On Sat, 28.03.09 17:06, Finn Thain (fthain at telegraphics.com.au) wrote: > The only interpretation here was my suggestion that STREAMS was the cause. > That is guesswork on my part. > > If you are asking "am I sure that my observation of this behaviour is > correct?" then yes, I am. > > Perhaps a better solution is to use blocking writes. Would that be OK with > you? Nope, not really. Blocking writes are not a good idea here. > > > > One last thing: it would probably be a good idea to allocate a > > > > pa_card object and attach the sink and the source to it. > > > > > > It is possible to open /dev/audio twice by loading the solaris module > > > twice -- once for the sink (passing record=0) and once for source > > > (passing playback=0), thus giving seperate threads/LWPs for source and > > > sink. It might be misleading to allocate two cards in that situation? > > > > Why would you want to do this? We should make the common case work. > > Are you saying that the common case cannot work without a pa_card > abstraction? Oh, it can work perfectly without allocating a pa_card. But it would be good if it did. That would clients to figure out which sink belongs to which source. > Are you referring to the claim in the Darwin and Linux man pages that > Solaris may return a relative path? FWIW, that claim contradicts the > present Solaris man page. No, I was referring to the MAX_PATH issue: MAX_PATH is not defined everywhere since some OS have no limit on the filename length. So relying on MAX_PATH is unportable as is passing NULL as second parameter. > OK. I'll search out or write an m4 macro for the realpath(NULL) feature > test and send another patch. Please disregard this one. I have now merged your patch. I had to change a few things to make it apply cleanly. Since I have no access to Solaris I am unable to test this though, so please check if things still work for you. I also worked around the realpath() issue mostly. It should work fine on Solaris now, as well. Thanks again for your contributions. Lennart -- Lennart Poettering Red Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4