Στις 08-03-2017, ημέρα Τετ, και ώρα 16:55 +0100, ο/η John Paul Adrian Glaubitz έγραψε: > It should. But it has been like that for ages and no one has > complained so far unlike with the ppc64 port. So I really think it > should stay that way. Most PowerPC machines are already underpowered, > so taking away Altivec on applications where it really makes a > difference would be suboptimal. For many years, Altivec enablement has been debated in both ports (powerpc & ppc64). AFAIR, the consensus is that Altivec should be disabled by default and either enabled in a separate package (eg. atlas-altivec) or used through runtime detection (eg. ffmpeg, vlc). If a package does enable altivec by default, then that's a bug and a policy violation. On the other hand, being a SIMD fanatic that I am, I wholeheartidly agree on enabling Altivec -and SIMD support in general- on a distribution-wide scale. If a CPU does not support SIMD (Altivec, NEON, SSE/AVX, etc), then tough luck (applies to Marvell, e5500, old Intel/AMD CPUs, etc). But until that happens, we have to support the decision, if Firefox is built with altivec enabled (and I don't mean runtime detection, but failure to run on non-altivec CPUs), then that's a bug, period. Regards Konstantinos
Attachment:
signature.asc
Description: This is a digitally signed message part