GitLab

Fixes #13

Invoke `make check` at the end of the default build, since we
don't do `make distcheck` for all the jobs. At the moment this
does nothing, but it should give better coverage if any test
harnesses are hooked up to the target.

We only really need distcheck to succeed for the default autotools
build, so save time by only running that step once.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>
Signed-off-by: Marvin Scholz <epirat07@gmail.com>

Do additional builds to test the various configure switches.
This gives us better parity with the travis ci builds.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>
Signed-off-by: Marvin Scholz <epirat07@gmail.com>

Do a basic autoconf build and check to verify commits
in gitlab reposistories.
Signed-off-by: Tristan Matthews <tmatth@videolan.org>
Signed-off-by: Marvin Scholz <epirat07@gmail.com>

Fixes ubsan error in fuzzer:
"runtime error: signed integer overflow: 9223372036854775807 - -1 cannot be represented in type 'long'"

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/speex

Fixes ubsan error in fuzzer:
   "runtime error: signed integer overflow: 2147483616 + 160 cannot be represented in type 'int'

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/speex

Fixes ubsan error:
"runtime error: signed integer overflow: 51200 - -9223372036854767360 cannot be represented in type 'long'"

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/speex

Signed-off-by: Tristan Matthews <tmatth@videolan.org>

Fixes ubsan error:
"runtime error: negation of -2147483648 cannot be represented in type 'int'; cast to an unsigned type to negate this value to itself"

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/speex

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/speex

Co-Authored by Tyson Smith <twsmith@mozilla.com>

This avoids an integer overflow.

Backported from opus commit 1ee139bca076a1a3606a8a924a698d356d40cbb8

This should yield output comparable to the wavs at https://www.speex.org/samples/

Fixes hang discovered by fuzzing: https://github.com/xiph/speex/issues/9

Fixes #2037

Signed-off-by: Tristan Matthews <tmatth@videolan.org>

_WIN32 is always defined by the compiler automatically when targeting
that platform, while WIN32 only is defined automatically in some
configurations, and e.g. in MSVC only ever is defined in project files
(if at all).

Some other checks in the codebase already check for both WIN32 and
_WIN32; those are left untouched.

This fixes warnings in calling code about "this function declaration
is not a prototype".

Based on the config from the opus project.

Reported by Roger Pack.

And we can drop the mkdir here now too, since we actually include things
of our own in that directory now so it does already exist.

Checking for upper or lower case RIFF fourcc makes sense, checking
twice for the same one does not.