https://sourceware.org/bugzilla/show_bug.cgi?id=24873
Alan Modra changed:
What|Removed |Added
CC||amodra at gmail dot com
Summa
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #3 from dilyan.palauzov at aegee dot org ---
Yes, LTO is involved, but the compiler does not get an explict -flto, so one of
the linked libraries, compiled in the past and not contained in the tarball of
libinput, uses LTO and the
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #4 from Alan Modra ---
> For me is surprizing that the order within --start-group…--end-group does
> matter.
Yes, that is a bug. In suggesting that just having the second -lm present
might work, I was merely trying to characteri
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #5 from Alan Modra ---
I tried to recreate the problem locally on my x86_64 Ubuntu 18.04.2 box using
current libinput git source, but
meson --prefix=/usr -Ddocumentation=false builddir/
results in a failure:
Running test binary co
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #6 from Alan Modra ---
OK, so I fixed the build dependency but when I build locally using CFLAGS="-O2
-g -flto" CXXFLAGS="-O2 -g -flto" and gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0
I don't reproduce the problem. Oh yes, lots more
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #7 from dilyan.palauzov at aegee dot org ---
With meson 0.48 and below, the configure phase terminates with
Meson encountered an error in file meson.build, line 672, column 1:
Native dependency 'check' not found
With meson 0.49 th
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #8 from dilyan.palauzov at aegee dot org ---
Why doesn’t LD -Y/--TRACE-SYMBOL provide useful information about the location
of the symbol. Transient LTO files as result of -Y are not useful information.
--
You are receiving this
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
Alan Modra changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #9 from Alan Modra ---
> At https://mail.aegee.org/dpa/ld24873/
Thanks, and nicely included your libcheck.a, which contains the LTO objects
triggering this bug. After building myself a new gcc-8 (debugging LTO linker
problems req
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #10 from Alan Modra ---
OK, so the LTO recompiled object is being inserted before libcheck.a, which is
fine, inside --start-group/--end-group. However, reloading of archives (and
checking --as-needed DSOs again) is disabled until
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #12 from cvs-commit at gcc dot gnu.org ---
The master branch has been updated by Alan Modra :
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=128bf1fe608badb59d27f9c5c8ffb1a6a6d9d811
commit 128bf1fe608badb59d27f9c5c8ff
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #11 from cvs-commit at gcc dot gnu.org ---
The master branch has been updated by Alan Modra :
https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;h=b02c4f162335ed055964f603ba3aca18b0e5c5bf
commit b02c4f162335ed055964f603ba3a
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
Alan Modra changed:
What|Removed |Added
Status|NEW |RESOLVED
CC|amodra at gm
https://sourceware.org/bugzilla/show_bug.cgi?id=24873
--- Comment #14 from dilyan.palauzov at aegee dot org ---
I have removed https://mail.aegee.org/dpa/ld24873/ .
--
You are receiving this mail because:
You are on the CC list for the bug.
___
bug-b
14 matches
Mail list logo