Stefan Seyfried 8ef9807b87 spark: make cRecord more error resilient
* use aio to improve things for slow recording media (heavily loaded
 NFS server for example)
* in case of a buffer overflow, don't stop the recording but simply
 drop a buffer
A proper aio implementation would involve multiple buffers that could
be queued in paralled, but that's much more complex, so let's see if
the current code works out well enough.


Origin commit data
------------------
Branch: master
Commit: f3d93833e1
Author: Stefan Seyfried <seife@tuxbox-git.slipkontur.de>
Date: 2012-04-08 (Sun, 08 Apr 2012)

Origin message was:
------------------
spark: make cRecord more error resilient

* use aio to improve things for slow recording media (heavily loaded
 NFS server for example)
* in case of a buffer overflow, don't stop the recording but simply
 drop a buffer
A proper aio implementation would involve multiple buffers that could
be queued in paralled, but that's much more complex, so let's see if
the current code works out well enough.


------------------
This commit was generated by Migit
2012-04-08 14:43:40 +02:00
Languages
C 58.3%
C++ 39.8%
M4 1.1%
Makefile 0.7%