Created attachment 800899 [details] buildlog sys-libs/binutils-libs-2.38-r2: finishes the strip stage with x86_64-pc-linux-gnu-strip: /var/tmp/portage/sys-libs/binutils-libs-2.38-r2/image/usr/lib64/stIEIWxU/regex.o: plugin needed to handle lto object can be solved by adding -ffat-lto-objects to the compiler flags
*** Bug 866416 has been marked as a duplicate of this bug. ***
*** Bug 866419 has been marked as a duplicate of this bug. ***
*** Bug 866500 has been marked as a duplicate of this bug. ***
"plugin needed to handle lto object" message should now be completely harmless. Workaround (calling ranlib on archive) was added to Portage over 3 years ago for bug #603594: https://gitweb.gentoo.org/proj/portage.git/commit/?id=2404ddca9d5db7992bf6853cbde8ca944224560c See also: https://sourceware.org/bugzilla/show_bug.cgi?id=21479 https://github.com/InBetweenNames/gentooLTO/issues/49
I suppose we can close this then?
Let's keep this open in case someone runs into it again.
Hey, just wanted to open a new bug for this issue. It occurs to any package built with LTO with USE="static-libs". Are you sure it's harmless?
(In reply to zurabid2016 from comment #7) > Hey, just wanted to open a new bug for this issue. It occurs to any package > built with LTO with USE="static-libs". Are you sure it's harmless? I've had kde-frameworks/kwayland fail to build with -fno-lto if dev-qt/qtgui:5 wasn't built with -ffat-lto-objects, but that's only happened when using the mold linker.
H.J. Lu is kindly working on this. Testing his patch.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/proj/toolchain/binutils-patches.git/commit/?id=dbbd30a6cf0ba45b316f11a01f25428230fe2768 commit dbbd30a6cf0ba45b316f11a01f25428230fe2768 Author: Sam James <sam@gentoo.org> AuthorDate: 2025-05-04 10:14:41 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2025-05-04 10:14:41 +0000 9999: add H.J.'s patch for strip --plugin Bug: https://sourceware.org/PR21479 Bug: https://bugs.gentoo.org/866422 Bug: https://bugs.gentoo.org/926120 Signed-off-by: Sam James <sam@gentoo.org> 9999/0006-strip-lto-plugin.patch | 751 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 751 insertions(+)
(In reply to Kostadin Shishmanov from comment #8) > (In reply to zurabid2016 from comment #7) > > Hey, just wanted to open a new bug for this issue. It occurs to any package > > built with LTO with USE="static-libs". Are you sure it's harmless? > > I've had kde-frameworks/kwayland fail to build with -fno-lto if > dev-qt/qtgui:5 wasn't built with -ffat-lto-objects, but that's only happened > when using the mold linker. That should happen either way if -ffat-lto-objects wasn't used, as you can't opt-out then (nothing to fallback on). This bug is still a real issue if calling strip manually though and the ranlib hack never felt like something to rely on. It's also noise that scares people ;) Anyway, bug 926120 handles all of that. I think we're nearly there on it. Just testing the eclass more...
(In reply to Sam James from comment #11) > That should happen either way if -ffat-lto-objects wasn't used, as you can't > opt-out then (nothing to fallback on). This bug is still a real issue if > calling strip manually though and the ranlib hack never felt like something > to rely on. > It also makes a few things start to make more sense with the eclass and behave more consistently (all of my tests pass now even when using archives, not just raw objects).
commit 70a5c8b6e74d26c9dd3571950a08761a72751d21 Author: Andreas K. Hüttel <dilfridge@gentoo.org> Date: Thu May 22 23:43:11 2025 +0200 sys-devel/binutils: add 2.44-r2 (patchlevel 4), no keywords Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org> commit 2166943fbf9a78333b5ec4f31017e79e68806be0 Author: Andreas K. Hüttel <dilfridge@gentoo.org> Date: Fri May 23 01:21:07 2025 +0200 sys-devel/binutils: keyword 2.44-r2 Signed-off-by: Andreas K. Hüttel <dilfridge@gentoo.org>
