Use a 120 KiB per-function threshold. This roughly matches the default
thread stack size on Musl, which is 128 KiB, leaving some headroom for
caller frames.
This warning draws attention to functions in the codebase that may have
excessively large stack frames. A few functions could benefit from
reducing their frame size.
120 KiB is a reasonable limit. If a function requires more stack, it
likely would benefit from restructuring the code.
Signed-off-by: Kacper Michajłow <kasper93@gmail.com>
I tried it, and it broke horribly. We should definitively rule out this.
Unfortunately, it's not as easy as just setting `disable asm` here, because
asm having been enabled will already have affected the build process further
upstream - conversely, we don't know whether or not memalign exists until
fairly late in the build process, and I'm not about to go destroying
the current organization of the file just to work around this.
So an error message it is.
When an array of pointers to strings is converted into
an array of constant-length strings, one has to find out
the maximum size of the strings. Currently no warning will
be emitted If one forgets to account for the trailing zero,
making this optimization dangerous (think of the scenario
where the array will be modified without adjusting the maximum
size of the strings).
The -Wunterminated-string-initialization warning catches these cases.
It is supported by GCC 15.1 and Clang 21. It only requires to mark
the few instances of arrays initialized via string literals that
are not supposed to be strings with the nonstring attribute; this
has already been done in the previous commit.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Clang by default allows unknown warnings and merely emits
a Wunknown-warning-option warning which is non-fatal by default,
making it appear as if Clang supported any warning. This patch
rewrites the tests to always add the arguments to error out
in case an unrecognized option is encountered.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
E.g. when CC is clang, the unknown warnings flags
are -Werror=unused-command-line-argument and
-Werror=unknown-warning-option. These flags are currently
also used for testing the C++ and ObjC compilers, yet
g++ does not recognize these flags and errors out because
of them, so that the tests fail and the parentheses,
switch etc. warnings are not disabled for C++.
Fix this by using per-language flags.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Remove the improperly named check functions that were added for the
opencolorio detection in 677cf95ea4.
Those are not needed anymore and there are already require_cxx and
check_lib_cxx for this purpose.
When checking for C++ classes in check_class_headers_cxx, the test code
required the class to be default-constructible, which is not always the
case. For the sake of the test, it is enough to check that referencing
the class succeeds.
This reverts commit aa20d7b3e8.
Adding these avpriv functions is absolutely overblown: Muxers
can get the desired duration in a few lines themselves.
In particular, using the parse functions from this file
necessitated parsing the extradata (and entailed exporting
the parsing function), although it was only used to know
whether the frames are self-delimiting, but everything of
interest to a muxer does not depend on this at all.
The commit to be reverted also made several structures
part of the ABI, which should be avoided in general.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Implement ff_ssl_*_key_cert()
Generate self-signed cert and key in server
mode if there're no key and cert input.
Implement ff_tls_set_external_socket() and
ff_dtls_export_materials()
Add gnutls as dtls protocol deps.
Signed-off-by: Jack Lau <jacklau1222gm@gmail.com>
This patch adds screen capture support via AMF vsrc_amf:
ffmpeg -y \
-filter_complex "vsrc_amf=framerate=120:capture_mode=keep_framerate" \
-c:v hevc_amf \
-frames:v 300 \
output_grab.mp4
If the HW frames pool is insufficient,
increase extra_hw_frames in filter_complex.
Removed in f2affdfafb by mistake.
Also, fix the logic in probe_glslc() so it doesn't return false when debug is disabled.
Signed-off-by: James Almer <jamrial@gmail.com>
Name the feature "arm_crc" rather than plain "crc", to make it
clear that this is about a CPU feature extension, not CRC
implementations in general.
This requires dealing with the extension slightly differently
than other extensions, as the name of the feature and the
".arch_extension" extension name differ.
Naming it with an "arm" prefix rather than "aarch64", as the
CPU extension also is available in 32 bit ARM form, even though
we don't intend to use it there.
