There are three types of style entries which are redundant:
a) Entries with length zero. They are already discarded.
b) Entries that are equivalent to the default style:
They can be safely discarded.
c) Entries that are equivalent to the immediately preceding style
if the start of the current style coincides with the end of the
preceding style. In this case the styles can be merged.
This commit implements discarding/merging in cases b) and c).
This fixes ticket #9548. In said ticket each packet contained
exactly one style entry that covered the complete packet with
the exception of the last character (probably created by a tool
that didn't know that the style's end is exclusive). Said style
coincided with the default style, leading to a superfluous reset,
which is now gone.
Reviewed-by: Philip Langdale <philipl@overt.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Both TextSampleEntry and TextSample can contain StyleRecords;
yet both the code as well as the structures for them were duplicated.
This commit changes this.
Reviewed-by: Philip Langdale <philipl@overt.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Giving elements of a structure called StyleBox names like
"style_start" or "style_end" is redundant, especially given
that the relevant variables are also called style.
Reviewed-by: Philip Langdale <philipl@overt.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Don't use different src and dst in av_tea_crypt(); use in-place
modifications instead. Also let av_tea_crypt() encrypt all three
blocks in one call.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
We use ECB, not CBC mode here, so one does not need to reinitialize
the context; for the same reason, one can also just let av_tea_crypt()
loop over the blocks, avoiding a loop here.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Up until now, the packets have been read in blocks of at most
eight bytes at a time; then these blocks have been decrypted
and copied into a buffer on the stack (that was double the size
needed...). From there they have been copied to the dst packet.
This commit changes this: The data is read in one go; and
the decryption avoids temporary buffers, too, by making
use of the fact that src and dst of av_tea_crypt() can coincide.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Due to this bush.aa (from the FATE suite) exported garbage metadata
with key "_040930".
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
av_image_copy() expects an array of four pointers according to its
declaration; although it currently only touches pointers that
are actually in use (depending upon the pixel format) this might
change at any time (as has already happened for the linesizes
in d7bc52bf45).
This fixes a -Wstringop-overflow= warning with GCC 11.2.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The concatf protocol returns an opaque error on open if
concatf list file contains trailing newlines.
Signed-off-by: Gyan Doshi <ffmpeg@gyani.pro>
Reviewed-by: James Almer <jamrial@gmail.com>
If the input stream framerate is known, it will be configured on the
relevant filtergraph input and get propagated to the output stream in
the above line. That makes these assignments redundant.
Instead of storing the protocol pointer in the opaque iteration state,
store just the index of the next protocol, similarly to how
ff_urlcontext_child_class_iterate() works.
Silences e.g. the following warning in gcc 10:
src/libavformat/ftp.c: In function ‘ftp_move’:
src/libavformat/ftp.c:1122:46: warning: ‘%s’ directive output may be truncated writing up to 4095 bytes into a region of size 4091 [-Wformat-truncation=]
1122 | snprintf(command, sizeof(command), "RNTO %s\r\n", path);
| ^~ ~~~~
src/libavformat/ftp.c:1122:5: note: ‘snprintf’ output between 8 and 4103 bytes into a destination of size 4096
1122 | snprintf(command, sizeof(command), "RNTO %s\r\n", path);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Silences the following warning with gcc 10:
src/libavdevice/v4l2.c: In function ‘v4l2_get_device_list’:
src/libavdevice/v4l2.c:1042:64: warning: ‘%s’ directive output may be truncated writing up to 255 bytes into a region of size 251 [-Wformat-truncation=]
1042 | ret = snprintf(device_name, sizeof(device_name), "/dev/%s", entry->d_name);
| ^~
src/libavdevice/v4l2.c:1042:15: note: ‘snprintf’ output between 6 and 261 bytes into a destination of size 256
1042 | ret = snprintf(device_name, sizeof(device_name), "/dev/%s", entry->d_name);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Previous patches intending to silence it have proposed increasing the
buffer size, but doing that correctly seems to be tricky. Failing on
truncation is simpler and just as effective (as excessively long device
names are unlikely).
device and cap are local to the loop iteration, there is no need for
them to retain their values. Especially for device it may be dangerous,
since it points to av_malloc'ed data.
The FD opened here is local to the loop iteration, there is no reason to
store it in the context. Since read_header() may have already been
called, this may ovewrite an existing valid FD.
Maximum output size with a 32-bit int is 17 bytes, or 26 with a 64-bit
int.
Silences the following gcc 10 warning:
src/libavdevice/jack.c: In function ‘audio_read_header’:
src/libavdevice/jack.c:171:45: warning: ‘snprintf’ output may be truncated before the last format character [-Wformat-truncation=]
171 | snprintf(str, sizeof(str), "input_%d", i + 1);
| ^
src/libavdevice/jack.c:171:9: note: ‘snprintf’ output between 8 and 17 bytes into a destination of size 16
171 | snprintf(str, sizeof(str), "input_%d", i + 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Do this by switching from the dynamic buffer API to the AVBPrint API;
the former has no defined way to check for errors.
This also avoids allocating an AVIOContext.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Do this by switching from the dynamic buffer API to the AVBPrint API;
the former has no defined way to check for errors.
This also avoids allocating an AVIOContext.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
av_image_copy() expects an array of four pointers and linesizes
according to its declaration; it currently only pointers that are
actually in use (depending upon the pixel format), but this might
change at any time. It has already happened for the linesizes in
d7bc52bf45 and so increasing their
array fixes a stack-buffer overread.
This fixes a -Wstringop-overflow= and -Wstringop-overread warning
from GCC 11.2.
Reviewed-by: Linjie Fu <linjie.justin.fu@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This reduces codesize because the offsets of commonly used elements
are now smaller and thus need less bytes to encode in ptr+offset
addressing modes (with GCC 11.2 on x64: 0x1b8b -> 0x1a7b).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Currently, it also tests whether extended_data points to something
different than the AVFrame's data array and frees extended_data
if it is different. Yet this is only necessary for one of its three
callers, namely av_frame_unref(); meanwhile the other two callers
took measures to avoid this (or rather, to make it to an av_free(NULL)).
This commit moves this chunk to av_frame_unref() (so that
get_frame_defaults() now treats its input as uninitialized)
and removes the now superfluous code in the other two callers.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The only caller of do_video_out() doesn't need the frame afterwards,
ergo one can replace an av_frame_ref() by av_frame_move_ref().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
