[MPlayer-cvslog] r19341 - in trunk/stream: Makefile asf_mmst_streaming.c asf_streaming.c network.c network.h pnm.c stream_ftp.c stream_netstream.c stream_rtsp.c stream_vstream.c tcp.c tcp.h

[Diego Biurrun]

Sorry for letting the issue rest for so long, I'm both busy and getting
distracted with other things..

On Sun, Aug 13, 2006 at 01:02:13PM +0200, Michael Niedermayer wrote:
> 
> On Sun, Aug 13, 2006 at 12:05:53PM +0200, Diego Biurrun wrote:
> > On Fri, Aug 11, 2006 at 08:34:26PM +0200, Michael Niedermayer wrote:
> > > 
> > > and do you have any arguments for svn merge which is the same as 
> > > local reverse + commit and against svn cp ?
> > 
> > Yes, svn merge works while svn cp doesn't.  With 'svn cp' you can only
> > duplicate objects within the repository.  If you attempt to copy a file
> > over itself Subversion will complain about "foo is in the way".
> > 
> > Don't take my word for it, try for yourself, I've made a copy of the
> > MPlayer repo to play with:
> > 
> >   svn checkout svn://svn.mplayerhq.hu/mplayer.foo/trunk
> 
> while it wasnt easy, the following worked
> 
> 1. change mplayer.c in an unacceptable way :)
> 2. svn ci mplayer.c
> 
> 3. svn rm mplayer.c
> 4. svn ci mplayer.c
> 5. svn cp -r19361 svn://svn.mplayerhq.hu/mplayer.foo/trunk/mplayer.c mplayer.c
> 6. svn ci mplayer.c
> 
> if there are no objections then ill add that to svn-howto.txt soon

I don't really like this solution because it requires two commits and it
will create a known-broken revision.  Without mplayer.c the project will
not compile, leading to avoidable problems and bogus bug reports.  Yes,
this does bite people, see

http://bugzilla.mplayerhq.hu/show_bug.cgi?id=557

I don't much mind having to go through another step when examining svn
annotate output.  I also don't think examining the diffs is that hard.

Neither solution is perfect, so we have to make a tradeoff here.  It
depends on what people value more.

Thoughts?  Opinions?

Diego