libffi: update to 3.5.2 (let's try this again!)#59704
Conversation
amadaluzia
force-pushed
the
libffi
branch
4 times, most recently
from
87917ae to
43e4800
Compare
amadaluzia
force-pushed
the
libffi
branch
6 times, most recently
from
9b0dd8c to
1ab9910
Compare
amadaluzia
force-pushed
the
libffi
branch
2 times, most recently
from
aa9109d to
c15de40
Compare
amadaluzia
force-pushed
the
libffi
branch
4 times, most recently
from
f1476d2 to
8f62e76
Compare
amadaluzia
force-pushed
the
libffi
branch
2 times, most recently
from
56617fa to
8c5cf1f
Compare
| version=3.5.2 | ||
| revision=1 | ||
| build_style=gnu-configure | ||
| configure_args="--includedir=/usr/include --disable-multi-os-directory --with-pic" |
There was a problem hiding this comment.
--includedir defaults to $prefix/include so it shouldn't be necessary
--without-gcc-arch to prevent march and mtune auto-detection and instead build for generic target
There was a problem hiding this comment.
Probably shouldn't be marked as a "Git LFS" file
There was a problem hiding this comment.
I have no idea why or who made it an LFS file but I'm hoping it had a reason.
There was a problem hiding this comment.
Something needs to be done to libcxx, libcxxabi, libunwind and their -devel — those symlinks point to llvm22 instead, so while build just fine, the installation phase fails.
For testing purposes I temporarily swapped these to llvm21 and after building it, back to llvm22.
With that, both llvm21 and llvm22 packaged just fine:
- x86_64
- x86_64-musl
- i686
- armv6l *
- armv6l-musl *
- armv7l *
- armv7l-musl *
- aarch64 *
- aarch64-musl *
Using * to mark cross
There was a problem hiding this comment.
Separate it out
Might make sense to update, 115.36.0 built fine.
Or remove the package altogether — the only template depending on it is 0ad, and upstream moved to mozjs128 instead now.
There was a problem hiding this comment.
Separated, just haven't updated it yet. It's worth it's own pull request.
There was a problem hiding this comment.
Commit message doesn't fit
Currently mozjs128 fails while running testsuite.
Can be solved by updating, i.e. to v128.14.0 and dropping 019-python-3.11.patch which has been merged by upstream.
With that, built fine.
There was a problem hiding this comment.
Fixed the commit message, the update should be it's own PR.
There was a problem hiding this comment.
Wrong commit message
Should probably also update, i.e. 140.11.0
Required these to build:
https://github.com/mozilla-firefox/firefox/commit/4a8e9c0b1b6924f734c90486790a7c011faeda84.patch
https://github.com/mozilla-firefox/firefox/commit/dde4db6dea5c1504db58eee58b05b849600a0c5b.patch
https://github.com/mozilla-firefox/firefox/commit/b09b27487e5d649612e0baae309608954ff9d542.patch
https://github.com/mozilla-firefox/firefox/commit/d497aa4f770ca02f6083e93b94996a8fe32c2ff4.patch
https://github.com/mozilla-firefox/firefox/commit/e6db063cdccce8ace1fc7f7bc881175f9d10b255.patch
Probably fine to combine these together into "python-3.14.patch"
There was a problem hiding this comment.
Corrected the commit message, this should also be it's own PR.
There was a problem hiding this comment.
FYI: p11-kit fails in 2 tests when running with XBPS_CHROOT_CMD=uunshare, but builds and runs the test suite without failure with XBPS_CHROOT_CMD=uchroot
This is observable on master too.
2 participants
Testing the changes
[ci skip]
Local build testing
Additional notes
I have currently only built libffi. I have yet to build the other bumped packages. I'm currently writing a script to build them all, although I doubt it'll be finished even after the afternoon (it's 5:10 at the moment, and I have not slept yet).