Isn't it time to drop manylinux2014 (PEP 599)? It is based on EOLed CentOS 7, and we don't support Python 3.7 and early 3.8/9, anyway, so what's the point?
<https://github.com/pypa/manylinux> As far as gp2c specifically, we can if needed add flags into its spkg-install (and upgrade, still) Dima On August 7, 2025 11:38:34 AM CDT, Marc Culler <marc.cul...@gmail.com> wrote: >I will check that, but I suspect it is a compiler issue. > >I was able to get it to build by adding -fcommon to CFLAGS. Apparently gcc >10, was the first version of gcc to make -fno-common be the default, and >later versions may handle it in a way which allows gp2c to declare external >variables in a header file which is included multiple times and has no >guard. Who knows, maybe >version 0.0.14 uses guards against multiple includes. > >But I am seeing other issues that look like they are compiler issues. The >meataxe spkg also did not build, and a fix was to delete the word inline >from a function definition in window.c. It seems that gcc 10 does not >always create a symbol for an inline function, so an inline function >defined in a library is not accessible to modules which link against the >library. > >I guess gcc 10 may be too old for sage. But it does work if no optional >packages are enabled. > >- Marc > >On Thursday, August 7, 2025 at 10:48:09 AM UTC-5 Dima Pasechnik wrote: > >> On Wed, Aug 6, 2025 at 4:41 PM Marc Culler <marc....@gmail.com> wrote: >> > >> > I am unable to build gp2c on a manylinux2014 docker image with this >> release. There are 3210 errors of the form "multiple definition of X" >> starting with: >> > >> > [spkg-install] >> /opt/rh/devtoolset-10/root/usr/libexec/gcc/x86_64-redhat-linux/10/ld: >> context.o:(.bss+0x0): multiple definition of `debug_stack'; >> gp2c.o:(.bss+0x0): first defined here >> > [spkg-install] >> /opt/rh/devtoolset-10/root/usr/libexec/gcc/x86_64-redhat-linux/10/ld: >> context.o:(.bss+0x8): multiple definition of `typemax'; gp2c.o:(.bss+0x8): >> first defined here >> >> please check whether the upgrade to gp2c 0.0.14 helps: >> https://github.com/sagemath/sage/pull/40552 >> (the upgrade that should have been done as soon as we switched to pari >> 2.17) >> >> Dima >> >> > >> > - Marc >> > >> > >> > On Tuesday, August 5, 2025 at 6:17:49 PM UTC-5 Volker Braun wrote: >> >> >> >> As always, you can get the latest beta version from the "develop" git >> branch. Alternatively, the self-contained source tarball is at >> http://www.sagemath.org/download-latest.html >> >> >> >> 4a2dac04955 (tag: 10.7.rc1, github/develop) Updated SageMath version to >> 10.7.rc1 >> >> bcdffce78b7 gh-40519: typo + guava >> >> 94fe540d332 (tag: 10.7.rc0) Updated SageMath version to 10.7.rc0 >> > >> > -- >> > You received this message because you are subscribed to the Google >> Groups "sage-release" group. >> > To unsubscribe from this group and stop receiving emails from it, send >> an email to sage-release...@googlegroups.com. >> > To view this discussion visit >> https://groups.google.com/d/msgid/sage-release/e0957cda-c67b-46f6-92af-dd891ba0ae40n%40googlegroups.com >> . >> > >-- >You received this message because you are subscribed to the Google Groups >"sage-release" group. >To unsubscribe from this group and stop receiving emails from it, send an >email to sage-release+unsubscr...@googlegroups.com. >To view this discussion visit >https://groups.google.com/d/msgid/sage-release/bb299e93-086e-4b7e-abd1-230bde30210cn%40googlegroups.com. -- You received this message because you are subscribed to the Google Groups "sage-release" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-release+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/sage-release/A8EA6BF4-6B10-4292-9B92-0EEFAA905DF4%40gmail.com.
