Bug#865721: openarena: Please add platform support for m68k

[Simon McVittie]

On Sat, 24 Jun 2017 at 17:19:22 +0200, John Paul Adrian Glaubitz wrote:
> On 06/24/2017 04:54 PM, Simon McVittie wrote:
> > Has the game resulting from this change been tested and demonstrated
> > to work?
> 
> No. But is this a requirement for each and single patch now?

If a downstream patch introduces things specific to one architecture,
then I am not happy about applying that change without at least trivial
testing on that architecture. For patches that merely make something
generic (like ioquake3 PR #129) I'm OK with it having been tested on
*some* architecture.

If someone (whether that someone is you or not) is interested in running
this leaf package on their favourite ports architecture, it should be
straightforward to provide confirmation that a patched build works.

> > I find it hard to believe that OpenArena is particularly useful on
> > m68k.
> 
> There are FGPA-based m68k CPUs now with 600 MHz and more, so yes, I think
> running OpenArena on m68k is possible.

OK, that's good to know.

> Why do you explicitly want to discriminate against ports architectures?

I don't. What I want is to not give the impression that I'm specifically
supporting architectures where nobody has ever demonstrated this package
to be functional.

> OpenArena would run perfectly fine on powerpcspe. The reason why the
> build currently fails there is because upstream thought it's a good
> idea to hardcode -maltivec on PowerPC even though not all PowerPC
> targets automatically support AltiVec.

The issue with powerpc(spe) (which doesn't apply to m68k) is that there
is a PowerPC bytecode JIT, and I don't know any PowerPC machine code, so
I don't know whether the output of the JIT relies on Altivec. If it does,
then the binary compiled without -maltivec might well compile perfectly,
but not actually work. I would rather have a build configuration that
is believed to work on at least Altivec PowerPCs (which, in practice,
seem to be well-correlated with those that have enough 3D processing
for a game engine) than something that hasn't been tested anywhere.

> If you remove -maltivec on powerpcspe conditionally (you can apply
> that particular patch only if $(dpkg --print-architecture) == powerpcspe,
> the package will build just fine.

If you can try OpenArena with QVM bytecode on a powerpcspe (or any older
powerpc with no Altivec, like a G3 Powermac) according to the instructions
at <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=701561#32> and
tell me that it does in fact work, I'd be happy to turn off -maltivec
on either powerpcspe, or powerpcspe and powerpc. There's an untested
patch on that bug.

> You don't think [a 1.2 GHz CPU] will run a 3D shooter from 1999?

With 3D hardware supported by Mesa, sure. With software rendering,
it won't be pretty or fast, but it might work (although if I remember
correctly the idTech3 engine specifically doesn't support software
rendering).

If you don't have 3D hardware, building mesa with llvmpipe might help,
although that's gated by an architecture whitelist in debian/rules
that doesn't currently include m68k or powerpcspe (if I remember the
Mesa maintainers' policy correctly, they'll enable it for architectures
where someone has tried it).

    S