seems like this flag only exists in llvm 17: https://github.com/llvm/llvm-project/commit/a4d3bc68142e67c1cbceef3ac33cdaf9b78fd84a. AFAICT it's not specific to my configuration, just build chromium-115.0.5790.24[lto] with <llvm-16.
*** Bug 908491 has been marked as a duplicate of this bug. ***
I don't think that'll work..? Building this with lto with llvm<16 that is Or do you know it works?
Dropping the flag is harmless - we did it in 114 for now. But I don't know if 115 has other LLVM 17 needed stuff.
I was able to build and run www-client/chromium-115.0.5790.24 with these settings: * Package: www-client/chromium-115.0.5790.24:0/beta * Repository: gentoo * Maintainer: chromium@gentoo.org Matt.Jolly@footclan.ninja * USE: abi_x86_64 amd64 cups custom-cflags elibc_glibc hangouts kernel_linux libcxx lto official pic proprietary-codecs pulseaudio screencast system-harfbuzz system-png wayland * FEATURES: compressdebug network-sandbox preserve-libs sandbox splitdebug userpriv usersandbox * FEATURES='assume-digests binpkg-dostrip binpkg-multi-instance buildpkg buildpkg-live candy compressdebug config-protect-if-modified distlocks ebuild-locks fixlafiles ipc-sandbox merge-sync multilib-strict network-sandbox news noinfo parallel-fetch parallel-install pid-sandbox preserve-libs protect-owned qa-unresolved-soname-deps sandbox sfperms split-log splitdebug strict unknown-features-warn unmerge-orphans userfetch userpriv usersandbox usersync xattr' * CFLAGS='-O2 -flto=thin -march=native -pipe -fno-math-errno' * CXXFLAGS='-O2 -flto=thin -march=native -pipe -fno-math-errno' * FFLAGS='-O2 -flto=thin -march=native -pipe -fno-math-errno' * FCFLAGS='-O2 -flto=thin -march=native -pipe -fno-math-errno' * F77FLAGS='-O2 -flto=thin -march=native -pipe -fno-math-errno' * LDFLAGS='-fuse-ld=lld -Wl,-O2 -Wl,--as-needed -Wl,-z,now -Wl,-z,pack-relative-relocs -Wl,--undefined-version -O2 -flto=thin -march=native -pipe -fno-math-errno' * MAKEOPTS='-j10 -l13' * USE_NONGNU='1' * clang version 16.0.5 plus EXTRA_GN="is_cfi=true" and patching out -disable-auto-upgrade-debug-info.
But wouldn't the correct thing to do for 115 be to depend on LLVM >=17? Since Chromium >=115 seems to depend on it? I'm currently in the process of building the LLVM 17 prerelease that is available in Portage to try building Chromium with it.
(In reply to Mads from comment #5) > But wouldn't the correct thing to do for 115 be to depend on LLVM >=17? > Since Chromium >=115 seems to depend on it? > > I'm currently in the process of building the LLVM 17 prerelease that is > available in Portage to try building Chromium with it. 1. Upstream already disable it for ChromeOS until LLVM 17 is there 2. As you noted, LLVM 17 is not released, so we wouldn't want to keyword it, so we'd have unkeyworded new Chromium 3. For 114 at least (and hopefully 115), it really is a harmless single option which it's fine to remove.
Ah, good to know. I guess you maybe should increase LLVM_MAX_SLOT to 17 in 115 and up so the chromium build will use 17 if you have it...
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=d6bbecf1257852a8129762ad4f0d2bbef16eafc0 commit d6bbecf1257852a8129762ad4f0d2bbef16eafc0 Author: Stephan Hartmann <sultan@gentoo.org> AuthorDate: 2023-06-15 20:25:44 +0000 Commit: Stephan Hartmann <sultan@gentoo.org> CommitDate: 2023-06-15 20:25:55 +0000 www-client/chromium: beta channel bump to 115.0.5790.32 Closes: https://bugs.gentoo.org/908432 Closes: https://bugs.gentoo.org/908467 Signed-off-by: Stephan Hartmann <sultan@gentoo.org> www-client/chromium/Manifest | 4 ++-- .../{chromium-115.0.5790.24.ebuild => chromium-115.0.5790.32.ebuild} | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-)
