On 12/22/2012 07:58 PM, Luca Barbato wrote: > On 12/22/2012 09:55 PM, Justin Ruggles wrote: >> On 12/10/2012 11:49 AM, Justin Ruggles wrote: >>> From: Michael Niedermayer <[email protected]> >>> >>> This avoids a potential infinite loop due to not consuming any packet data. >>> >>> Signed-off-by: Michael Niedermayer <[email protected]> >>> Signed-off-by: Justin Ruggles <[email protected]> >>> >>> CC:[email protected] >>> --- >>> libavcodec/wmadec.c | 5 +++-- >>> 1 files changed, 3 insertions(+), 2 deletions(-) >>> > > The open questions are: > > - which values we accept for block_align? ( <=0 are impossible? )
No, they're certainly not impossible. > - where should we validate it? I think the question is, are there files with wma which do not have block_align set in the demuxer but will decode properly without it? I don't know if there are any. The samples provided so far seem to be very strange and will not decode properly. The fix just avoids an infinite loop at the user level due to the decoder consuming 0 bytes. I would be ok at this point with either trying to decode as if it's 1 frame-per-packet or just returning an error. Either way...we need to fix it. -Justin _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
