[MPlayer-dev-eng] [PATCH] Remove OpenBSD bits from Sun audio backend

[Reimar Döffinger]

> On 20 Mar 2022, at 02:03, Alexandre Ratchov <alex at caoua.org> wrote:
> 
> On Sat, Mar 19, 2022 at 09:45:33PM +0100, Reimar Döffinger wrote:
>> Hi!
>> 
>>> On 18 Mar 2022, at 20:57, Brad Smith <brad-at-comstyle.com at ffmpeg.org> wrote:
>>> 
>>> ping.
>>> 
>>> On 3/2/2022 8:18 PM, Brad Smith wrote:
>>>> Remove some OpenBSD bits from the Sun audio backend. We have been using sndio
>>>> for 12+ years.
>> We were wondering just out of principle, is that just because the
>> code is outdated/unnecessary, or because ao_sun can not work on
>> OpenBSD at all anymore?
> 
> Hi,
> 
> The SunAudio-style ioctls and associated structures don't exist
> anymore, so the code doesn't build on OpenBSD.

Yes, I see it ends up disabled by configure.

>> In the latter case it would be a bit the question, what's the harm
>> in leaving it in as-is...
> 
> I see no harm: ao_sun.c is not used on OpenBSD, and __OpenBSD__ is not
> defined on OSes using ao_sun.c.

While this is a fairly harmless case, I don't agree that that's the right way to argue.
For every change there should be a reason why it's a GOOD thing to do, and that should be part of the commit message.
Here the argument is weak either way, but not exactly much in favour of the patch IMO:
Plus:
- save 3 words of code
Minus:
- can't compile on truly ancient/historical OpenBSD versions anymore (while configure still thinks it is supported)

And to be completely honest I am not sure I see the point of such trivial changes with no real impact when it seems MPlayer doesn't even compile on OpenBSD at all, because of issues in tvi_bsdbt848 (soundcard.h and _oss_ioctl issues), plus linker issues in fft_neon.S on Arm.
Which also means that despite setting up a VM I can't even properly test if any of these fully work.