[FFmpeg-devel] [PATCH] RTSP muxer, round 3

[Ronald S. Bultje]

Hi,

On Mon, Feb 8, 2010 at 10:00 AM, Martin Storsj? <martin at martin.st> wrote:
> On Sun, 31 Jan 2010, Martin Storsj? wrote:
>> On Sat, 30 Jan 2010, Ronald S. Bultje wrote:
>> > On Fri, Jan 22, 2010 at 3:17 AM, Martin Storsj? <martin at martin.st> wrote:
>> > > On Wed, 20 Jan 2010, Ronald S. Bultje wrote:
>> > >> (Will look at the rest in a bit.)
>> > >
>> > > Here's an updated series, to work with the latest SVN. 0001 is updated to
>> > > fit with the updated parameter signature to rtsp_send_cmd_async, and 0010
>> > > is updated to use the _async version for sending the commands in
>> > > rtsp_write_close (as you did for rtsp_read_close), the rest is unmodified.
>> >
>> > General comments:
>> > - should we split this in 3 files (rtsp, rtspdec, rtspenc.c) for
>> > easier maintenance?
>>
>> That might be a good idea. Would it be ok to defer that to after getting
>> this applied?
>>
>> > - the priv_data is not the same for demuxer and muxer, should we
>> > document that? Did you double check all possible cases that this is
>> > handled OK?
>>
>> Documented it. Double checked - transport_priv is used only in the
>> setup/close routines, where is_output is checked, or in read/write
>> functions (rtsp_fetch_packet, rtsp_write_packet), where only either one of
>> the alternatives is applicable.
>>
>> Updated patch series attached. Except for rebasing it onto the latest
>> version and adding the documentation mentioned above, I check the lower
>> transport mask in 0009, since only UDP output is supported at the moment.
>
> Ping

You need an if (is_output) around the newly added NAT puncher. :-P.

And yes I promise I'll look at this again soon (this week), I must
honestly admit I didn't expect so many patches (most from you :-) ) so
I'm a bit overwhelmed ATM.

Ronald