while determining if the device is in use.


git-svn-id: http://svn.xiph.org/trunk/ao@2505 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2488 0101bb08-14d6-0310-b084-bc0e0c8e3800

channel that is).


git-svn-id: http://svn.xiph.org/trunk/ao@2487 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2433 0101bb08-14d6-0310-b084-bc0e0c8e3800

size.  So we'll set it lower and multiply by number of output channels.


git-svn-id: http://svn.xiph.org/trunk/ao@2432 0101bb08-14d6-0310-b084-bc0e0c8e3800

blocking mode; there's no reason to use non-blocking.


git-svn-id: http://svn.xiph.org/trunk/ao@2431 0101bb08-14d6-0310-b084-bc0e0c8e3800

new canonical location.


git-svn-id: http://svn.xiph.org/trunk/ao@2429 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2428 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2427 0101bb08-14d6-0310-b084-bc0e0c8e3800

Mike Castle <dalgoda@ix.netcom.com> for noticing the problem.


git-svn-id: http://svn.xiph.org/trunk/ao@2392 0101bb08-14d6-0310-b084-bc0e0c8e3800

requested.  Thanks to Joe Drew <hoserhead@woot.net> for pointing this out.


git-svn-id: http://svn.xiph.org/trunk/ao@2365 0101bb08-14d6-0310-b084-bc0e0c8e3800

finish writing to the audio device when it is interrupted by a signal.


git-svn-id: http://svn.xiph.org/trunk/ao@2321 0101bb08-14d6-0310-b084-bc0e0c8e3800

output, so this is a cleanup, not a bugfix.


git-svn-id: http://svn.xiph.org/trunk/ao@2257 0101bb08-14d6-0310-b084-bc0e0c8e3800

workaround is now turned off by default.  If the --enable-broken-oss
option is passed to configure or if ALSA is detected, the workaround is
turned on during compilation.  This will resolve bug 48.


git-svn-id: http://svn.xiph.org/trunk/ao@2146 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2121 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2092 0101bb08-14d6-0310-b084-bc0e0c8e3800

compile farm.


git-svn-id: http://svn.xiph.org/trunk/ao@2091 0101bb08-14d6-0310-b084-bc0e0c8e3800

Timothy J. Wood <tjw@omnigroup.com> to the vorbis-dev list nearly a year
ago.  See http://www.xiph.org/archives/vorbis-dev/200009/0139.html for the
original posting.

I ported this blind, and have only verified that I haven't broken compile
on existing platforms.  It probably doesn't work right now.


git-svn-id: http://svn.xiph.org/trunk/ao@2089 0101bb08-14d6-0310-b084-bc0e0c8e3800

autodetection.  Should fix Debian bug #110858.


git-svn-id: http://svn.xiph.org/trunk/ao@2066 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@2015 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1995 0101bb08-14d6-0310-b084-bc0e0c8e3800

Made buffer mechanism more generic -- it's now a circular block buffer.


git-svn-id: http://svn.xiph.org/trunk/ao@1993 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1992 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1989 0101bb08-14d6-0310-b084-bc0e0c8e3800

suggests this is more or less optimal.  Added new option to allow the
number of periods to be adjusted without needing to recompile.  Also fixed
an incorrect sound writing loop as per Jack's suggestion.


git-svn-id: http://svn.xiph.org/trunk/ao@1981 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1971 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1970 0101bb08-14d6-0310-b084-bc0e0c8e3800

associated with a particular file output driver.  The additional
plugin function is optional, so drivers that don't need to implement it
don't have to.  That's why we don't need to bump the plugin API version number.


git-svn-id: http://svn.xiph.org/trunk/ao@1946 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1941 0101bb08-14d6-0310-b084-bc0e0c8e3800

separate subdirectories of $libdir/ao.


git-svn-id: http://svn.xiph.org/trunk/ao@1905 0101bb08-14d6-0310-b084-bc0e0c8e3800

like "make debug" (which switch the CFLAGS) will still work.


git-svn-id: http://svn.xiph.org/trunk/ao@1902 0101bb08-14d6-0310-b084-bc0e0c8e3800

on device.


git-svn-id: http://svn.xiph.org/trunk/ao@1862 0101bb08-14d6-0310-b084-bc0e0c8e3800

rather than ao_initialize().  This has the benefit of avoiding autodetection
unless is it necessary.


git-svn-id: http://svn.xiph.org/trunk/ao@1854 0101bb08-14d6-0310-b084-bc0e0c8e3800

opens the dsp between plugin_test() and plugin_open().  Now we always
open it with O_NONBLOCK and remove the blocking later if needed.  The
ALSA docs claim that this is required by the OSS spec, but that is
not so obvious.


git-svn-id: http://svn.xiph.org/trunk/ao@1756 0101bb08-14d6-0310-b084-bc0e0c8e3800

files from being piped to other programs.  We now write the wav header
at the beginning with bogus (max signed 32 bit int - header size) size
info and go back and fix it if the file is seekable at the end.


git-svn-id: http://svn.xiph.org/trunk/ao@1733 0101bb08-14d6-0310-b084-bc0e0c8e3800

- Fixed problem caused by the OSS emulation in ALSA.  open() would block
if the dsp was already in use, causing the driver autodetect code to
stall.


git-svn-id: http://svn.xiph.org/trunk/ao@1718 0101bb08-14d6-0310-b084-bc0e0c8e3800

I was on when I did that earlier.


git-svn-id: http://svn.xiph.org/trunk/ao@1681 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1677 0101bb08-14d6-0310-b084-bc0e0c8e3800

the arts Makefile from clobbering custom CFLAGS.


git-svn-id: http://svn.xiph.org/trunk/ao@1540 0101bb08-14d6-0310-b084-bc0e0c8e3800

git-svn-id: http://svn.xiph.org/trunk/ao@1467 0101bb08-14d6-0310-b084-bc0e0c8e3800