Re: Bug#1115385: vlc: FTBFS on powerpc architecture
Jeroen Diederen <[email protected]>
| Newsgroups | gmane.linux.debian.ports.powerpc |
|---|---|
| Message-ID | <[email protected]> |
Hello all, I was not suggesting to turn off Altivec for all powerpc based machines. All I am suggesting is to turn it off for the package VLC, so it can be successfully autobuilt. I built VLC for powerpc that way successfully. It was shown to work on the Mirari board. Regards, Jeroen Ken Cunningham schreef op 2025-09-17 04:00: > On 2025-09-16, at 1:45 PM, Stan Johnson wrote: > >> On 9/16/25 12:36 PM, Christian Zigotzky wrote: >>> >>>> On 16 September 2025 at 5:10 pm, John Paul Adrian Glaubitz >>>> <[email protected]> wrote: >>>> >>>> Another idea could be to disable Altivec for the 32-bit port and >>>> enable it on the 64-bit port so that the 64-bit embedded CPUs could >>>> use the 32-bit port. >>>> >>> >>> That sounds good. >>> >> >> PowerMac G4 systems, such as PPC G4 MDD, are 32-bit, but they also >> support Altivec. >> > > Given the option of > > a) disabling altivec for all PPC machines, vs > b) just for that one PPCG4 system > > it would easily be the less crippling option to sacrifice Altivec for > that one PPC G4 system rather than sacrifice it for everyone. > > Of course, upstreaming runtime Altivec detection is the best. Not at > all likely to happen 20 years after these machines left the gate, but > still, best if they will accept it. There are many examples of how to > do runtime Altivec detection that present themselves with 5 minutes of > Googling, including some compelling patches in FIreFox added by Cameron > Kaiser. > > I have no idea how likely it would be to sell a separate linux PPC > build farm of +altivec and -altivec to the debian powers that be. > Probably not much enthusiasm for maintaining that, but who knows. > > Ken