regcomp() was failing to parse them. This changeset improves the algorithm
by using less assumptions.
This commit is a backport of similar (but bigger) commit in jack2.
b3394f4dce
for python loading libjackserver and using controlAPI.
drivers are linked to a second instance of libjackserver.so and the
global symbol _jack_get_microseconds showed up uninitialised.
by providing a function pointer to right instance in the engine struct,
we can fix this problem. since _jack_get_microseconds is a in systemdeps,
we need a generic way to get the right function pointer.
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@4421 0c269be4-1314-0410-8aa9-9f06e86f4224
this is a quick fix for the problem reported by Felix Pfeifer.
whitespace cleanup and more thorough fixes of this code follow.
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@4193 0c269be4-1314-0410-8aa9-9f06e86f4224
connection information can be used again.
* add disgusting kludge so that "ALSA:...." port names continue to
work, thus not breaking connection information for 0.109
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@1086 0c269be4-1314-0410-8aa9-9f06e86f4224
Description:
In r1051 jack_engine_munge_backend_port_names() is being called during
startup. ALSA raw midi backend does port discovery in dynamic way and
thus midi ports appear after call to
jack_engine_munge_backend_port_names() and thus are not munged.
This patch contains:
* munge during port registration not during startup. It is done only
for internal clients. MIDI ports are midi_capture_N and
midi_playback_N in order to not interfere with audio port names.
* ALSA backend name is changed from "alsa_pcm" to "ALSA"
* engine->driver is initialized before call to driver attach() method
to make munge during port registration work during startup.
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@1071 0c269be4-1314-0410-8aa9-9f06e86f4224
* COMPLEX_MMAP patch for ALSA driver from Frank van der Pol
* BIG_ENDIAN patch by A. Steinmetz with followup patch by M. Habets
for ALSA backend (allows JACK to run on devices like M-Audio
Audiophile USB that only support big endian sample formats),
modified and fixed by me
* remove unused argument from jack_shmalloc to ease x86_64 errors
* fixes to printf formatting and on x86_64
* fix argument signedness warnings from gcc4
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@945 0c269be4-1314-0410-8aa9-9f06e86f4224
This is done by hacking driver->capture_interleave_skip
in usx2y_driver_start(). Other changes in usx2y.c improve
rawusb mode debugging.
An "if (unlikely(x))" optimization commonly found in kernel
code, applied to alsa_driver_run_cycle(), as proposed by
Lee Revell.
Signed-off-by: Karsten Wiese <annabellesgarden@yahoo.de>
Signed-off-by: Rui Nuno Capela <rncbc@rncbc.org>
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@918 0c269be4-1314-0410-8aa9-9f06e86f4224
This adds to general correctness as it passes the error status to the caller, which does not happen for the nframes == 0 case without it.
Signed-off-by: Karsten Wiese <annabellesgarden@yahoo.de>
Signed-off-by: Rui Nuno Capela <rncbc@rncbc.org>
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@912 0c269be4-1314-0410-8aa9-9f06e86f4224
+ format search problem introduced yesterday
+ bug:00031 "crash when using more than 32 channels"
git-svn-id: svn+ssh://jackaudio.org/trunk/jack@864 0c269be4-1314-0410-8aa9-9f06e86f4224
falkTX
Paul Davis
Nedko Arnaudov
torben
paul
sletz
marco
nedko
pieterpalmers
dsbaikov
pbd
rncbc
essej
joq