[FFmpeg-devel] [PATCH] Make AVRtpPayloadTypes static and constant
Sat Jan 5 18:58:49 CET 2008
On Sat, Jan 05, 2008 at 06:48:44PM +0100, Luca Abeni wrote:
> Hi Michael,
> On Sat, 2008-01-05 at 18:20 +0100, Michael Niedermayer wrote:
> > > > I'm all in favour of this change
> > >
> > > Double thinking about it, I have some doubts about the "const" part. I
> > > agree that right now AVRtpPayloadTypes is not modified by ffmpeg's code,
> > > but I suspect that if rtp_get_payload_type() does not find a PT and a
> > > dynamic PT is used, then the corresponding entry in AVRtpPayloadTypes
> > > should be modified... Otherwise, there is no point in having the entries
> > > > 35 in the table.
> > It is absolutely not possible to change any entry in any global or static
> > table. This would totally break multithreaded apps.
> Uhmmm... You are right (as usual :)... But then I think
> AVRtpPayloadTypes is misdesigned: why having 128 entries? And, what's
> the purpose of the "dynamic" entries from pt 96 to 127? (since the table
> cannot be modified, the only useful entries seem to be the first 35,
ask the one whos patch it was ...
> > That said, the table should be const. If the type is const or not as another
> > question.
> Well, in this case the "const" qualifier looks ok, and I think the patch
> can be applied.
then apply it!
or maybe we should give diego#2 svn write access, i would be ok with that as
well, if he agrees to our development policy ...
> > And rtp.h should not be installed IIRC
> > And any application depending on anything from it is broken, this stuff is
> > not part of the public API IMHO
> Some months ago I proposed to avoid installing rtp.h, and I was told
> that it is needed.
Needed by what?
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
No snowflake in an avalanche ever feels responsible. -- Voltaire
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: Digital signature
More information about the ffmpeg-devel