a/b:

movea.w    ax,+(ay)
movea.w    ax,-(ay)
movea.l    ax,+(ay)
movea.l    ax,-(ay)

All these lines result in an "Adress reg. expected" error.

movea    ax,+(ay)
movea    ax,-(ay)

These give an "Undefined symbol" error.

[StingRay]

Originally Posted by pintcat

@ a/b:

movea.w    ax,+(ay)
movea.w    ax,-(ay)
movea.l    ax,+(ay)
 movea.l    ax,-(ay)

All these lines result in an "Adress reg. expected" error.

Error message is correct, movea requires an address register as destination.

movea    ax,+(ay)
movea    ax,-(ay)

These give an "Undefined symbol" error.

First instruction is invalid (should be (ay)+), second works with ASM-Pro 1.16d at least (did not check with latest ASM-Pro version) but will be silently converted to the correct move.w ax,-(ay) version.

Edit: Checked with ASM-Pro 1.19b, no error message, it also silently converts the instruction to its correct version.

[a/b]

movea ax,-(ay) works fine here in 1.19c, and should in any new or old version. Did you try only the second (predec) instruction, without the invalid preinc?
I'm using asm-pro to generate itself (3x actually, e.g.: x1.19c in 1.18, then y1.19c in x1.19c, and finally z1.19c in y1.19c, and all 3 must be 100% identical, except timestamps) so these definitely wouldn't go unnoticed.

[pintcat]

Ah got it, thanks. It was my mistake the whole time and now I got confused since this is the first assembler which dares to complain. Life is tricky.
In fact, "ax,+(ay)" is always wrong and should be "ax,(ay)+". It was my typo and I didn't notice because of being focused on the error messages.

[pintcat]

@a/b: Works here too. I checked every line on its own, but probably forgot to check the predec instruction after all.