Summary: | sys-cluster/galera-25.3.10 - src_configure(): Checking for C library pthread... no // Error: pthread library not found | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Jeroen Roovers (RETIRED) <jer> |
Component: | [OLD] Server | Assignee: | Gentoo Linux MySQL bugs team <mysql-bugs> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | hppa |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
See Also: | https://bugs.gentoo.org/show_bug.cgi?id=543114 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 544378 | ||
Attachments: |
sys-cluster:galera-25.3.10:20150328-072608.log.xz
config.log sys-cluster:galera-25.3.10:20150408-041135.log |
Description
Jeroen Roovers (RETIRED)
2015-04-03 04:45:52 UTC
--- ./ChangeLog +++ ./ChangeLog @@ -4,0 +5,3 @@ + 03 Apr 2015; Brian Evans <grknight@gentoo.org> galera-25.3.5.ebuild: + Add append-flags -D_REENTRANT wrt bug 545376 + (In reply to Brian Evans from comment #1) > --- ./ChangeLog > +++ ./ChangeLog > @@ -4,0 +5,3 @@ > + 03 Apr 2015; Brian Evans <grknight@gentoo.org> galera-25.3.5.ebuild: > + Add append-flags -D_REENTRANT wrt bug 545376 1) I'm not sure that's the right fix. 2) 25.3.10 still fails. Please post the config.log that is generated in $S Sconstruct says: if not conf.CheckLib('pthread'): print 'Error: pthread library not found' Exit(1) On amd64: scons: Configure: Checking for C library pthread... .sconf_temp/conftest_0.c <- | | | |int |main() { | |return 0; |} | x86_64-pc-linux-gnu-gcc -o .sconf_temp/conftest_0.o -c -O2 -march=native -pipe -fomit-frame-pointer -mno-tls-direct-seg-refs -std=c99 -fno-strict-aliasing -pipe -DNDEBUG -m64 -Wall -Wextra -Wno-unused-parameter -pthread -D_XOPEN_SOURCE=600 -DHAVE_COMMON_H -I. -Iasio -Icommon -Igalerautils/src -Igcomm/src -Igcomm/src/gcomm -Igcache/src -Igcs/src -Iwsdb/src -Igalera/src .sconf_temp/conftest_0.c x86_64-pc-linux-gnu-gcc -o .sconf_temp/conftest_0 -m64 -Wl,-melf_x86_64 -Wl,-O1 -Wl,--as-needed .sconf_temp/conftest_0.o -lpthread scons: Configure: yes Created attachment 400744 [details]
config.log
Added a patch to remove all hardware specific CFLAGS. Please retest again. Created attachment 400802 [details]
sys-cluster:galera-25.3.10:20150408-041135.log
./chromium/build_config.h:129:2: error: #error Please add support for your architecture in build/build_config.h
That file lays out support for amd64, x86, ARM, MIPS and PPC. It doesn't look like it's trying to figure out what architecture it's being compiled for very well - it just looks for some keywords and sets some variables. In the case of __PPC64__, for instance, it magically knows the system must be bigendian. It shouldn't be hard to teach it to do this properly, I guess. But it's still scons.
(In reply to Jeroen Roovers from comment #6) > Created attachment 400802 [details] > sys-cluster:galera-25.3.10:20150408-041135.log > > ./chromium/build_config.h:129:2: error: #error Please add support for your > architecture in build/build_config.h > > That file lays out support for amd64, x86, ARM, MIPS and PPC. It doesn't > look like it's trying to figure out what architecture it's being compiled > for very well - it just looks for some keywords and sets some variables. In > the case of __PPC64__, for instance, it magically knows the system must be > bigendian. It shouldn't be hard to teach it to do this properly, I guess. > But it's still scons. The sad part is.. they borrowed this code from chromium and the only part that is used outside that file is chromium/aligned_memory.h:#if defined(__GNUC__) && !defined(ARCH_CPU_X86_FAMILY) The only define that is used is COMPILER_GCC. So I can patch out that error and have no ill effects. Then again, #if defined(__GNUC__) && !defined(ARCH_CPU_X86_FAMILY) CHROMIUM_COMPILE_ASSERT(ALIGNOF(T) <= 16, crbug_115612); #endif will probably trigger an error since CHROMIUM_COMPILE_ASSERT does not seem to be defined anywhere :( (In reply to Brian Evans from comment #1) please revert that as the bug lies in gcc (see bug 543114) (In reply to SpanKY from comment #8) > (In reply to Brian Evans from comment #1) > > please revert that as the bug lies in gcc (see bug 543114) This part is reverted.. still need to investigate the other definition. I've added galera-25.3.12 and patched out the #error preprocessor since that section is only used as: #if defined(__GNUC__) && !defined(ARCH_CPU_X86_FAMILY) CHROMIUM_COMPILE_ASSERT(ALIGNOF(T) <= 16, crbug_115612); #endif I found the assert and it does a simple check. If that fails, then there will be no support. Please retry No response for test request and I believe all problems are resolved. Reopen if there are still issues |