On 01/08/15 20:32, Anton Khirnov wrote: > Quoting Luca Barbato (2015-08-01 14:16:45) >> On 31/07/15 07:59, Anton Khirnov wrote: >>> Quoting Luca Barbato (2015-07-29 22:51:05) >>>> And restrict the string to ascii text. >>>> >>>> CC: libav-sta...@libav.org >>>> >>>> Signed-off-by: Luca Barbato <lu_z...@gentoo.org> >>>> --- >>>> libavcodec/h264_sei.c | 57 >>>> ++++++++++++++++++++++++++++++++++++++++++--------- >>>> 1 file changed, 47 insertions(+), 10 deletions(-) >>>> >>>> diff --git a/libavcodec/h264_sei.c b/libavcodec/h264_sei.c >>>> index 1f3844b..7361941 100644 >>>> --- a/libavcodec/h264_sei.c >>>> +++ b/libavcodec/h264_sei.c >>>> @@ -191,24 +191,61 @@ static int decode_registered_user_data(H264Context >>>> *h, int size) >>>> return 0; >>>> } >>>> >>>> +static const uint8_t x264_version_uuid[] = { >>>> + 0xdc, 0x45, 0xe9, 0xbd, 0xe6, 0xd9, 0x48, 0xb7, >>>> + 0x96, 0x2c, 0xd8, 0x20, 0xd9, 0x23, 0xee, 0xef >>>> +}; >>>> + >>>> +static int decode_x264_version(H264Context *h, int size) >>>> +{ >>>> + int build = 0, i = 0; >>>> + uint8_t x264_string[256]; >>>> + >>>> + if (size < 13) >>>> + goto skip; >>>> + >>>> + for (i = 0; i < size && i < sizeof(x264_string) - 1; i++) >>>> + x264_string[i] = get_bits(&h->gb, 8) & 0x7f; >>> >>> I don't think this is enough, there are still control characters in this >>> range (or the caller can have a weird character set). Either filter out >>> everything before ' ', or use isprint() perhaps? >>> >>>> + >>>> + x264_string[i] = 0; >>>> + >>>> + if (!memcmp(x264_string, "x264 - core ", 12)) >>>> + build = atoi(x264_string + 12); >>>> + >>>> + if (build > 0) >>>> + h->x264_build = build; >>>> + >>>> + if (h->avctx->debug & FF_DEBUG_BUGS) >>>> + av_log(h->avctx, AV_LOG_DEBUG, "x264 version string:\"%s\"\n", >>>> x264_string); >>>> +skip: >>>> + for (; i < size; i++) >>>> + skip_bits(&h->gb, 8); >>> >>> Why duplicate this? >>> >> >> Duplicate what exactly? > > This exact same loop is also at the end of > decode_unregistered_user_data() >
I can rework to return the consumed size from both probably. _______________________________________________ libav-devel mailing list libav-devel@libav.org https://lists.libav.org/mailman/listinfo/libav-devel