On Fri, Dec 5, 2014 at 8:21 AM, Christophe Gisquet <[email protected]> wrote: > HuiHi, > > 2014-12-04 21:00 GMT+01:00 Nicolas Dufresne <[email protected] >>: >> av_fast_malloc() signature is void*, but this is just a convenience >> not to have to cast. In fact it expect a pointer to pointer. This >> fixes a crash on NULL pointer when playing back H263 Sorensen. >> >> Signed-off-by: Nicolas Dufresne <[email protected]> >> --- >> libavcodec/svq1dec.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/libavcodec/svq1dec.c b/libavcodec/svq1dec.c >> index 14ff41c..13bbdd1 100644 >> --- a/libavcodec/svq1dec.c >> +++ b/libavcodec/svq1dec.c >> @@ -637,7 +637,7 @@ static int svq1_decode_frame(AVCodecContext *avctx, > void *data, >> return AVERROR_INVALIDDATA; >> } >> >> - av_fast_malloc(s->pkt_swapped, &s->pkt_swapped_allocated, >> + av_fast_malloc(&s->pkt_swapped, &s->pkt_swapped_allocated, >> buf_size); >> if (!s->pkt_swapped) >> return AVERROR(ENOMEM); > > Isn't that (same patch context, different line numbers?) a subset of > "[PATCH] svq1dec: Add padding, clear size, use the correct pointer"? If > yes, I'd say the other is better.
The only difference is that the other patch zeroes the pointer on deinit, which is not really required. I prefer the Nicolas' patch as it offers more context and a wider commit message. -- Vittorio _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
