[FFmpeg-devel] Broken third party software

[Michael Niedermayer]

On Sun, Sep 14, 2008 at 03:15:25PM +0200, Diego Biurrun wrote:
> On Sun, Sep 14, 2008 at 03:08:53PM +0200, Michael Niedermayer wrote:
> > On Sun, Sep 14, 2008 at 02:53:40PM +0200, Luca Abeni wrote:
> > > 
> > > Ramiro Polla wrote:
> > > [...]
> > > >>    svn annotate or svn log can easily point to r15262 as the version which
> > > >>    broke the old API, so developer of applications using libavcodec can
> > > >>    point to r15261 until their programs are updated.
> > > >>    These changes do not happen very often (I think the last one was some
> > > >>    years ago?)
> > > > 
> > > > I think we could do other developers a favor and at least announce
> > > > somewhere that there was a major version bump (and that if I
> > > > understand correctly more things might still change in the very near
> > > > future).
> > > > 
> > > > I attached a patch for index.html, but maybe it's not the best place
> > > > (and the text could be rewritten)...
> > > 
> > > This looks like a good idea.
> > > Maybe we could go even further, and do something like
> > > svn cp -r r15261 svn://svn.mplayerhq.hu/ffmpeg/trunk svn://svn.mplayerhq.hu/ffmpeg/tags/libavcodec_v51
> > > or similar (I did not check the correct svn syntax and similar
> > > details).
> > > Note: this is just an idea, if other people think it is bad, I
> > > am not going to defend it too strongly.
> > 
> > Iam not against it but i must note that the "tags are copies" design is a
> > step back from cvs.
> 
> Uh, have you ever tagged something (big) with CVS?  I had Godot pay a
> visit, have a drink and leave again while waiting for the CVS tag command
> to finish doing its magic to MPlayer ...

having to wait to create a tag isnt that bad IMHO because its not done that
often (its the first time since we switch to svn ...), compared to the extra
disk space checkouts of the tags directory would need.

But as already said iam not against it, users shouldnt check out the tags
directory in theory ...

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

Complexity theory is the science of finding the exact solution to an
approximation. Benchmarking OTOH is finding an approximation of the exact
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080914/9c3c43c8/attachment.pgp>