rpms
/
python-graph-tool
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Re-enable s390x (fix RHBZ#1771034) 

Work around resource limits to re-enable s390x, and get closer to the
root causes on the remaining excluded architectures
This commit is contained in:
Benjamin A. Beasley 2021-08-27 16:16:09 -04:00
parent b20f1242c2
commit 1a9573c615
1 changed files with 55 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

63
python-graph-tool.spec
View File

@ -42,24 +42,67 @@ documentation.}
Name: python-%{modname}
Version: 2.33
Release: 1%{?dist}
Release: 2%{?dist}
Summary: Efficient network analysis tool written in Python
License: LGPLv3+
URL: https://graph-tool.skewed.de/
Source0: https://downloads.skewed.de/%{modname}/%{modname}-%{version}.tar.bz2
# Fails on i686: issue filed: https://git.skewed.de/count0/graph-tool/issues/617
# https://bugzilla.redhat.com/show_bug.cgi?id=1771023
# Fails on armv7hl: virtual memory exhausted
# Need to reduce debugging symbols to make compiling possible on these
# architectures at all. See:
# https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/message/GIBNORLNGSUHAMX6PZ7XUPQ2SQUK7AX4/
#
# Works around:
# Fails on armv7hl:
# virtual memory exhausted
# https://bugzilla.redhat.com/show_bug.cgi?id=1771024
# Fails on ppc64le: note: variable tracking size limit exceeded with '-fvar-tracking-assignments', retrying without
# Fails on ppc64le:
# note: variable tracking size limit exceeded with
# '-fvar-tracking-assignments', retrying without
# https://bugzilla.redhat.com/show_bug.cgi?id=1771031
# Fails on s390x: note: variable tracking size limit exceeded with '-fvar-tracking-assignments', retrying without
# Fails on s390x:
# note: variable tracking size limit exceeded with
# '-fvar-tracking-assignments', retrying without
# https://bugzilla.redhat.com/show_bug.cgi?id=1771034
# Takes ~23 hours on x86_64 if we get unlucky and get a 6 core 16gig machine, ~4 hours if we get a 48 core 128gig machine
%ifarch armv7hl ppc64le s390x
%global optflags %(echo %{optflags} | sed 's/-g /-g1 /')
%endif
# Fails on i686, armv7hl
# ../../../src/pcg-cpp/include/pcg_random.hpp:1247:40: error: call to
# non-'constexpr' function 'pcg_extras::uint_x4<U, V>
# pcg_extras::operator-(const pcg_extras::uint_x4<U, V>&, const
# pcg_extras::uint_x4<U, V>&) [with UInt = unsigned int; UIntX2 = long
# long unsigned int]'
# 1247 | (state_type(1U) << table_pow2) - state_type(1U);
# | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~
# In file included from ../../../src/pcg-cpp/include/pcg_extras.hpp:84,
# from ../../../src/pcg-cpp/include/pcg_random.hpp:114,
# from ../../../src/graph/random.hh:21,
# from graph_motifs.hh:26,
# from graph_motifs.cc:24:
# ../../../src/pcg-cpp/include/pcg_uint128.hpp:642:22: note:
# 'pcg_extras::uint_x4<U, V> pcg_extras::operator-(const
# pcg_extras::uint_x4<U, V>&, const pcg_extras::uint_x4<U, V>&) [with
# UInt = unsigned int; UIntX2 = long long unsigned int]' declared here
# 642 | uint_x4<UInt,UIntX2> operator-(const uint_x4<UInt,UIntX2>& a,
# | ^~~~~~~~
# issue filed: https://git.skewed.de/count0/graph-tool/issues/617
# https://bugzilla.redhat.com/show_bug.cgi?id=1771023
# https://bugzilla.redhat.com/show_bug.cgi?id=1771024
#
# Fails on ppc64le:
# /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.
# https://bugzilla.redhat.com/show_bug.cgi?id=1771031
#
# Takes ~23 hours on x86_64 if we get unlucky and get a 6 core 16gig machine,
# ~4 hours if we get a 48 core 128gig machine
# Takes ~45 hours on aarch64
ExcludeArch: i686 armv7hl ppc64le s390x
ExcludeArch: %{ix86} armv7hl ppc64le
# Workaround for https://git.skewed.de/count0/graph-tool/issues/613
# CGAL 5.x is now header only
@ -133,6 +176,10 @@ find $RPM_BUILD_ROOT -name "*.la" -delete
%{_libdir}/pkgconfig/%{modname}-py%{python3_version}.pc
%changelog
* Sat Aug 28 2021 Benjamin A. Beasley <code@musicinmybrain.net> - 2.33-2
- Work around resource limits to re-enable s390x, and get closer to the root
causes on the remaining excluded architectures
* Fri Sep 04 2020 Ankur Sinha <ankursinha AT fedoraproject DOT org> - 2.33-1
- Update to latest release
- Disable LTO