[FFmpeg-devel] [PATCH] Allow mpjpeg demuxer to process MIME parts which do not include Content-Length header.
Nicolas George
george at nsup.org
Mon Nov 30 08:41:42 CET 2015
Le nonidi 9 frimaire, an CCXXIV, Alexander Agranovsky a écrit :
> Please see the updated patches attached. The trimming loop that was subject
> of the discussion had been rewritten to use indices rather than pointer
> arithmetics.
This kind of drastic change was not necessary, you can do the same with
pointers. IMHO, the best way of dealing with that situation is this: when
dealing with the end of the string, have the pointer point AFTER the end of
the string, i.e.:
char *p = string + strlen(string); // not -1
if (p > string && av_isspace(p[-1]))
*(--p) = 0;
> + char* boundary;
Here and in all new code, please use "char *var" instead of "char* var" for
consistency. There is a good reason for that: "char* a, b" means that a is a
pointer and b is not, grouping the pointer mark with the type is misleading.
> + "Expected boundary '%s' not found, instead found a line of %lu bytes\n",
> + expected_boundary,
> + strlen(line));
"%lu" is not correct for size_t. The correct type would be %zu, but it is
possible that we have to use another construct to avoid bugs from microsoft
libraries, see other instances in the code for examples.
> - size = parse_content_length(value);
> - if (size < 0)
> - return size;
> + *size = parse_content_length(value);
Did you remove the check on purpose?
> + if (!av_strncasecmp(start, "boundary=", 9)) {
> + start += 9;
It has already be pointed out: av_stristart() to avoid the duplicated magic
number.
Can not comment on the functional aspect, sorry.
Regards,
--
Nicolas George
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20151130/a0dfff7a/attachment.sig>
More information about the ffmpeg-devel
mailing list