> Is there an ambiguity issue with server strings containing unix domain socket > paths that in turn contain embedded whitespace? Or are such whitespace- > containing socket paths prohibited by some construction rule that I'm probably > not aware of? I think there is (haven't tried); this is a separate issue other than the documentation mismatch I'm trying to address please file a bug for it thanks, p. > Here's what I'm getting at: > > The Wiki writeup on server strings > > http://www.freedesktop.org/wiki/Software/PulseAudio/Documentation/User/ServerStrings/ > > says that "[t]he server string is a space-seperated list of server addresses", > and then gives rules according to which each component 'server address' is > interpreted. But what if a unix domain socket contains embedded spaces? For > example: > > PULSE_SERVER='/annoying socket path/mysocket myhost:myport' > > or something similar as specified on the commandline (e.g. via the "--server" > option to pulseaudio or pactl)? > > At the C subroutine level, obviously there is no problem with embedded spaces, > but the issue arises (at least, in my head) about parsing of the --server > option or the various envars (e.g. $PULSE_SERVER) used to provide server > strings. > -- Peter Meerwald +43-664-2444418 (mobile)
