While updating webkit-gtk it just failed with the following: FAILED: /usr/bin/x86_64-pc-linux-gnu-g++ -DBUILDING_GTK__=1 -DBUILDING_WITH_CMAKE=1 -DBUILDING_WTF -DDATA_DIR=\"share\" -DGETTEXT_PACKAGE=\"WebKit2GTK-4.0\" -DHAVE_CONFIG_H=1 -DUSER_AGENT_GTK_MAJOR_VERSION=602 -DUSER_AGENT_GTK_MINOR_VERSION=1 -DWEBKITGTK_API_VERSION_STRING=\"4.0\" -DENABLE_JIT=0 -DENABLE_YARR_JIT=0 -DENABLE_ASSEMBLER=0 -DNDEBUG -march=native -O2 -pipe -fno-strict-aliasing -std=c++11 -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/bmalloc -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/dtoa -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/text -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/threads -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/unicode -I/var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/ThirdParty -I. -IDerivedSources -isystem /usr/include/glib-2.0 -isystem /usr/lib64/glib-2.0/include -Wall -Wextra -Wcast-align -Wformat-security -Wmissing-format-attribute -Wpointer-arith -Wundef -Wwrite-strings -fPIC -MMD -MT Source/WTF/wtf/CMakeFiles/WTF.dir/ParkingLot.cpp.o -MF Source/WTF/wtf/CMakeFiles/WTF.dir/ParkingLot.cpp.o.d -o Source/WTF/wtf/CMakeFiles/WTF.dir/ParkingLot.cpp.o -c /var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/ParkingLot.cpp In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/atomic:41:0, from /var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/Atomics.h:29, from /var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/ParkingLot.h:31, from /var/tmp/portage/net-libs/webkit-gtk-2.10.7/work/webkitgtk-2.10.7/Source/WTF/wtf/ParkingLot.cpp:27: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/bits/atomic_base.h: In member function ‘void WTF::WordLock::lock()’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/bits/atomic_base.h:538:70: error: failure memory model cannot be stronger than success memory model for ‘__atomic_compare_exchange’ return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 1, __m1, __m2); ^ /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/bits/atomic_base.h: In member function ‘void WTF::WordLock::unlock()’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/bits/atomic_base.h:538:70: error: failure memory model cannot be stronger than success memory model for ‘__atomic_compare_exchange’ return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 1, __m1, __m2); ^ ninja: build stopped: subcommand failed. Reproducible: Always Actual Results: failed while building Expected Results: builds fine
Created attachment 426592 [details] Output from emerge --info =net-libs/webkit-gtk-2.10.7 I can confirm this bug.
Created attachment 426814 [details] Output from emerge --info "=net-libs/webkit-gtk-2.10.7" I can confirm as well this bug.
Does anyone know of a work-around for this issue? It seems that this issue prevents gnome/stable updates from finishing and masking this version fails because dependencies are not met.
I cannot reproduce this Also, I see you both are using -Os CFLAG... can you try only with -O2? (also, I see that in one case you are specifying -O2 and -Os at the same time... that doesn't make much sense to me, -Os is like -O2 without the options that will increase the size)
(In reply to Pacho Ramos from comment #4) > I cannot reproduce this > > Also, I see you both are using -Os CFLAG... can you try only with -O2? > (also, I see that in one case you are specifying -O2 and -Os at the same > time... that doesn't make much sense to me, -Os is like -O2 without the > options that will increase the size) ... a bit late responding to this: Yeah, for me changing C(XX)FLAGS to "-O2" works for me. I sometimes forget I have "-Os" (this is for a virtual machine build) and so I forget to first try "-O2" and it's been a while since "-Os" presented any problems for me. webkit-gtk finally build with that C(XX)FLAGS change and I'll just put that in /etc/portage/env.
Adding -O2 fixed the problem for me, as well. Thanks.
I reported this upstream at https://bugs.webkit.org/show_bug.cgi?id=155371 Moving to -O2 is a workaround but not a solution (as setting this affects all packages).
(In reply to Johannes Buchner from comment #7) > I reported this upstream at https://bugs.webkit.org/show_bug.cgi?id=155371 > > Moving to -O2 is a workaround but not a solution (as setting this affects > all packages). I wasn't able to compile webkit-gtk-2.10.7 with -Os with gcc 4.9.3, but when I upgraded gcc to 5.3.0 it compiled with -Os.
I can confirm that this build fail is still present with net-libs/webkit-gtk-2.10.9 versus gcc-4.9.3 (both latest "stable" for amd64 at time of writing). Changing -Os to -O2 in CXXFLAGS seems to be working around it for me as well.
*** Bug 589366 has been marked as a duplicate of this bug. ***
Upstream produces now a patch.
What is the status of this with 2.14.x? It seems upstream considers it solved... but per some reports in upstream bug report... maybe it's still happening in some situations :/
I confirm ---cut---- ParkingLot.cpp In file included from /usr/lib/gcc/i686-pc-linux-gnu/4.9.4/include/g++-v4/atomic:41:0, from /var/tmp/portage/net-libs/webkit-gtk-2.14.5/work/webkitgtk-2.14.5/Source/WTF/wtf/Atomics.h:29, from /var/tmp/portage/net-libs/webkit-gtk-2.14.5/work/webkitgtk-2.14.5/Source/WTF/wtf/ParkingLot.h:31, from /var/tmp/portage/net-libs/webkit-gtk-2.14.5/work/webkitgtk-2.14.5/Source/WTF/wtf/ParkingLot.cpp:27: /usr/lib/gcc/i686-pc-linux-gnu/4.9.4/include/g++-v4/bits/atomic_base.h: In member function 'void WTF::WordLockBase::lock()': /usr/lib/gcc/i686-pc-linux-gnu/4.9.4/include/g++-v4/bits/atomic_base.h:538:70: error: failure memory model cannot be stronger than success memory model for '__atomic_compare_exchange' return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 1, __m1, __m2); ^ /usr/lib/gcc/i686-pc-linux-gnu/4.9.4/include/g++-v4/bits/atomic_base.h: In member function 'void WTF::WordLockBase::unlock()': /usr/lib/gcc/i686-pc-linux-gnu/4.9.4/include/g++-v4/bits/atomic_base.h:538:70: error: failure memory model cannot be stronger than success memory model for '__atomic_compare_exchange' return __atomic_compare_exchange_n(&_M_i, &__i1, __i2, 1, __m1, __m2); ^ ninja: build stopped: subcommand failed. * ERROR: net-libs/webkit-gtk-2.14.5::gentoo failed (compile phase): * (no error message) * * Call stack: * ebuild.sh, line 115: Called src_compile * environment, line 5425: Called cmake-utils_src_compile * environment, line 1491: Called _cmake_execute_optionally 'src_compile' * environment, line 532: Called enable_cmake-utils_src_compile * environment, line 2075: Called cmake-utils_src_make * environment, line 1506: Called _cmake_ninja_src_make * environment, line 588: Called die * The specific snippet of code: * "$@" || die * * If you need support, post the output of `emerge --info '=net-libs/webkit-gtk-2.14.5::gentoo'`, * the complete build log and the output of `emerge -pqv '=net-libs/webkit-gtk-2.14.5::gentoo'`. * The complete build log is located at '/var/log/portage/net-libs:webkit-gtk-2.14.5:20170407-222247.log'. * The ebuild environment file is located at '/var/tmp/portage/net-libs/webkit-gtk-2.14.5/temp/environment'. * Working directory: '/var/tmp/portage/net-libs/webkit-gtk-2.14.5/work/webkit-gtk-2.14.5_build' * S: '/var/tmp/portage/net-libs/webkit-gtk-2.14.5/work/webkitgtk-2.14.5'
Created attachment 469448 [details] emerge --info net-libs/webkit-gtk
Is this still an issue with webkit-gtk-2.18.6 (or newer if looking later)?
webkit-gtk-2.18.6 compiles now fine in my system. Most likely because the currently stable gcc version is now 6.4.0, so this bug shouldn't be hitting stable users anymore.
Lets call this obsolete then (especially with icu + 17.0 profiles). Glad it's working now.
