[Ffmpeg-devel] [PATCH] http_seek feature added

[Michael Niedermayer]

Hi

On Wed, Jul 19, 2006 at 02:52:00PM +0000, I. Po wrote:
> Hi,
> On 7/18/06, Michael Niedermayer <michaelni at gmx.at> wrote:
> 
> >why not move http_connect_loop before http_open so this wont be needed?
> I think, it's better to move http_open to the rest of function,
> declared in URLProtocol

moving functions around is a non functional change, and these must be in
a seperate patch (makes reviewing and reading svn log easier)

[...]

> I corrected the patch, but it have a side effect: if
> URLContext->is_streamed field set to 0 (like in patch), ffserver fails
> last step of a regression test. If I set it to 1, like before, ffmpeg
> can't use seek with http protocol, but ffserver pass all tests. I
> think it isn't a good idea to apply such patch, but it's better to try
> to know, why ffserver fails. It's strange for me, because it means,
> ffserver tryes to make some seeks in stream, if it possible. As I
> understand, ffserver works through http always and previously it
> doesn't seek (is_streamed was == 1).
> I don't have now a solution of this trouble, so I quoted code, which
> enables the seek feature, with #ifdef brackets. By default, it's
> disabled.

hmm, thats a problem, if the code is disabled like that its useless
and breaking ffserver is not good either, i guess you will have to
fix ffserver so it doesnt fail before this patch can be applied

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

In the past you could go to a library and read, borrow or copy any book
Today you'd get arrested for mere telling someone where the library is