Support the URL_FLAG_NONBLOCK semantic and uniform the protocol.
The quick retry loop is already part of retry_transfer_wrapper.
The polling routine is common to the network protocols:
udp, tcp and, once merged, sctp.
Functions interrupted by url_interrupt_cb should not be restarted.
Therefore using AVERROR(EINTR) was wrong, as it did not allow to distinguish
when the underlying system call was interrupted and actually needed to be
restarted.
This fixes roundup issues 2657 and 2659 (ffplay not exiting for streamed
content).
Signed-off-by: Nicolas George <nicolas.george@normalesup.org>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Map EAGAIN and EINTR from ff_neterrno to the normal AVERROR()
error codes. Provide fallback definitions of other errno.h network
errors, mapping them to the corresponding winsock errors.
This eases catching these error codes in common code, without having
to distinguish between FF_NETERRNO(EAGAIN) and AVERROR(EAGAIN).
This fixes roundup issue 2614, unbreaking blocking network IO on
windows.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
This makes it possible to abort a blocking connect call.
Patch by Thomas Guillem, thomas dot guillem at gmail
Originally committed as revision 26014 to svn://svn.ffmpeg.org/ffmpeg/trunk
On windows, a connection failure doesn't trigger wfds as it does on unix.
This fixes issue 2237, based on code by yeyingxian.
Originally committed as revision 25154 to svn://svn.ffmpeg.org/ffmpeg/trunk
ff_url_split() is retained as an alias, as it was used by ffserver,
to avoid breaking ABI compatibility with it.
Originally committed as revision 23822 to svn://svn.ffmpeg.org/ffmpeg/trunk
If ff_neterrno() is checked instead of errno, that's probably what should be
used in the return value.
As a sideeffect, this gives better compatibility with Windows (CE), where
network errors aren't visible in errno.)
patch by Martin Storsjö, martin martin st
Originally committed as revision 19433 to svn://svn.ffmpeg.org/ffmpeg/trunk
associated with the I/O handle (e.g. the fd returned by open()). See
"[RFC] rtsp.c EOF support" thread.
There were previously some URI-specific implementations of the same idea,
e.g. rtp_get_file_handles() and udp_get_file_handle(). All of these are
deprecated by this patch and will be removed at the next major API bump.
Originally committed as revision 17779 to svn://svn.ffmpeg.org/ffmpeg/trunk
actual hostname. This functionality already exists (and always existed) in
url_split() and is therefore useless. See discussion in "[PATCH] tcp.c/udp.c
memleak?" thread on ffmpeg-devel.
Originally committed as revision 14946 to svn://svn.ffmpeg.org/ffmpeg/trunk
only allocated if initialization worked. This means that on failure, we
don't have to deallocate it.
Originally committed as revision 14924 to svn://svn.ffmpeg.org/ffmpeg/trunk
AVERROR_ defines are moved to avcodec.h as they are needed in there as well. Feel free to move that to avutil/common.h.
Bumped up avcodec/format version numbers as though it's binary compatible we will want to rebuild apps as error values changed.
Please from now on use return AVERROR(EFOO) instead of the ugly return -EFOO in your code.
This also removes the need for berrno.h.
Originally committed as revision 7965 to svn://svn.ffmpeg.org/ffmpeg/trunk
Patch by Ramiro Polla % ramiro A lisha P ufsc P br %
Original thread:
date: Jan 20, 2007 8:56 PM
subject: [Ffmpeg-devel] [PATCH] [RFC] change socket read and write functions
Originally committed as revision 7613 to svn://svn.ffmpeg.org/ffmpeg/trunk
tested and submitted by (Torsten Spindler <spindler at hbt dot arch dot ethz dot ch>)
Originally committed as revision 3381 to svn://svn.ffmpeg.org/ffmpeg/trunk
tcp/http in ffmpeg.c patch by (Leon van Stuivenberg <l dot vanstuivenberg at chello dot nl>)
Originally committed as revision 2891 to svn://svn.ffmpeg.org/ffmpeg/trunk
Whenever there was an EINTR/EAGAIN return, then a byte in the data stream
would be duplicated!! This fix should allow ffserver to work again.
Originally committed as revision 2317 to svn://svn.ffmpeg.org/ffmpeg/trunk
