The project appears to have gone source-unavailable and all-in on AI.
There are 6.x versions available mysteriously as some git branch with no
distfiles (Fedora packages this) and a 7.x version is announced on their
website but you have to contact sales to find out more. They haven't
communicated on github in years.
Not only is reporting a bug an exercise in futility, I'm not super
interested in trying to begin with.
Closes: https://bugs.gentoo.org/863737
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
Problems with firefox that led to masking 1.1.14 seem
solved with recent firefox+1.1.15, but that doesn't
seem to be the end of issues, let's keep it masked for
a while still.
Alternatively could try __NV_DISABLE_EXPLICIT_SYNC=1
but it is currently unclear if this really solves all
issues.
Bug: https://bugs.gentoo.org/937773
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
This is obvious in hindsight. The flags we were passing *precisely because*
we wanted e.g. -Wno-format to be passed to every test also included
-O2 or whatever from the user's flags which breaks tests that may require
no optimisation and so on.
Instead of trying to filter out options in a whack-a-mole game, let's
introduce special GCC_TESTS_* *FLAGS variables which we append
the needed -Wno-* to (etc.) which users can also specify if they really want
or need to.
Note that this isn't as scary or as weird as it sounds. We were only
trying to jam these flags in **purely** to counteract some defaults
we set, these tests really aren't supposed to be run with arbitrary
flags stuck in, but a workaround we added started to introduce way
more than intended.
The torture tests are fine with being run with various optimisation flags
but the rest of the testesuite isn't.
This fixes 361c375864 and some of the
attempts afterwards, although not everything in those commits was bad.
Fixes: 361c375864
Signed-off-by: Sam James <sam@gentoo.org>
Make clear it's not a typo as it comes from gcc's testsuite (could
look like an accidental 'TEST' instead of 'TESTS' like the eclassvars
we use for tests.)
Signed-off-by: Sam James <sam@gentoo.org>
Using git:// is a really bad idea and the eclass yells at you for it.
And the second uri does not actually exist (anymore).
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
Using git:// is a really bad idea and the eclass yells at you for it.
And the second uri does not actually exist (anymore).
Note: the live ebuild still doesn't actually build. The patch is stale
and then there are a bunch of weird "XXX does not name a type" and
"YYY was not declared in this scope" errors that I have no patience to
debug. Perhaps an upstream bug. I was able to determine that existing
LTO errors are still present and then gave up. Baby steps... :D
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
Zac Medico
Eli Schwartz
Matt Jolly
Michał Górny
Ionen Wolkens
Sam James
Jason Zaman
Filip Kobierski