RE: switch to svn

SVN can use external source of files too, via the svn:external
properties. But AFAIK it can only do it from other SVN sources. But
that's very convenient in some cases. Not sure that's what you're
looking for.

-----Original Message-----
From: ffmpeg-devel-bounces <at> mplayerhq.hu
[mailto:ffmpeg-devel-bounces <at> mplayerhq.hu] On Behalf Of Joshua Varner
Sent: lundi 31 octobre 2005 18:15
To: FFMpeg development discussions and patches
Subject: Re: [Ffmpeg-devel] switch to svn

On 10/30/05, matthieu castet <castet.matthieu <at> free.fr> wrote:
> Attila Kinali wrote:
> >
> >>* merging stuff, so for example we can import the cvs or svn of lets
say
> >>libmpeg2 into mplayer while preserving the history, with cvs thats
easy
> >
> >
> > I dont know whether it is possible to merge svn repos into
> > svn, but it's possible with cvs repos, given that the repo files
> > are available.
> >
> I think he wants something like cp :
>
> copy (cp): Duplicate something in working copy or repository,
> remembering history.
> usage: copy SRC DST

Re: [PATCH] Add DESTDIR support to Makefiles

2005/10/19, İsmail Dönmez <ismail <at> uludag.org.tr>:
> On Wednesday 19 October 2005 10:50, Sam Hocevar wrote:
> > On Wed, Oct 19, 2005, İsmail Dönmez wrote:
> > > >    No. These -D flags are compile-time flags. DESTDIR is only specified
> > > > for "make install" and is set to nothing during the build phase.
> > >
> > > No but you change lib_dir, conf_dir in config.mak which is legal to use
> > > in code. But for now its not used at all.
> >
> >    Again: these variables and files are *not* changed when DESTDIR is
> > not specified, and DESTDIR should *not* be specified when compiling (and
> > it isn't in any of the distributions I know), so it does not affect code
> > at all.
> >
> > > But if someone uses that as variable in code your patch would break.
> >
> >    Uh, no it wouldn't.
>
> Uh,ok got it. It would only break if DESTDIR is set before configuring (isn't
> that still a bad practise? ;) ). Sorry for the misunderstanding. Well ok lets
> apply Sam's patch :)

Sometimes we need DESTDIR to be defined during configure & build in
order to use e.g. headers and libs from the previously build packages.
This is usual when building an entire system, like bsd port system and
maybe gentoo(somebody please check).

BTW what's the problem to hard code it into the install section of
Makefile, instead of using hack that will fail on something like

Re: [PATCH] Add DESTDIR support to Makefiles

Ivan Kalvachev wrote:
> Sometimes we need DESTDIR to be defined during configure & build in
> order to use e.g. headers and libs from the previously build packages.
> This is usual when building an entire system, like bsd port system and
> maybe gentoo(somebody please check).

Not really. Gentoo uses the installed headers and libs and would use 
DESTDIR just on install stage.

> 
> BTW what's the problem to hard code it into the install section of
> Makefile, instead of using hack that will fail on something like
> 
> ./configure
> DESTDIR=/tmp make install

make DESTDIR=${place} install

Is the wanted behavior. Since you compile it as it was and would install 
in the live system but you want to install on a separate dir to be 
packed for later usage.

> Anyway destdir was in my todo, gonna take a closer look.

Nice =)

lu

--

-- 

libavcodec option description typo fix

s/comapare/compare/

-- 
Antoine Cellerier
dionoea

Index: libavcodec/utils.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/utils.c,v
retrieving revision 1.160
diff -u -r1.160 utils.c
--- libavcodec/utils.c	20 Sep 2005 21:43:45 -0000	1.160
+++ libavcodec/utils.c	1 Nov 2005 16:45:29 -0000
 <at>  <at>  -701,7 +701,7  <at>  <at> 
 {"frame_skip_threshold", NULL, OFFSET(frame_skip_threshold), FF_OPT_TYPE_INT, DEFAULT, INT_MIN,
INT_MAX, V|E},
 {"frame_skip_factor", NULL, OFFSET(frame_skip_factor), FF_OPT_TYPE_INT, DEFAULT, INT_MIN,
INT_MAX, V|E},
 {"frame_skip_exp", NULL, OFFSET(frame_skip_exp), FF_OPT_TYPE_INT, DEFAULT, INT_MIN, INT_MAX, V|E},
-{"skipcmp", "frame skip comapare function", OFFSET(frame_skip_cmp), FF_OPT_TYPE_INT,
FF_CMP_DCTMAX, INT_MIN, INT_MAX, V|E, "cmp_func"},
+{"skipcmp", "frame skip compare function", OFFSET(frame_skip_cmp), FF_OPT_TYPE_INT,
FF_CMP_DCTMAX, INT_MIN, INT_MAX, V|E, "cmp_func"},
 {"border_mask", NULL, OFFSET(border_masking), FF_OPT_TYPE_FLOAT, DEFAULT, FLT_MIN, FLT_MAX, V|E},
 {"mb_lmin", NULL, OFFSET(mb_lmin), FF_OPT_TYPE_INT, DEFAULT, INT_MIN, INT_MAX, V|E},
 {"mb_lmax", NULL, OFFSET(mb_lmax), FF_OPT_TYPE_INT, DEFAULT, INT_MIN, INT_MAX, V|E},

_______________________________________________

Re: libavcodec option description typo fix

Hi,

On 11/1/05, Antoine Cellerier <dionoea <at> via.ecp.fr> wrote:
> s/comapare/compare/

Thanks, applied.

Guillaume
--
Reading doesn't hurt, really!
 -- Dominik 'Rathann' Mierzejewski

libavcodec help

This patch changes the libavcodec help output in ffmpeg from :
-bit_rate          E.VA.
-bit_rate_tolerance E.V..
-flags             EDVA.
-mv4               E.V.. use four motion vector by macroblock (mpeg4)
(...)

to :
-bit_rate <int>       E.VA.
-bit_rate_tolerance <int> E.V..
-flags <flags>        EDVA.
-mv4                  E.V.. use four motion vector by macroblock (mpeg4)
(...)

can be helpfull if you don't have a clue about what an option is
supposed to have as an argument. 

--

-- 
Antoine Cellerier
dionoea

Index: libavcodec/opt.c
===================================================================
RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/opt.c,v
retrieving revision 1.9
diff -u -r1.9 opt.c
--- libavcodec/opt.c	20 Sep 2005 12:57:20 -0000	1.9
+++ libavcodec/opt.c	1 Nov 2005 18:38:25 -0000
 <at>  <at>  -254,16 +254,45  <at>  <at> 

Re: libavcodec help

Hi,

On 11/1/05, Antoine Cellerier <dionoea <at> via.ecp.fr> wrote:
> This patch changes the libavcodec help output in ffmpeg from :
> -bit_rate          E.VA.
> -bit_rate_tolerance E.V..
> -flags             EDVA.
> -mv4               E.V.. use four motion vector by macroblock (mpeg4)
> (...)
>
> to :
> -bit_rate <int>       E.VA.
> -bit_rate_tolerance <int> E.V..
> -flags <flags>        EDVA.
> -mv4                  E.V.. use four motion vector by macroblock (mpeg4)
> (...)
>
> can be helpfull if you don't have a clue about what an option is
> supposed to have as an argument.

I like the idea. I'll commit this patch in the coming days if nobody
objects (and once the CVS is back to business (right now there's a
lock in /cvsroot/ffmpeg/ffmpeg/libavcodec which prevents any commit
from happening).

Guillaume
--
Reading doesn't hurt, really!
  -- Dominik 'Rathann' Mierzejewski

Re: libavcodec help

Hi,
On Tue, Nov 01, 2005 at 07:48:05PM +0100, Guillaume POIRIER wrote:
> I like the idea. I'll commit this patch in the coming days if nobody
> objects (and once the CVS is back to business (right now there's a
> lock in /cvsroot/ffmpeg/ffmpeg/libavcodec which prevents any commit
> from happening).

The idea might be good, the patch is not. It contains two cosmetics, and
esp. it uses asprintf, which is an non-portable GNU-extension.

Greetings,
Reimar Döffinger

Re: libavcodec help

Hi,

On 11/1/05, Reimar Döffinger <Reimar.Doeffinger <at> stud.uni-karlsruhe.de> wrote:
> Hi,
> On Tue, Nov 01, 2005 at 07:48:05PM +0100, Guillaume POIRIER wrote:
> > I like the idea. I'll commit this patch in the coming days if nobody
> > objects (and once the CVS is back to business (right now there's a
> > lock in /cvsroot/ffmpeg/ffmpeg/libavcodec which prevents any commit
> > from happening).
>
> The idea might be good, the patch is not. It contains two cosmetics, and
> esp. it uses asprintf, which is an non-portable GNU-extension.

I saw the cosmetics and I meant to remove it myself before committing it.

What function do you suggest instead of asprintf? snprintf?

Guillaume
--
Reading doesn't hurt, really!
  -- Dominik 'Rathann' Mierzejewski

Re: libavcodec help

Hi

On Tue, Nov 01, 2005 at 08:02:11PM +0100, Guillaume POIRIER wrote:
> Hi,
> 
> On 11/1/05, Reimar Döffinger <Reimar.Doeffinger <at> stud.uni-karlsruhe.de> wrote:
> > Hi,
> > On Tue, Nov 01, 2005 at 07:48:05PM +0100, Guillaume POIRIER wrote:
> > > I like the idea. I'll commit this patch in the coming days if nobody
> > > objects (and once the CVS is back to business (right now there's a
> > > lock in /cvsroot/ffmpeg/ffmpeg/libavcodec which prevents any commit
> > > from happening).
> >
> > The idea might be good, the patch is not. It contains two cosmetics, and
> > esp. it uses asprintf, which is an non-portable GNU-extension.
> 
> I saw the cosmetics and I meant to remove it myself before committing it.
> 
> What function do you suggest instead of asprintf? snprintf?

none

just write something like

switch(opt->type){
case FF_OPT_TYPE_FLAGS: str="<flags>"; break;
...
}
av_log(..., str, ...)