Commit graph

1348 commits

Author SHA1 Message Date
Wim Taymans
0694be2a82 pulsesink: use acquired flag when checking valid state
Use the acquired field of the ringbuffer in get_time to know when we are in an
invalid state. We don't clear the rate flag when releasing the ringbuffer so
this values is not usable.

Avoids some error messages being posted because the pulseaudio connection is
down.
2009-11-18 17:27:19 +01:00
Wim Taymans
f85cf3e200 pulsesink: implement some more callbacks
Implement some more callbacks for debugging purposes.
2009-11-18 12:01:52 +01:00
Tim-Philipp Müller
7af7d93b27 jpegdec: free temporary buffer when changing state to NULL
Free temporary allocations in the state change function and not
only when the object is finalised.
2009-11-09 15:20:00 +00:00
Tim-Philipp Müller
26e7fe27d7 jpegdec: only allocate as much temporary memory as needed for indirect decoding
When we can't decode directly into the output buffer, make our temp buffers
only as big as needed instead of allocating for the worst case scenario (well,
we still alloc more than strictly needed for some cases, but significantly
less than before).
2009-11-09 15:18:58 +00:00
Tim-Philipp Müller
2270d3359a pulsesink: printf format fix 2009-11-05 23:46:58 +00:00
Tim-Philipp Müller
f50f516eca raw1394: printf format fixes 2009-11-05 23:44:27 +00:00
Stefan Kost
e3e910e971 jpegdec: don't allocate big arrays on the stack
Add the arrays to the instance data and allocate on first use.
2009-11-03 17:49:24 +02:00
Sebastian Dröge
3f541452c4 pulsesink: Only set the volume on stream connection if pulse >= 0.9.20 is available
In older versions the volume set during stream connection had
no defined sematic and usually it was a relative volume. What
was needed for our use case is an absolute volume though, otherwise
the volume will be always decreased on stream connection if it's
less than 100%.
Since pulse 0.9.20 that volume is always an absolute volume if
flat volumes are used and relative otherwise, which is the same
as for pa_context_set_sink_input_volume().

Relevant pulse changesets:
http://git.0pointer.de/?p=pulseaudio.git;a=commit;h=f27a50691c8fe45bac7dd6b21fac91a359def3a1
http://git.0pointer.de/?p=pulseaudio.git;a=commit;h=2501687579e359d5032a4d165b2ffc8f5b1b8ba6
2009-10-30 09:27:59 +01:00
Priit Laes
34a7a90f57 cacasink: minor cleanups for header.
Use G_BEGIN_DECLS macros, remove unused variables and fix typo.
See #599018.
2009-10-22 02:03:12 +01:00
Priit Laes
45042dffbc cacasink: exit properly when invalid driver has been selected.
See #599018.
2009-10-22 02:03:01 +01:00
Bastien Nocera
a1ac55c7a5 Fix the StreamVolume interface not being advertised
gst_pulsesink_interface_supported() was missing a check for it.

https://bugzilla.gnome.org/show_bug.cgi?id=598933
2009-10-19 15:36:20 +01:00
Lennart Poettering
e4d6a2aa2c pulse: never apply volume more than once
Generally decisions on the volume of the stream should be done inside of
PA, not inside of Gst. Only PA knows how volumes translate between
devices and s on.

This patch makes sure that all volumes set via the volume property are
only applied *once* to the underlying stream. After applying them the
client side will not store them anymore. This should make sure that
really only user-triggered volume changes are forwarded to server, but
the client never tries to save/restore the volume internally.

Fixes bug #595231.
2009-10-18 10:15:59 +02:00
Sebastian Dröge
92fd59024b pulsesink: Initialize gettext for the translated strings in plugin_init() 2009-10-17 09:00:00 +02:00
Lennart Poettering
89b0c2fc4f pulse: use 'performer' as a fallback for 'artist' tag 2009-10-17 08:48:23 +02:00
Lennart Poettering
76c8ca4b00 pulse: when constructing a stream title from tag data make sure it is translatable 2009-10-17 08:48:23 +02:00
Lennart Poettering
fd75286dd1 pulse: loop while connecting to server
pthread does not guarantee that there are no spurious condition variable
wakeups, neither does pa_threaded_mainloop_xxx() which is a wrapper
around it. So we need to loop around the _wait() function to make sure
we get the right wakeup.

Also, unify the order of the wait loops across the file.
2009-10-17 08:48:23 +02:00
Lennart Poettering
48c3bd303f pulse: mainloop creation can fail too, so handle that 2009-10-17 08:48:23 +02:00
Lennart Poettering
3188f5699b pulse: adjust CHECK_DEAD_GOTO macro to glib style 2009-10-17 08:48:23 +02:00
Lennart Poettering
d64200f46b pulse: make a few things smaller by making them bitfields 2009-10-17 08:48:21 +02:00
Tim-Philipp Müller
c83ea23a03 jpegdec: fix branch hints
Remove inappropriate branching hints and add some new ones.
2009-10-16 13:41:45 +01:00
Tim-Philipp Müller
4fb3f2519c jpegdec: fix regression in indirect decode path
Revert variable name back to what it was before the G_LIKELY was
added (in commit 69c24fb9). The code works better that way.
2009-10-16 13:36:42 +01:00
Tim-Philipp Müller
0c92e70f6c jpegdec: fix regression with certain formats
Fix regression introduced by previous commit (#598517).
2009-10-16 13:00:12 +01:00
Tim-Philipp Müller
30cb82da96 jpegdec: don't use decompress structure members we shouldn't be using 2009-10-16 13:00:12 +01:00
Tim-Philipp Müller
19b4f66014 jpegdec: remove some unused members from jpegdec instance structure 2009-10-16 13:00:12 +01:00
René Stadler
f9274226e8 pulsesink: set desired minreq value to segsize/latency-time
If we let the daemon decide freely by passing -1, we end up always getting 20ms.
We want to set this value because in some cases we want to select a higher
latency-time in order to save power.

Fixes #597601
2009-10-14 12:05:14 +02:00
Tim-Philipp Müller
87bd79f4ba jpegdec: fix bogus warning about discont flag on first buffer
The very first buffer should always have the DISCONT flag set, no
need to warn about that. Only warn if we get a DISCONT buffer in
non-packetised mode and we already have some data.
2009-10-13 01:10:59 +01:00
Tim-Philipp Müller
6ae6af9dab jpegdec: fix crash for unusual vertical chroma subsampling factors
Fixes #597351.
2009-10-13 01:10:59 +01:00
Josep Torra
c581e6f205 jpegdec: fixes warning building in snow leopard 2009-10-10 00:37:08 +02:00
Stefan Kost
f1fe1f52fd flac: apparently on some platforms a FLAC__uint64!=guint64 2009-10-09 14:37:32 +03:00
Stefan Kost
f41d7e7bd5 build: don't cast, but use the right format specified instead
This correct some of the previous macos fixes.
2009-10-09 13:54:24 +03:00
Josep Torra
696e350c7b dv: fix warnings on macosx 2009-10-09 12:40:47 +02:00
Josep Torra
133ebe76a3 flac: fix warnings on macosx 2009-10-09 12:25:19 +02:00
Josep Torra
6320de853b annodex: fix warnings in macosx 2009-10-09 12:19:35 +02:00
Stefan Kost
e0cdd879b4 build: fprintf, sprintf, sscanf need stdio.h 2009-10-07 14:03:20 +03:00
René Stadler
15c6175044 pulsesrc: guard fragment size with a lower limit based on latency-time
In case that the pulse daemon runs the source device at a relatively low fixed
fragment size compared to the requested latency-time, configure the ring buffer
segsize to the largest integer multiple of the fragment size that is still
smaller than or equal to the requested latency-time.

Fixes bug #597463.
2009-10-06 20:54:20 +03:00
Stefan Kost
86b8935b07 jpegdec: comment/logging cleanups and more branch guides 2009-10-06 17:46:49 +03:00
René Stadler
ccddf0643d pulse: rename pa_buffer_attr variables
Makes it much easier to see what is going on and is a lot less error prone.
2009-10-02 23:21:02 +03:00
Christian F.K. Schaller
e2623815d1 Update makefile with missing header file 2009-09-30 18:06:07 +01:00
Stefan Kost
b1feeee166 jpegdec: don't leak output buffers on decoding errors
The setjmp handles libjpeg error. Free the outputbffer if we don't need it.
2009-09-30 09:19:48 +03:00
Tim-Philipp Müller
477cb58640 jpegdec: fix 'unused variable' compiler warning when compiling with GST_DISABLE_GST_DEBUG 2009-09-29 00:01:59 +01:00
Stefan Kost
69c24fb991 jpeg: handle more libjpeg return values, add some more branch hints
Also remove unused size variable in _chain().
2009-09-28 17:25:35 +03:00
Alessandro Decina
195883b30a Fix compile warnings with gcc 4.0.1. 2009-09-22 15:04:36 +02:00
Stefan Kost
253f43203d jpegdec: add a G_UNLIKELY and put perf-cat log to code path that copies 2009-09-21 16:07:50 +03:00
Mark Nauwelaerts
50c41acdff speexdec: allow for unknown varying number of frames per buffer
In particular, this caters for RTP payloads with multiple frames
per packet.
2009-09-18 14:46:38 +02:00
Mark Nauwelaerts
35f9efac88 speexdec: use correct sample size in conversions 2009-09-18 14:46:38 +02:00
Mark Nauwelaerts
f4482e4ae3 speexenc: fix buffer time and duration for multiple frames per packet 2009-09-18 14:46:38 +02:00
Sebastian Dröge
317d6e18a2 pulsesrc: Don't dereference NULL pointers
pa_stream_get_timing_info() can return NULL.

Fixes bug #595220.
2009-09-15 10:05:51 +02:00
David Henningsson
bb787f1030 pulsesink: Don't dereference NULL pointers
pa_stream_get_timing_info() can return NULL.

Fixes bug #595220.
2009-09-15 10:05:51 +02:00
Wim Taymans
ccda761562 pulsesink: handle stream events
Handle stream events and request a PAUSE/PLAY state change from the application
when we receive a CORK/UNCORK event.
2009-09-14 16:05:30 +02:00
David Schleef
903c79e767 dvdemux: Use values from decoder structure directly
Don't store the same values in the GstDvDemux.  This
fixes a bug where dvdemux would detect a stream as PAL
instead of NTSC, and silently parse it wrong.
2009-09-13 19:59:45 -07:00