New Aranym release?

[Eero Tamminen]

Hi,

Any chance of there being a new Aranym release, as last v1.1.0 one is
from 2019, 6 years ago...?


- Eero

[Thorsten Otto]

On Mittwoch, 21. Mai 2025 23:20:12 CEST Eero Tamminen wrote:

> Any chance of there being a new Aranym release,

Petr started an attempt more than a year ago, but it did not happen...

Best we could do is to get some snapshots build again. I'll start looking into it, but can't promise anything.

[Francois LE COAT]

Hi,

Thorsten Otto writes:

Eero Tamminen wrote:

> Any chance of there being a new Aranym release,

Petr started an attempt more than a year ago, but it did not happen...

Best we could do is to get some snapshots build again. I'll start looking into it, but can't promise anything.

A new version number would be great, since all this time. Binary snapshots are required for Windows and macOS, because for GNU/Linux binaries are not really distributed, apart from the OS distributions...

For macOS, Philipp Donzé is building binary snapshots, but rarely (once a year). For Windows, there's no binaries since the ARAnyM 1.1.0 version.

For GNU/Linux, a new version number would mean that binaries would be updated in various OS distributions.

I made a demo recently with the 1.1.0 ARAnyM version under macOS:

Pops On the channel from Francois LE COAT 2025/05/08 https://www.youtube.com/watch?v=_pNcVdRGcRg

All is running right for me. But I think that a new version number would be welcome. Hatari will be soon version 2.6 :-)

Wishing the best for ARAnyM.

Best regards,

-- 
François LE COAT
Author of Eureka 2.12 (2D Graph Describer, 3D Modeller)
https://eureka.atari.org/

[Miro Kropáček]

On Thu, 22 May 2025 at 20:13, 'Thorsten Otto' via ARAnyM <ara...@googlegroups.com> wrote:

Petr started an attempt more than a year ago, but it did not happen...

Once a year, Petr tries to 'recruit' me for that but I have always had the same answer: Aranym's issue with a release isn't lack of horsepower to actually compile and release everything but to thoroughly test and fix at least the biggest showstoppers (usually related to SDL and/or cursor and/or screen handling). IIRC some of them are on github but usually in "works for me" or no reaction at all state.

Best we could do is to get some snapshots build again. I'll start looking into it, but can't promise anything.

Yes, that would be certainly better than nothing for sure. Definitely for Windows and perhaps Mac OS X users.

--

http://mikro.atari.org

[Thorsten Otto]

On Samstag, 24. Mai 2025 19:00:32 CEST Francois LE COAT wrote:

> For GNU/Linux, a new version number would mean that binaries would be

> updated in various OS distributions.

Yes of course, but even with a new release, it will take some time for it do show up in distributions. If you look at https://repology.org/project/aranym/versions you will see that there are still distributions which ship with the much older 1.0.2 version

[Francois LE COAT]

Hi,

Thorsten Otto writes:

Looking at the link, most of the GNU/Linux distributions have adopted the ARAnyM 1.1.0 version.

Anyway, if there's not a new version number, ARAnyM 1.1.0 can't take advantage of the new EmuTOS 1.4 version. Only recent ARAnyM binary snapshots are taking advantage of it. There's recent binaries for macOS, but not for Windows. And none of GNU/Linux distributions have enough recent binaries for supporting EmuTOS 1.4.

We really need a new ARAnyM version number, after all these years staying at version 1.1.0.

I hope it makes sense =)

[Eero Tamminen]

Hi,

On 9.6.2025 8.45, Francois LE COAT wrote:
> Anyway, if there's not a new version number, ARAnyM 1.1.0 can't take advantage
> of the new EmuTOS 1.4 version. Only recent ARAnyM binary snapshots are taking
> advantage of it. There's recent binaries for macOS, but not for Windows. And
> none of GNU/Linux distributions have enough recent binaries for supporting
> EmuTOS 1.4.

Hatari has targets for several "native" and Linux cross-compile Windows
targets in the CI files:
https://github.com/hatari/hatari/blob/main/.cirrus.yml#L103
https://github.com/hatari/hatari/blob/main/.gitlab-ci.yml#L27

I would assume at least one of them would work also for Aranym?


- Eero

[Francois LE COAT]

Hi,

I don't know what you're talking about. It seams that it is about
Hatari's sources. But I don't have a GitHub account. This is quite
obscure to me.

I'm sorry.

Regards,

[Philipp Donzé]

Hi François

> Anyway, if there's not a new version number, ARAnyM 1.1.0 can't take advantage of the new EmuTOS 1.4 version. Only recent ARAnyM binary snapshots are taking advantage of it. There's recent binaries for macOS, but not for Windows. And none of GNU/Linux distributions have enough recent binaries for supporting EmuTOS 1.4.
>
> We really need a new ARAnyM version number, after all these years staying at version 1.1.0.

I don’t think doing ARAnyM releases is about bumping the version number up and the eventual technical challenges to build the various binaries so everybody can download easily.

It’s more about proper organisation:
- The various builds have to be gathered for testing: many binaries come from different developers.
- The number of build variants does not help either (Libraries: SDL 1 vs SDL 2, OpenGL, MesaGL, … CPU implementations: plain 68040 with FPU, MMU and JIT?) and target environments (Linux, Windows, macOS (versions?), Intel x86, Apple Silicon, Arm, 32 bit, 64 bit).
- Proper testing of the builds has to be orchestrated (and it's even unclear what „proper“ means - what features should be tested because they are considered the bare minimum).
What would be the test OS? TOS 4, EmuTOS 1.4, MiNT? Based on Disk Images or HostFS
There are so many features in ARAnyM and all should be tested to ensure people know what works and what is broken.
- Issues have to be verified, discussed, if possible fixed before doing a final release.

This would all need a lot of time and resources… and as dead as the discussion group here is, I don’t think many would be willing to put in the effort.

I would help with the macOS developer part. But I’m not a good tester as I don’t use ARAnyM on a regular basis and I have too many “test setups“ and not a single “productive one“.
So I think we need many more volunteers to do it properly.

Kind regards
Philipp

[Miro Kropáček]

On Mon, 9 Jun 2025 at 12:41, 'Philipp Donzé' via ARAnyM <ara...@googlegroups.com> wrote:

This would all need a lot of time and resources… and as dead as the discussion group here is, I don’t think many would be willing to put in the effort.

I 100% agree with everything you have written.

On a more pragmatic side, perhaps we should just release (build) current code as is and see what happens. It doesn't help to get bug reports for a several years old release anyway.

--

http://mikro.atari.org

[Vincent Rivière]

On 09/06/2025 at 12:48, Miro Kropáček wrote:
> On a more pragmatic side, perhaps we should just release (build) current
> code as is and see what happens. It doesn't help to get bug reports for a
> several years old release anyway.

Just keep in mind that virtually all users only use the binaries provided
with releases, for their favorite OS. As there hasn't been any ARAnyM
release since 2019, you can be sure that almost no one has seen the
improvements done since then.

If the ARAnyM team just bumps the version number and publish the new
binaries for all supported operating systems, people will use it.

- If people are satisfied with the new release binaries, they will use it.

- If they aren't, they will report bugs. And they could still use the older
release, until they are fixed.

So my advice would be to publish a release as is, with current Git sources.
Of course, it won't be perfect. But people will report major bugs, then they
could be fixed for the next release, which should be satisfying for anyone.

--
Vincent Rivière

[Francois LE COAT]

Hi,

Philipp Donzé writes:
> So I think we need many more volunteers to do it properly.

Well I think that Vincent Rivière can test Windows version. He did it
successfully in the past. You Philipp can build macOS versions. We need
a SDL1 JITc version for old Mac (OS X). We also need a SDL2 (3?)
non-JITc version for recent Mac (>OS X). I can test those versions with
my iMac 27" and MacBook Pro 16". For GNU/Linux that will take more time,
but no binaries are required for the moment. This is not too demanding
conditions for the ARAnyM Team. And thus we could test the support for
EmuTOS 1.4. This is the main point =)

So far, everything was running alright.

Best regards,

[Eero Tamminen]

Hi,

On 9.6.2025 11.30, Francois LE COAT wrote:
>> Hatari has targets for several "native" and Linux cross-compile
>> Windows targets in the CI files:
>> https://github.com/hatari/hatari/blob/main/.cirrus.yml#L103
>> https://github.com/hatari/hatari/blob/main/.gitlab-ci.yml#L27
>>
>> I would assume at least one of them would work also for Aranym?
>
> I don't know what you're talking about. It seams that it is about
> Hatari's sources. But I don't have a GitHub account. This is quite
> obscure to me.

While clicking on a link to view a text file, and determining whether
environment variables and commands / package names (required for Windows
builds) are under "env" or "script" heading in it, may seem as daunting
/ obscure tasks if one is unfamiliar with modern Internet, may I just
suggest boldly trying it? :-)


- Eero

[Francois LE COAT]

Hi,

It seems the solution is to choose a 512kb version of EmuTOS 1.4, for
ARAnyM 1.1.0. But I don't know precisely which one. That's because
ARAnyM 1.1.0 is not supporting 1mb versions of EmuTOS. Do you know which
512kb version of EmuTOS 1.4 I should use with ARAnyM 1.1.0 ?

Thanks,

[Eero Tamminen]

Hi,

On 10.6.2025 11.51, Francois LE COAT wrote:
> It seems the solution is to choose a 512kb version of EmuTOS 1.4, for
> ARAnyM 1.1.0. But I don't know precisely which one. That's because
> ARAnyM 1.1.0 is not supporting 1mb versions of EmuTOS. Do you know which
> 512kb version of EmuTOS 1.4 I should use with ARAnyM 1.1.0 ?

That's easy as there's only single 512k EmuTOS version:
https://sourceforge.net/projects/emutos/files/emutos/1.4/

EmuTOS v1.4 512k version should work also with older Aranym, just be a
bit less optimal, as it's built for 680x0, not specifically for 040, and
at boot tests also for HW that Aranym does not support.

Main new feature in EmuTOS 1.4, is TC support in VDI/AES. However, if
you're using Aranym with fVDI or NVDI, not EmuTOS VDI, you could just
use the 512k Aranym ROM from previous EmuTOS v1.3 release:
https://sourceforge.net/projects/emutos/files/emutos/1.3/


- Eero

PS. EmuTOS config options selected for Aranym builds:
https://github.com/emutos/emutos/blob/master/include/config.h#L59

[Roger Burrows]

On 10 Jun 2025 at 10:51, Francois LE COAT wrote:
>
> It seems the solution is to choose a 512kb version of EmuTOS 1.4, for
> ARAnyM 1.1.0. But I don't know precisely which one. That's because
> ARAnyM 1.1.0 is not supporting 1mb versions of EmuTOS. Do you know which
> 512kb version of EmuTOS 1.4 I should use with ARAnyM 1.1.0 ?
>

The following link is for a zip containing all the 512K ROM versions of EmuTOS:
https://sourceforge.net/projects/emutos/files/emutos/1.4/emutos-512k-1.4.zip

Any of these will work, but note that there is a separate ROM for every
language because of lack of space for multi-language support. That's why the
ARAnyM ROM is now 1024K.

Roger