Amiga QuickBasic Compiler

[amiwolf]

Originally Posted by cale_tucker

config sounds fine to me though the issue might be 3.2 related

Yes...

Originally Posted by cale_tucker

The 68000 most likely isn't the issue -

and no.

I tried a few configs in WinUAE.

68000 2+2MB Kickstart 3.1.4, Workbench 3.1.4 - OK
68000 2+2MB Kickstart 3.2, Workbench 3.1.4 - FAIL
68000 2+2MB Kickstart 3.2, Workbench 3.2- FAIL
68020 2+2MB Kickstart 3.2, Workbench 3.2 - OK

So the problem seems to be when a 68000 and Kickstart 3.2 are combined. AQB 0.8.0 loads fine on 3.2 if you have an 020.

Thank you for AQB, by the way. It has great potential.

[Pyromania]

It’s nice to see more development tools being created. Developers on the Amiga doing this should be encouraged and supported. There are so few of them they are like royalty.

[Pyromania]

Anyone doing a comparison to Quick Basic and AMOS Pro? This might make a great YouTube video. I would do it but I don’t know Basic programming.

[cale_tucker]

Originally Posted by amiwolf

I tried a few configs in WinUAE.

68000 2+2MB Kickstart 3.1.4, Workbench 3.1.4 - OK
68000 2+2MB Kickstart 3.2, Workbench 3.1.4 - FAIL
68000 2+2MB Kickstart 3.2, Workbench 3.2- FAIL
68020 2+2MB Kickstart 3.2, Workbench 3.2 - OK

So the problem seems to be when a 68000 and Kickstart 3.2 are combined. AQB 0.8.0 loads fine on 3.2 if you have an 020.

interesting, thanks for you thorough investigation. Really seems to be the combination 68000+3.2 that fails. Since I don't have any 3.2 roms, don't think I can do much about this at this point.

[amiwolf]

Originally Posted by cale_tucker

interesting, thanks for you thorough investigation.

You're welcome. More detail than my intial comment was needed. I hate vague bug reports then went ahead and made one myself! Sorry about that.

Originally Posted by cale_tucker

Since I don't have any 3.2 roms, don't think I can do much about this at this point.

No worries. If it helps, I should also add that 0.7.3 worked well on 68000 2+2MB OS3.2.

[cale_tucker]

Originally Posted by amiwolf

No worries. If it helps, I should also add that 0.7.3 worked well on 68000 2+2MB OS3.2.

...just ordered a copy of AmigaOS 3.2 (individual computers) - I will try to reproduce the problem when it arrives. Thanks again for you bug report - I will try to keep you posted on my progress.

[cale_tucker]

Originally Posted by amiwolf

So the problem seems to be when a 68000 and Kickstart 3.2 are combined. AQB 0.8.0 loads fine on 3.2 if you have an 020.

there is a new github issue that might be related:

https://github.com/gooofy/aqb/issues/12

can you confirm the crash happens only when AQB is launched without a file on startup?

[amiwolf]

Originally Posted by cale_tucker

...just ordered a copy of AmigaOS 3.2 (individual computers) - I will try to reproduce the problem when it arrives. Thanks again for you bug report - I will try to keep you posted on my progress.

Cool beans.

Originally Posted by cale_tucker

there is a new github issue that might be related:

https://github.com/gooofy/aqb/issues/12

can you confirm the crash happens only when AQB is launched without a file on startup?

Your tester on github beat me to it! I can confirm AQB 0.8.0 doesn't stumble on prefs.ini with 68000/OS3.2 when loaded as a default tool from Workbench. It's an easy workaround until you can squash the bug.

[cale_tucker]

AmigaOS 3.2 has arrived and with it could reproduce the problem. Address error happens deep inside the 3.2 ROM, not sure how to debug that one. Could be a caching/timing issue with the two consecutive write()s to the ini file or something along those lines.

Anyway, for now I have decided to work around the problem, aqb will (re-)write the prefs.ini file only once during shutdown (deinit). I have created a 0.8.1preview1 release and uploaded it here:

http://zamiga.org/aqb-0.8.1preview1.lha

could you give this one a try and see if it fixes the issue for you?

[amiwolf]

Originally Posted by cale_tucker

AmigaOS 3.2 has arrived and with it could reproduce the problem. Address error happens deep inside the 3.2 ROM, not sure how to debug that one. Could be a caching/timing issue with the two consecutive write()s to the ini file or something along those lines.

Anyway, for now I have decided to work around the problem, aqb will (re-)write the prefs.ini file only once during shutdown (deinit).

You tracked down the cuplrit rather swift. Great work!

Originally Posted by cale_tucker

I have created a 0.8.1preview1 release and uploaded it here:

http://zamiga.org/aqb-0.8.1preview1.lha

could you give this one a try and see if it fixes the issue for you?

Ah, Server not found.

[cale_tucker]

Originally Posted by amiwolf

You tracked down the cuplrit rather swift. Great work!



Ah, Server not found.

interesting typo I got in that URL - this should work:

http://zamia.org/aqb-0.8.1preview1.lha

[amiwolf]

Originally Posted by cale_tucker

interesting typo I got in that URL -

Only zamiga makes it possible.

Originally Posted by cale_tucker

- this should work:

http://zamia.org/aqb-0.8.1preview1.lha

That is does, the download and the 0.8.1preview1 itself. No issues so far on 68000 and 3.2. Thank you for the time and coin you spent to fix it all.

Happy Christmas!

[cale_tucker]

amiga aqb basic compiler+ide release 0.8.1 is now available for download, this bugfix release addresses AmigaOS 3.2.x and RTG issues but also comes with new commands and functions

https://github.com/gooofy/aqb/releases/tag/0.8.1

Improvements:

* runtime: SPRITE() function added
* runtime: SPRITE SHOW command added
* runtime: SPRITE MOVE command added
* runtime: SPRITE FREE command added
* runtime: ILBM LOAD SPRITE command added
* runtime: POINTER SPRITE command added
* runtime: POINTER CLEAR command added
* runtime: FONT and FONT FREE commands added
* runtime: FONT() function added
* runtime: TEXTWIDTH() function added
* runtime: FONSTYLE command added
* runtime: FONSTYLE() function added
* runtime: ON BREAK CALL command added
* runtime: BOB() function x/y offset arguments added
* ide: editor will use the system text font now
* ide: editor horizontal scrolling implemented

Bug Fixes:

* ide: editor RTG high/true color rendering fixed
* ide: write INI file on deinit only (seems to help with AmigaOS 3.2 68000 stability)
* ide: disable unavailable/unimplemented menu items
* help: fix function node refs in amigaguide help

[Pyromania]

Thanx for the update.

[amiwolf]

Good to see sprite commands added, excellent progress. Recently, I managed to salvage some ancient BASIC programs from my Spectravideo days so if I can find the time, AQB will be perfect to port them to my Amiga.

If you are following a roadmap, any idea on when sound commands will be implemented?

[cale_tucker]

Thanks for the positive feedback porting some of my old programs is definitely something I am planning myself as well. Would love to hear about your progress and let me know about issues you encounter. AQB is still in early stages of development, please be prepared for compiler as well as runtime bugs and changing APIs.

If you come up with something interesting or entertaining, please get it touch if you would consider contributing it to the demo programs included in the AQB distribution.

I have added a few basic (hehe) sound commands now:

in _aqb:

• WAVE
• WAVE()
• WAVE FREE
• SOUND
• SOUND WAIT
• SOUND STOP
• SOUND START

in IFFSupport:

• IFF8SVX LOAD WAVE
• IFF8SVX READ WAVE

you probably won't be playing cool MOD music with these yet, however they should be enough for some game sound effects as well as simple background music. I have decided to stick to OS compliant programming with these as well so they're purely based on what audio.device has to offer.

All this will be included in the upcoming 0.8.2 release - if you would like to try them right now, I have uploaded a preview release here:

http://zamia.org/aqb-0.8.2preview1.lha

As always I am interested in feedback.

[amiwolf]

Originally Posted by cale_tucker

I have added a few basic (hehe) sound commands now:

in _aqb:

• WAVE
• WAVE()
• WAVE FREE
• SOUND
• SOUND WAIT
• SOUND STOP
• SOUND START

in IFFSupport:

• IFF8SVX LOAD WAVE
• IFF8SVX READ WAVE

Wow! Had you already started on the sound commands or do you just work incredibly quickly? Either way, thank you for the 0.8.2 preview. AQB is really coming along in leaps and bounds.

Originally Posted by cale_tucker

If you come up with something interesting or entertaining, please get it touch if you would consider contributing it to the demo programs included in the AQB distribution.

I think you'd be better served by someone who knows how to code and can showcase your speedy new compiler, although thanks for the kind offer. I'm looking forward to what the talented folks of EAB can come up with using AQB.

[cale_tucker]

Originally Posted by amiwolf

Wow! Had you already started on the sound commands or do you just work incredibly quickly? Either way, thank you for the 0.8.2 preview. AQB is really coming along in leaps and bounds.

had a bit of spare time and since sound support was pretty high on the list I thought I'd have a go at that right away. Turned out to be not too much effort at least for the set of commands I have implemented for now which more or less just map what audio.device offers.

Thanks again for the positive feedback, really helps keeping me motivated

[cale_tucker]

https://github.com/gooofy/aqb/releases/tag/0.8.2

Improvements:

* runtime: WAVE, WAVE FREE, SOUND, SOUND WAIT, SOUND STOP, SOUND START commands added
* runtime: WAVE() function added
* runtime: IFF8SVX LOAD WAVE, IFF8SVX READ WAVE commands added
* runtime: BITMAP MASK statement added
* runtime: CLEAR statement added
* runtime: MID$, UCASE$, LCASE$, INSTR, LEFT$, RIGHT$ functions added
* runtime: ABS function added
* runtime: EOF() function added
* runtime: CLS, LOCATE, SLEEP FOR, PRINT, PRINT#, INPUT, LINE INPUT moved from \_aqb to \_brt
* runtime: INPUT#, LINE INPUT#, WRITE, WRITE# statements added
* runtime: DATE$, POINT functions added
* compiler: support pure interface modules that have no code

Bug Fixes:

* ide: EZRequest on source write fails instead of a plain exit()
* ide: ENDIF auto-indentation fixed
* examples: tetris code cleanup, use custom fonts
* compiler: fix float handling in DATA statements
* compiler: fix coord/coord2 error handling
* compiler: do not abort on type system inconsistencies (e.g. unresolved forwarded types)
* compiler: check and resolve all forward ptrs
* compiler: fix err msg position for constant declaration expression
* compiler: fix string type coercion (fixes #17, thanks to Tom Wilson for reporting this one)
* compiler: fix negative numeric literal handling in DATA statements
* compiler: fix string handling in DATA statements (fixes #18, thanks to Tom Wilson for reporting this one)
* compiler: fix ENDIF SLE stack underflow
* runtime: fix INT() behavior (matches ACE now), add CLNG() to online help
* runtime: fix LINE INPUT
* use EXIT\_FAILURE for fatal error conditions (fixes issue #13 by polluks)
* add "$VER" version string

[Etze]

Any news? It’s very promising so I‘d be glad if there‘d be improvements.