Hello, > > allow to disable SIMD optimizations with PULSE_NO_SIMD; > > to give a consistent meaning to that environment variable > The SBC code is pulled straight from the BlueZ code, so it's nicer to > not change that code downstream. If the BlueZ folks would be amenable to > taking something like this upstream (say BLUEZ_NO_SIMD), we could pull > then setenv() that early in PA init if PULSE_NO_SIMD is set). > > I know it's a bit of a circuitous route to making changes, but this > split keeps things cleaner/more central. I understand your concern; my motivation was to provide a consistent behaviour of PULSE_NO_SIMD not sure if the effort to go upstream is worth it, another env var is probably not desirable there the rule that upstream code should not be modified is broken for rtkit.c (96a52257a90f03478976fe762f544145cf1d681a, pa_streq()), reserve.h, reserve-monitor.h (c5dca7cf2b0423f152c6e81a503cc9639b08f09f, spelling) the update-ffmpeg target of src/Makefile.am doesn't work anymore, instead wget --no-check-certificate https://raw.github.com/FFmpeg/FFmpeg/master/libavcodec/resample2.c seems to work; resample2 deviated a bit and needs some work to bring it up to date regards, p. -- Peter Meerwald +43-664-2444418 (mobile)