[FFmpeg-cvslog] vf_overlay: prevent premature freeing of cur_buf

[Michael Niedermayer]

On Fri, Jul 27, 2012 at 10:13:38AM +0200, Nicolas George wrote:
> Le nonidi 9 thermidor, an CCXX, compn a écrit :
> > the problem is that Nicolas George made a patch, committed the
> > patch to ffmpeg, and libav has committed it with incorrect attribution.
> 
> Not exactly, but almost. More precisely:
> 
> Michael committed a patch, fixing a problem.
> They committed a very similar patch, with Michael's name on it and the exact
> same commit message; my guess is Anton cherry-picked it.
> But while Michael's patch is correct, their version of it is not; probably
> cherry-pick was confused because vf_overlay has diverged quite a lot.
> 
> I had nothing to do with it except that, in the meantime I proposed and
> pushed a patch similar to their wrong version, then noticed it was wrong and
> proposed a fix.
> 
> Sorry if my initial wording was so confusing as to let think that Michael
> was the one who wrote something wrong.

no offense taken :)

btw, speaking of tasteless
todays libav tastelessness, this time the more common self attribution
of other peoples changes: (the date should clarify beyond doubt who
was the author)

commit 02ac28229aa6824455e2cbc2389355575913041a
Author:     Ronald S. Bultje <rsbultje at gmail.com>
AuthorDate: Wed Jul 25 20:42:04 2012 -0700
Commit:     Ronald S. Bultje <rsbultje at gmail.com>
CommitDate: Thu Jul 26 15:56:32 2012 -0700

    eval: fix printing of NaN in eval fate test.

    This fixes "make fate-eval" on MSVC builds. Without this, the test outputs
    "-1.#NaN" instead of "nan" on MSVS 2010.

diff --git a/libavutil/eval.c b/libavutil/eval.c
index ff3191d..d274272 100644
--- a/libavutil/eval.c
+++ b/libavutil/eval.c
@@ -636,7 +636,10 @@ int main(int argc, char **argv)
         av_expr_parse_and_eval(&d, *expr,
                                const_names, const_values,
                                NULL, NULL, NULL, NULL, NULL, 0, NULL);
-        printf("'%s' -> %f\n\n", *expr, d);
+        if (isnan(d))
+            printf("'%s' -> nan\n\n", *expr);
+        else
+            printf("'%s' -> %f\n\n", *expr, d);
     }

     av_expr_parse_and_eval(&d, "1+(5-2)^(3-1)+1/2+sin(PI)-max(-2.2,-3.1)",
-------------------------------
compare to:

commit 16a565d0cbb7e753f594236c60bbcc66657024f6
Author:     Michael Niedermayer <michaelni at gmx.at>
AuthorDate: Tue Jul 5 10:46:11 2011 +0200
Commit:     Michael Niedermayer <michaelni at gmx.at>
CommitDate: Tue Jul 5 10:46:11 2011 +0200

    eval/fate: try to fix nan printing format failures.

    Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

diff --git a/libavutil/eval.c b/libavutil/eval.c
index 066bf7b..3833ca8 100644
--- a/libavutil/eval.c
+++ b/libavutil/eval.c
@@ -659,7 +659,11 @@ int main(int argc, char **argv)
         av_expr_parse_and_eval(&d, *expr,
                                const_names, const_values,
                                NULL, NULL, NULL, NULL, NULL, 0, NULL);
-        printf("'%s' -> %f\n\n", *expr, d);
+        if(isnan(d)){
+            printf("'%s' -> nan\n\n", *expr);
+        }else{
+            printf("'%s' -> %f\n\n", *expr, d);
+        }
     }

     av_expr_parse_and_eval(&d, "1+(5-2)^(3-1)+1/2+sin(PI)-max(-2.2,-3.1)",

...

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Awnsering whenever a program halts or runs forever is
On a turing machine, in general impossible (turings halting problem).
On any real computer, always possible as a real computer has a finite number
of states N, and will either halt in less than N cycles or never halt.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-cvslog/attachments/20120728/5b59829c/attachment.asc>