Ramiro Polla ramiro.polla
Thu Jun 4 23:33:50 CEST 2009

On Thu, Jun 4, 2009 at 4:03 PM, Pavel Pavlov <pavel at summit-tech.ca> wrote:
>> -----Original Message-----
>> We prefer patches made with svn diff from the root directory,
>> so they can be applied with patch -p0 < blah. I for example
>> keep one clean svn tree just for the purpose of creating
>> diffs if they're part of some big changes, and it really
>> helps us out in applying & reviewing patches.
>> > Let me know if what I have sent is unusable so I wouldn't
>> waste my and
>> > others' time...
>> Your patches are welcome, but they'll be more welcome after a
>> few minor tweaks (it's just standard procedure around here).
> Well, I have tortoise svn, and I also have two c:\FFMPEG and C:\-FFMPEG,
> the second one is
> unmodified version for creating the diffs and comparing my changes.
> On windows I right click and from svn submenu there is "Create patch
> option", that's what
> I did.

Hmm, I had no such problem when I used Tortoise SVN. IIRC you can
choose which files you want to include in the diff.

> Maybe you can manually modify diffs I sent so that they could be
> applied?
> From
> Index: C:/FFMPEG/libswscale/swscale_internal.h
> ===================================================================
> --- C:/FFMPEG/libswscale/swscale_internal.h ? ? (revision 29346)
> +++ C:/FFMPEG/libswscale/swscale_internal.h ? ? (working copy)
> To
> Index: libswscale/swscale_internal.h
> ===================================================================
> --- libswscale/swscale_internal.h ? ? ? (revision 29346)
> +++ libswscale/swscale_internal.h ? ? ? (working copy)
> Is that's all required?

This way, every time someone takes this patch and tries to apply it,
they'll first see it failing, then have to go through the patch file
by file and fix accordingly. If it gets sent with the standard "svn
diff" from root source (or you fix it yourself before sending), this
is not necessary. So fixing it once by the sender is preferred over
fixing it every time by the reader.

Ramiro Polla

More information about the ffmpeg-devel mailing list