Opened this thread to collect information and discussions about Deluxe Music 2.0 issues on AROS/68k, to avoid polluting https://amigaworld.net/modules/newbb/viewtopic.php?topic_id=45830&forum=2&start=60&mode=viewtopic&viewmode=flat&order=0


@kolla

Quote:

kolla wrote: NOTES This function had previously been documented that it would accept a null font. This practice is discouraged. o Use of a RastPort with a null font with text routines has always been incorrect and risked the guru. o Keeping an obsolete font pointer in the RastPort is no more dangerous than keeping a zero one there. o SetFont(rp, 0) causes spurious low memory accesses under some system software releases. As of V36, the following Amiga font variants are no longer directly supported: fonts with NULL tf_CharSpace and non-NULL tf_CharKern. fonts with non-NULL tf_CharSpace and NULL tf_CharKern. fonts with NULL tf_CharSpace and NULL tf_CharKern with a tf_CharLoc size component greater than tf_XSize. Attempts to SetFont these one of these font variants will cause the system to modify your font to make it acceptable.

Good. Then a implementing the last part should fix this issue.

Notice, however, that this is the behaviour of V36, whereas for previous versions:
"This function had previously been documented that it would accept a null font."
and:
"This practice is discouraged."
Which does NOT bound other implementations of the same API to perform like the Amiga o.s. V36.

So, fixing it is "just" a way to allow badly written applications to properly work.


@Georg

Quote:

Georg wrote: @Hammer Messed around a bit with Deluxe Music on AROS 68k, today. Use "Typeface" from Aminet to load and resave the fonts DMusic/8 and DMusic/16. For some reason they have NULL tf_CharKern, altough they are proportional. AOS3 seems to correct this problem at load/run time. AROS doesn't so during use bad things happen. Like text extent calculations give negative width for template to be used by Text/BltTemplate.

As per above.
Quote:

Next, patch the DMusic program itself with a hex editor. There are four longs in there "80080007". Change to "88880007". 80080007 == ASLFR_HookFunc. The program relies on the function to get its small-data-whatever-base-pointer in A4 register. When MorphOS initialliy still used AROS asl.library, they added some code to deal with this, but it isn't compiled in for AROS. I have not checked if all occurencies of 80080007 in the program file are really all ASLFR_HookFunc usages. If one or more is not, you probably get a crash somewhere else as free bonus.

What's the difference between 80080007 and 88880007?
Quote:

After starting the program use Scout to ~disable the InputHandler "MKSoft ToolWindow" by setting it's priority to very low value. This thing doesn't work correctly in AROS, but is otherwise used to allow clicking into toolwindow gadgets without deactivating main window.

Is it allowed by the Amiga o.s. guidelines?

@cdimauro

Quote:

So, fixing it is "just" a way to allow badly written applications to properly work.

And that's what maintaining operating systems is all about.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro Quote: So, fixing it is "just" a way to allow badly written applications to properly work. And that's what maintaining operating systems is all about.

May I suggest you talk about things you have expertise with?

Maintaining systems is about your profession.

Developing oses is (also) about developers profession.

Quote:

May I suggest you talk about things you have expertise with?

Friend Cdimauro,

troglodyte manners notwithstanding - if you had to speak only of what you have expertise of,
then you would be limited to blank, zero characters posts.

Emojis, perhaps.

So please, be nice to your comrades in Amiga.
Or, be even nicer: remove yourself.



/MEGA!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

@cdimauro

Quote:

cdimauro wrote: May I suggest you talk about things you have expertise with?

I do, and it is true that legacy compatibility and even work-arounds for badly written software is something all OS development must deal with. Especially so for OSes with tons of legacy closed source software.

Quote:

Maintaining systems is about your profession. Developing oses is (also) about developers profession.

What are you even talking about? Who has OS development as a profession in Amiga land?

Last edited by kolla on 11-Apr-2026 at 07:11 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

[quote]What are you even talking about? Who has OS development as a profession in Amiga land?

Why oh why, friend kolla, you still humour that goblin at all.
He is a nonperson.
Disregard him.


NOW!!!





/mega

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

@cdimauro


Since this seems to be fairly well diagnosed at this point has anyone submitted it to oh I don't know the AROS bug tracker?

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect

@cdimauro

Quote:

What's the difference between 80080007 and 88880007

It's just a random change such that asl.library ignores the tag/hookfunc.

Quote:

Is it allowed by the Amiga o.s. guidelines?

I have not checked what exactly that InputHandler in DeluxeMusic is doing. I just noticed that it seems to mess up some InputEvents or something and things don't work as they should like when clicking on the "keys" ( ?) in this "piano" (?) window to produce sound. Or when placing notes in the notation window. Sometimes did not work and you needed to press/release control key to make it work again.

It works with the InputHandler disabled. It just looks uglier. As when clicking in tool windows, the main window is shortly made inactive and then active again.

Also there were gfx problems, unless telling DeluxMusic to run on WB screen which has also to be set to use some PAL screenmode.

@Georg




https://github.com/aros-development-team/AROS/commit/1dc3f432d991a55318e87bb8d85fa0f903970c38

Last edited by terminills on 15-Apr-2026 at 06:42 PM.
Last edited by terminills on 15-Apr-2026 at 06:38 PM.

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect

@terminills

There are a lot of architecture dependent macros, I hope that fix(?) is compliant to other architectures.

I hoped the parameter passing would follow some kind of 'natural' order and would be handled by every compiler in the same order it was on the original Amiga.

---

I saw one screenshot and don't know if this will fix other problems with EA applications and will not interfere with other applications using the asl.library.

----

What I forgot:

Thank you, for improving AROS.

Last edited by OneTimer1 on 16-Apr-2026 at 03:02 PM.

@terminills

If it's possible to tell gcc not to touch/use A4 register when generating code of asl.library, no real source code change would be necessary. Google says there may be something like "-ffixed-reg" (-ffixed-a4" ?) or as alternative "register volatile type *global_ptr asm ("a4")" in the source.

This "if (*funcptr >= (UWORD)0xFF00)" thing seems to be MorphOS specific check for TRAP_#? 68k opcode. That stuff they use for 68k PPC code transition. Not relevant to AROS 68k.

@OneTimer1

Quote:

by OneTimer1 on 16-Apr-2026 8:29:29 @terminills There are a lot of architecture dependent macros, I hope that fix(?) is compliant to other architectures. I hoped the parameter passing would follow some kind of 'natural' order and would be handled by every compiler in the same order it was on the original Amiga. --- I saw one screenshot and don't know if this will fix other problems with EA applications and will not interfere with other applications using the asl.library. ---- What I forgot: Thank you, for improving AROS.

I didn't fix it kalamatee did. I merely pointed him at the thread and another member of the community reported it to the bug tracker.

That being said he's not one to frequent the Amiga forums much so all the discussions in the world on these forums won't necessarily get bugs fixed.

What was different this time wasn't the bug but the ability to recreate it and a very good diagnostic in the thread.

Last edited by terminills on 16-Apr-2026 at 11:17 PM.

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect

@terminills

Quote:

terminills wrote: @cdimauro Since this seems to be fairly well diagnosed at this point has anyone submitted it to oh I don't know the AROS bug tracker?

Sorry, I was out traveling back home and then a lot of duties. But I've seen that everything is already sorted out. Great.


@kolla

Quote:

kolla wrote: @cdimauro Quote: cdimauro wrote: May I suggest you talk about things you have expertise with? I do, and it is true that legacy compatibility and even work-arounds for badly written software is something all OS development must deal with. Especially so for OSes with tons of legacy closed source software.

Again, that's NOT an OS duty!

Applications should follow the guidelines, and if they aren't doing it then it's responsibility of the authors to fix those bugs.

Or from some other kind-hearted person. Have you ever heard about WHDLoad, for example?
Quote:

Quote: Maintaining systems is about your profession. Developing oses is (also) about developers profession. What are you even talking about?

My expertise, of course: 44 years of development.
Quote:

Who has OS development as a profession in Amiga land?

Who cares? You don't need to work as OS developer to develop OSes. You only (!) need the right skills/expertise to do it.

That's elementary logic, kolla. And it's not the first time that you fail at it...

@cdimauro

in a perfect world and a living platform it would be fixed by the developers of the software. But on 68k we have a platform with mostly closed software so it is difficult to fix something. In this case either offer a workaround or accept that software not works. If it not breaks software and problem can be identified I am for the workaround.

@OlafS25

Quote:

OlafS25 wrote: in a perfect world and a living platform it would be fixed by the developers of the software. But on 68k we have a platform ....

It's even worse, we have old software that will be never improved. So if Apple or Microsoft are marking an API as 'depreciated' most software will be updated to a new API quickly.

On 68k we can't do this and every OS that should keep AOS Source code compatibility might run in the same problem.

Making it difficult to improve the OS, just think about the reasons preventing a new AOS getting SMP or Memory protection.

But that's a rant that is not really focusing on the problems with the ASL.library.

AROS 68k is considered an unstable system; the reason for this is that people want to use Amiga 3.1 software, which has poor compatibility, and if we then consider that OS 3.2 and 3.9 are completely incompatible.

In my opinion, AROS 68k should only be used with native AROS 68k software, and Amiga software should only be used if it has been recompiled for AROS 68k; this would make it more stable and more appealing.

There is a lot of AROS x86 software that could be ported to AROS 68k, as well as software from other platforms.

I have stopped distributing my AROS One 68k for this reason, mainly to prevent incompatible Amiga software from being installed and then being criticised for having an unstable distribution.

_________________

@AMIGASYSTEM

Quote:

... consider that OS 3.2 and 3.9 are completely incompatible.

So AOS3.2 and AOS3.9 are not AOS3.1 compatible?

Well, I saw GFX benchmarks where AOS3.2 was definitely slower than AOS3.1 but I never thought it will become incompatible.

@OneTimer1

they extend 3.1 further... whereas aros seems to be compatible only with 3.1 era software, sp aros might not necessary be able to run software that needs os3.2.

Last edited by pixie on 19-Apr-2026 at 11:29 AM.

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@OneTimer1
OS 3.9 is not compatible with OS 3.2 either; no OS 3.2 software works on OS 3.9, unless you create an OS Mix.

_________________

@AMIGASYSTEM

we agree to disagree here at least