| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Noticed a bit late that this started getting releases again,
will be getting notifications in the future.
Will no longer bother with development releases unless this
becomes severely outdated+broken again.
Still need the backtrace patch (removed in 9999) given this
hasn't got the USE_BACKTRACE change.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
| |
Shouldn't really need this warning anymore.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
| |
Signed-off-by: Yuri Konotopov <ykonotopov@gnome.org>
Closes: https://github.com/gentoo/gentoo/pull/37641
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
|
|
|
|
| |
Due to a missing header, likely only fails with some
specific combination of clang+libstdc+musl or so.
Closes: https://bugs.gentoo.org/935352
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Originally thought would only need _GNU_SOURCE, but there is a bit
more to the fix 7zip upstream did in bug #928730. Could still patch
but can go for the easy solution which will use neither cpu_set_t
nor (more importantly) pthread_attr_setaffinity_np that musl lacks.
Also fix a different musl issue only affecting >=pcsx2-1.7.5913,
workaround imported from Chimera Linux.
Closes: https://bugs.gentoo.org/935298
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
| |
Upstream has done away with all submodules and hardcoded the sources
in the repo which does work out for us given removes the ned for the
custom tarball.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
| |
Still applies but been fixed upstream, wasn't neccessary
for the last bump either.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
| |
Hardly tested beside running a game + a valid shader using vulkan,
please report bugs to Gentoo if there is anything up with this that
cannot be reproduced with upstream's AppImage.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
| |
Unlikely to be (both) missing on the average desktop system but warn
in case given upstream ranted about someone not having alsa-utils
on Arch in the linked issue and it could very well happen here too.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
| |
No longer links directly with it, but still using glad for dlopen.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
|
|
|
| |
Bundled glslang is finally gone (traded for system shaderc), so can
remove -fno-sa too.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
| |
cmake does not check for it but it makes some header-only usage.
Could be DEPEND-only but given the logic in the header prefer
to make that rdepend to be safe (that is hardly worth disabling
anyway). Not worth a revbump.
Closes: https://bugs.gentoo.org/926037
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
No need for a revbump given it does not *actually* link with
libGLX but looks for it during configure time.
It could be interesting to force it to be optional (or at least
not last time I tried, but given X libraries currently cannot be
optional either way (and upstream advise against using the wayland
support) it's likely not worth the extra trouble.
Will revisit if ever add a proper USE=X to pcsx2.
Bug: https://bugs.gentoo.org/924765
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Joonas Niilola <juippis@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
| |
Makes what is happening clearer to users, and gives the option to
force-disable it (through /etc/portage/profile) if really wanted.
Albeit doing so will currently fail to build without patches which
upstream is unlikely to accept given they took a clang-only stance
and ignore bugs (thus being forced).
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
| |
If reading this and wondering why this is newly requiring clang, see
the ebuild comment in src_configure (was previously added to 9999).
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
See bug #922149 for what motivated dropping it even though
it's the upstream stable release.
Note that this drops the x86 keyword, upstream does not
support that anymore (only amd64+sse4.1 is supported).
wrt bug #884825, there is no pcsx2 version left using system
soundtouch.
Closes: https://bugs.gentoo.org/884825
Closes: https://bugs.gentoo.org/922149
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1.6.0 is the current "stable" release upstream, but is:
* nearly 4 years old (lacks a *lot* of game fixes)
* 32bit-only (needs multilib on amd64)
* starting to hit issues like bug #922149 (which took a long
time to be reported, indicating it likely has rather few users)
* has some utf8 locale issues last time I tried it
* ebuild is bit dated and needs update for newer wxgtk
* not really getting used/tested by its gentoo maintainers
with no real motivation to fix problems
The 64bit Qt6 1.7.x may still be kind of WIP but has matured a fair
amount and 1.6.0 does not seem to offer much advantages over it with
the exception of working on older machines (x86, or anything lacking
sse4.1).
An upstream "real" release will likely take a long time still, and
the situation for 1.6.0 just get worse meanwhile.
So, opting to stable 1.7.x even though it's "nightly" arbitrary
releases and drop 1.6.0 given it's broken in current state.
Bug: https://bugs.gentoo.org/922149
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
| |
ODR issues seem resolved.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Likely an easy fix that could be done from our end, but no
longer have the patience to support things that upstream is
refusing to support. Albeit may drop the requirement if it
gets fixed by the time of next bump, but otherwise it'll
likely stick around.
Likely not worth using the llvm eclass here, not linking with
and only need to pick the current compiler.
fwiw forcing clang removes the need for the lto patch.
Also BUILD_DATE option was removed. GIT_TAG bits may also need
updates but will check that next time do a bump as it does not
matter in live. LICENSE may need a bit of reviewing too given
upstream added a compilation of all the licenses used.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Upstream likes to bump these to very recent ones at arbitrary
points without seemingly being truly required.
Needed to be able to stabilize pcsx2 without bothering to stable
libsdl2-2.28.4 "right now" when 2.28.3 seems just fine still.
pcsx2-9999 had already needed to switch to 2.28.5 too (why..?).
If a version is found problematic, will add the needed lower
bound to *DEPEND only. Odds are very old Qt6 (e.g. 6.2) is likely
broken but these are long gone and won't worry about it.
Using sed given versions keep changing and want to avoid rebasing
a patch every time.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
| |
Upstream has discovered that unbundling some dependencies reduces
build time.
ARCH_FLAG was removed, it was useful to pass the actual minimal
-m* flags, but -DDISABLE_ADVANCE_SIMD comes close enough and
prevents passing the more problematic -march=native.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As usual pcsx2 is picky about sdl2 versions, not that I can
see the reason why (pretty sure this works just fine with
stable libsdl2-2.28.3 and this just annoys stable users
accept keywording it).
May really have to consider just sedding out the version
check for Qt and SDL2 if these keep up (not a patch given
will likely keep getting changed).
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|
|
|
|
| |
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
|