On Sat, 2014-10-25 at 01:16 +0200, Max Kellermann wrote: > On 2014/10/14 07:35, Ian Scott <i...@polpo.org> wrote: > > + ss.format = PA_SAMPLE_FLOAT32NE; > [...] > > ss.format = PA_SAMPLE_S16NE; > > This cannot work. No matter what you write to ss.format, it will be > overwritten right after the switch.
Oops, you are right. I forgot to remove that line. With it removed I have verified the patch functions as expected. > But I like how simple this patch looks, compared to Steven's > overcomplicated probing code (which I don't know what it's good for). I looked at his patch and I'm not quite sure, but I am not a Pulse expert. Steven, when would that probing fail? There's also the byteswapping which I cannot test for because I don't have access to a little-endian box. > Though I believe you can remove the #ifdefs. All of my chroots > support these formats. If a specific libpulse version is needed, we > may add a minimum required version to the pkg-config check in > configure.ac. The sample formats between the #ifdefs were added in PulseAudio 0.9.15, which was released in 2009. I imagine any system that old won't be able to compile mpd anyway due to requiring newer versions of other libraries. So might as well remove the #ifdefs. Ian _______________________________________________ mpd-devel mailing list mpd-devel@musicpd.org http://mailman.blarg.de/listinfo/mpd-devel
