Hi all, Any progress on this issue? Seems like nobody is working on it. It also looks like the underlaying issue (pulseaudio & speech-dispatcher combination) has been known for years and hasn't received much attention. Is the solution just not possible because it requires a choice for the default of pulseaudio between non-secure and working, or secure and non-working (given the goals and designs of the tools involved)? And currently the behavior is more secure and thus not working out of the box for pulseaudio users (but working for alsa users)?
Maybe the "solution" is to acknowledge the situation, disable the init scripts by default (at least of speechd-up, although that would also cause alsa users to not have speechd-up) and describe it properly in the documentation (I originally wrote debconf, but I don't think that is allowed), such that the user can be told how to fix it (I don't know, but it seems that starting pulseaudio as system user may solve the issues). Obviously, if we are convinced that what I describe above is correct, then we could add some detection for alsa vs pulseaudio into the loop (no idea how that would work and unsure which package(s) should do that). Paul
signature.asc
Description: OpenPGP digital signature
