[FFmpeg-devel] Experimental MSVC port

[Ole Andre Vadla Ravnås]

> -----Original Message-----
> From: ffmpeg-devel-bounces at mplayerhq.hu [mailto:ffmpeg-devel-
> bounces at mplayerhq.hu] On Behalf Of Reimar D?ffinger
> Sent: Wednesday, March 26, 2008 3:53 PM
> To: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] Experimental MSVC port
> 
> On Wed, Mar 26, 2008 at 02:18:45PM +0100, Ole Andr? Vadla Ravn?s wrote:
> > FWIW, for those of you interested in building a somewhat semi-recent
> ffmpeg,
> > r11247 (the revision currently suggested by GStreamer's gst-ffmpeg)
> with
> > MSVC15, I've ported and included it in OABuild:
> >
> > http://people.collabora.co.uk/~oleavr/OABuild/
> >
> > Where the ffmpeg bazaar branch is at:
> > http://people.collabora.co.uk/~oleavr/OABuild/bzr/ffmpeg/
> 
> A diff (possibly just excerpts) or at least a web interface would
> probably be more useful, at least I don't expect many will learn bazaar
> usage just for this...

Not a bad idea, I'll see about creating a project at launchpad, google
code or similar at some point. For now each branch comes with a quilt
patch stack. I'm not entirely happy with the MSVC fixes as they stand
though. The patch should also be split up into general ffmpeg fixes like
not calling functions in disabled codecs (I'm guessing this isn't that
big an issue with gcc as branch elimination takes care of removing those
calls, i.e.
if (ENABLE_FOO_BAR && ...)
  call_foo_bar
means
if (0 && ...)
  call_foo_bar
and the branch gets eliminated.)
Anyway, I've attached the current patch. The bulk of changes are about
C99, which is quite amusing how Microsoft, with a compiler released
almost ten years later, still doesn't support much of. The biggest
changes are in imgconvert.c, where I ended up writing a quick and dirty
python script to parse the array initialization code and transform them
into the non-C99 equivalents.

> > At the moment it only includes H263 and H263+ encoding/decoding, but
> I'm
> > planning on porting codecs on an as-needed basis (doesn't appear to
> be
> > particularly much work per codec).
> 
> If there is any per-codec porting involved that sounds really, really
> bad to me...

Not much, just bits and pieces of C99, very trivial stuff.

Ole Andr?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: msvc_fixes.bz2
Type: application/octet-stream
Size: 14465 bytes
Desc: msvc_fixes.bz2
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080326/16a8ba2c/attachment.obj>