Created attachment 1634967 [details] build log for failed ppc64le build graph-tool does not currently build on ppc64le. Complete build log attached. Help appreciated.
Looks like it failed due out-of-memory on the builder. "g++: fatal error: Killed signal terminated program cc1plus" is the symptom. Workarounds are to reduce parallelism or debuginfo verbosity. Fedora infra is also reviewing the builder VM setup to have good ration between memory size and number of virtual CPUs.
Make was running with -j1 here, so a serial build. :(
This message is a reminder that Fedora 31 is nearing its end of life. Fedora will stop maintaining and issuing updates for Fedora 31 on 2020-11-24. It is Fedora's policy to close all bug reports from releases that are no longer maintained. At that time this bug will be closed as EOL if it remains open with a Fedora 'version' of '31'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, simply change the 'version' to a later Fedora version. Thank you for reporting this issue and we are sorry that we were not able to fix it before Fedora 31 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Fedora, you are encouraged change the 'version' to a later Fedora version prior this bug is closed as described in the policy above. Although we aim to fix as many bugs as possible during every release's lifetime, sometimes those efforts are overtaken by events. Often a more recent Fedora release includes newer upstream software that fixes bugs or makes them obsolete.
This bug appears to have been reported against 'rawhide' during the Fedora 34 development cycle. Changing version to 34.
This bug appears to have been reported against 'rawhide' during the Fedora 35 development cycle. Changing version to 35.
Updating this bug to reflect current status. I worked around the resource exhaustion following > note: variable tracking size limit exceeded with '-fvar-tracking-assignments', retrying without by reducing the level of debugging symbols on architectures that were hitting resource limits: > %ifarch armv7hl ppc64le s390x > %global optflags %(echo %{optflags} | sed 's/-g /-g1 /') > %endif Now the problem looks like: > /usr/include/boost/math/tools/promotion.hpp:138:10: > error: static assertion failed: > Sorry, but this platform does not have sufficient long double support for > the special functions to be reliably implemented. That looks pretty conclusive, so I’m closing this issue as CANTFIX. (It will still satisfy the ExcludeArch tracking requirement by blocking PPCTracker.)
I believe it's waiting on https://fedoraproject.org/wiki/Changes/PPC64LE_Float128_Transition to be implemented.
(In reply to Dan Horák from comment #7) > I believe it's waiting on > https://fedoraproject.org/wiki/Changes/PPC64LE_Float128_Transition to be > implemented. Thanks! I wasn’t aware of that change proposal. I’m reopening this bug, and blocking it on the change tracker bug, since that is a likely fix. Unfortunately, it looks like the change proposal was deferred, and it’s not clear when or if it will be resubmitted.
and the change was implemented for F-36, please give it a try
Thanks. I’m away from a suitable computer at the moment, but I’ll try it in the next week or two—or perhaps Ankur will have a chance.
I gave it another try[1], and found that the configure script could not find CGAL on ppc64le *only*. I know that CGAL became header-only in 5.4.0, but everything is working fine on x86_64, aarch64, and s390x. I’ll have to try to figure out what is different on PowerPC. checking for __gmpz_init in -lgmp... yes checking for __gmpz_init in -lgmp... (cached) yes checking whether CGAL is available in /usr... checking whether CGAL is available in /usr... no checking whether CGAL is available in /usr/local... checking whether CGAL is available in /usr/local... no checking whether CGAL is available in /opt... checking whether CGAL is available in /opt... no checking whether CGAL is available in /opt/local... checking whether CGAL is available in /opt/local... no configure: error: CGAL library not found. [1] https://koji.fedoraproject.org/koji/taskinfo?taskID=82254634
I guess it's because you can't "include" the CGAL headers on ppc64le, they will throw an error. See the links from bug 2045152 for details.
The most details are in https://github.com/boostorg/multiprecision/issues/419
Thanks! That makes sense. I’ll keep an eye on the upstream bugs (CGAL/Boost/GCC). I’ll also do an emulated ppc64le mock-build and dig through the config.log to verify that the messages therein are in fact consistent with bug 2045152.
Changing the version to Rawhide, as this won’t be fixed in F35.
Created attachment 1858474 [details] Detailed logs from from failing configure script It’s definitely the CGAL/Boost/GCC issue. From config.log: > configure:22319: g++ -o conftest -fopenmp -g -O2 -fexceptions -g1 -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mcpu=power8 -mtune=power8 -fasynchronous-unwind-tables -fstack-clash-protection -I/usr/include -pthread -I/usr/include -Wl,-z,relro -Wl,--as-needed -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -Wl,-dT,/builddir/build/BUILD/graph-tool-2.43/.package_note-python-graph-tool-2.43-11.fc36.ppc64le.ld -L/usr/lib64 conftest.cpp -lgmp -lgmp >&5 > In file included from /usr/lib/gcc/ppc64le-redhat-linux/12/include/immintrin.h:29, > from /usr/include/boost/multiprecision/cpp_int/intel_intrinsics.hpp:68, > from /usr/include/boost/multiprecision/cpp_int/add_unsigned.hpp:10, > from /usr/include/boost/multiprecision/cpp_int/add.hpp:12, > from /usr/include/boost/multiprecision/cpp_int.hpp:2286, > from /usr/include/CGAL/boost_mp.h:33, > from /usr/include/CGAL/Number_types/internal/Exact_type_selector.h:27, > from /usr/include/CGAL/Exact_kernel_selector.h:26, > from /usr/include/CGAL/Filtered_kernel.h:22, > from /usr/include/CGAL/Exact_predicates_inexact_constructions_kernel.h:21, > from conftest.cpp:58: > /usr/lib/gcc/ppc64le-redhat-linux/12/include/mmintrin.h:52:2: error: #error "Please read comment above. Use -DNO_WARN_X86_INTRINSICS to disable this error." > 52 | #error "Please read comment above. Use -DNO_WARN_X86_INTRINSICS to disable this error." > | ^~~~~ > In file included from /usr/lib/gcc/ppc64le-redhat-linux/12/include/immintrin.h:31: > /usr/lib/gcc/ppc64le-redhat-linux/12/include/xmmintrin.h:54:2: error: #error "Please read comment above. Use -DNO_WARN_X86_INTRINSICS to disable this error." > 54 | #error "Please read comment above. Use -DNO_WARN_X86_INTRINSICS to disable this error." > | ^~~~~ > […]
I just merged the PR to re-enable ppc64le[1] and kicked off a build[2]. That should close this bug when it finishes. Thank you, Dan, for helping debug this. [1] https://src.fedoraproject.org/rpms/python-graph-tool/pull-request/4 [2] https://koji.fedoraproject.org/koji/taskinfo?taskID=82371452
This bug appears to have been reported against 'rawhide' during the Fedora 36 development cycle. Changing version to 36.
Not sure why this didn’t auto-close, but it’s fixed now, for F36 and later.