e78acce652
Don't mix old- and new-style FIND_PACKAGE_HANDLE_STANDARD_ARGS calls
...
Some (all?) versions of cmake seem to get a bit confused when an old-style
FIND_PACKAGE_HANDLE_STANDARD_ARGS call fails, and the scripts attempt a new-
style call later on. It continues to think it's using old-style, acting as if
REQUIRED_VARS is the message and looks for a VERSION_VAR variable.
2014-05-10 09:34:16 -07:00
ddd1550a8b
Avoid locking the mixer while queueing buffers onto a source.
2014-05-10 08:55:28 -07:00
e9297d9744
Mark the buffer queue pointers volatile
2014-05-10 08:03:53 -07:00
3e28fb33a3
Add a missing pointer init
2014-05-10 07:57:35 -07:00
9e25ea8666
Update the output buffer pointer in the Write_* methods
2014-05-10 07:52:02 -07:00
67b4bd1031
Add a couple consts
2014-05-10 07:25:05 -07:00
9642bf637d
Explicit init the source queue and current_buffer pointers
2014-05-10 06:32:39 -07:00
a455704148
Use a RWLock to help protect the source's buffer queue
...
In some instances this allows to to remove the device/mixer lock, or reduce how
long it's held.
2014-05-10 05:07:13 -07:00
eebde08e65
Don't explicitly store the BuffersInQueue
2014-05-10 03:33:41 -07:00
c4383b65e2
Store the current buffer queue item, rather than played buffer count
2014-05-10 03:21:40 -07:00
eea46f268d
Avoid checking the previous buffer multiple times
2014-05-10 02:01:38 -07:00
bf97f58edb
Append the number of bits, not bytes, to types
...
Also use ALuint64SOFT from alext.h for uint64
2014-05-08 00:28:34 -07:00
8b7c71e20e
Always use unsigned int for RefCount
2014-05-07 23:49:06 -07:00
bd88e1743a
Build a common static lib for reusable code
2014-05-07 19:51:42 -07:00
a2bddb7b40
Move RWLock and UIntMap implementations to common
...
This should make the code in common completely self-reliant.
2014-05-07 19:16:49 -07:00
3e274c3a57
Make sure FILETIME and ULARGE_INTEGER sizes match
2014-05-07 16:12:34 -07:00
1c01e94237
Move the static_assert definition to its own header
2014-05-07 02:28:25 -07:00
d8d609b940
Use standard types for the RWLock
2014-05-07 01:03:06 -07:00
38b98de78e
Check for C99 _Bool support
2014-05-06 23:10:50 -07:00
1aff37114a
Move atomic method definitions to a separate common source
2014-05-06 19:07:40 -07:00
fa12855d19
Move threads.c to a separate source dir
...
This will eventually serve to build a static lib of common wrapper methods,
such as threads, mutexes, atomics, etc.
2014-05-06 18:57:42 -07:00
0ea979a262
Move some headers to include/
...
Note, these are not installed. Only headers in include/AL/ are installed.
2014-05-06 18:29:53 -07:00
1a76a3238c
Use gettimeofday if clock_gettime isn't available
2014-05-06 17:46:17 -07:00
cbb94405db
Test _Alignas with 16 instead of 4
2014-05-05 14:21:09 -07:00
2edd047178
C11's alignas is also sufficient for SSE aligning
2014-05-05 13:53:07 -07:00
2bc16e9947
Add a cast to silence an MSVC warning
2014-05-04 11:50:58 -07:00
a48d235ef4
Trace skipped info sub-chunks
2014-05-04 01:19:09 -07:00
bca7c682f1
Use ALuint for the amount to skip
2014-05-04 01:00:21 -07:00
2a7f3c015e
Don't use the next power of two when duplicating config values
...
Speed isn't terribly important here, and reallocs should be extremely rare.
2014-05-04 00:37:25 -07:00
db90985f6a
Remove unnecessary ifdefs
...
mixer_sse.c and mixer_neon.c are only compiled when the relavent headers are
found anyway.
2014-05-04 00:19:55 -07:00
a2bb4c3f13
Always use the current gains when mixing
...
The current gain gets explicitly set to the target when the stepping is
finished to ensure the target is still used. This way, however, will allow for
asynchronously 'canceling' a fade by setting the counter to 0.
2014-05-04 00:13:19 -07:00
d066c7b124
Better pack HRTF mixing properties
2014-05-03 18:59:26 -07:00
b2e533fbfc
Clamp the current and target gain lower bound to epsilon
...
Should give a bit more wiggle room for the gain stepping to get lower than the
silence threshold.
2014-05-03 17:51:06 -07:00
34e96aef60
Make sure all gain steps are applied with the SSE and Neon mixers
2014-05-03 17:24:46 -07:00
2660ecfd4a
Revert e9aee578a7
...
The vector iterator needs to be the start of the Data array even if the vector
itself is NULL, or else insertion can have a negative insert position.
2014-05-02 11:48:55 -07:00
3bdb3f9db7
Iniitialize some mmdevapi backend fields in the constructor
2014-05-02 10:47:31 -07:00
d90f684f16
Use a backup in case pthread_mutex_timedlock isn't available
2014-05-01 14:15:31 -07:00
804b08e0ca
Use ALboolean for al_string_empty's return type
2014-04-30 12:53:52 -07:00
f150676873
Add a helper VECTOR_FOR_EACH macro
2014-04-30 12:30:02 -07:00
f9d70aa9ef
Use _mm_set_ps() to set an __m128 instead of {}
2014-04-26 06:31:31 -07:00
149df6d4db
Convert the mmdevapi backend to the new backend API
2014-04-25 09:26:53 -07:00
28e9f0db8c
Use a vector for mmdevapi device lists
2014-04-24 04:02:05 -07:00
83970369e0
More device list cleanup cleanup
2014-04-24 02:33:54 -07:00
7dd4a2348f
Use a helper function to clear PulseAudio device lists
2014-04-24 01:38:24 -07:00
a4cf378af8
Use a helper method to clear winmm device lists
2014-04-24 01:25:26 -07:00
9af8e10c20
Avoid using superfluous LP* types
2014-04-23 10:08:33 -07:00
0b527621a5
Use a RefCount for counting committed winmm buffers
2014-04-23 05:08:17 -07:00
8fd224f841
Avoid forward-declaring backend vtables
2014-04-23 03:52:05 -07:00
420599f8e3
Avoid using assert() in createBackend
2014-04-23 02:18:07 -07:00
7fed6383c0
Include FindALSA.cmake
...
Older versions of CMake don't seem to have it.
2014-04-23 02:12:34 -07:00
Chris Robinson