Bug 654822 – dev-lang/python-2.7.14-r2 with sys-devel/gcc-8.1.0 - src_compile(): segmentation fault ./python -E -S -m sysconfig --generate-posix-vars generate-posix-vars failed

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Steffen Hau 2018-05-04 11:04:34 UTC]

Upstream patch [1] fixes the compilation. Strangely this does only occur on my Haswell system, my other two Skylake systems do not show the segfault.

[1] https://github.com/python/cpython/commit/0b91f8a668201fc58fa732b8acc496caedfdbae0

[Comment 1 Perfect Gentleman 2018-05-09 03:13:26 UTC]

confirm it on Haswell system. there is no this bug on AMD Phenom_II and Turion_II

[Comment 2 Perfect Gentleman 2018-05-11 15:12:04 UTC]

2.7.15 builds fine

[Comment 3 Vladimir Gavrilov 2018-05-11 17:34:45 UTC]

Also confirm this bug on Haskell.

[Comment 4 Althorion 2018-05-11 20:36:56 UTC]

Breaks down on Threadripper too.

[Comment 5 Matt Whitlock 2018-06-17 13:01:32 UTC]

Segfault in generate-posix-vars also occurs on Kentsfield (Intel Core 2 Quad Q6600) using GCC 8.1.0.

I can work around it by setting:

CC=x86_64-pc-linux-gnu-gcc-7.3.0
CXX=x86_64-pc-linux-gnu-g++-7.3.0

in /etc/portage/env/dev-lang/python-2.7.14.


Yes, I am still using a processor that's more than 11 years old.

[Comment 6 Pacho Ramos 2018-07-10 14:07:31 UTC]

Fedora ended up using https://mail.python.org/pipermail/python-dev/2018-January/152011.html I think

[Comment 7 Alex Xu (Hello71) 2018-07-10 14:25:30 UTC]

(In reply to Pacho Ramos from comment #6)
> Fedora ended up using
> https://mail.python.org/pipermail/python-dev/2018-January/152011.html I think

I'm pretty sure that was what landed in 2.7.15.

[Comment 8 Manfred Knick 2019-01-28 10:13:51 UTC]

(In reply to Matt Whitlock from comment #5)
> ... Intel Core 2 ...
  Still love my rock-solid ThinkPad T60 from 2007   :-)

@  Python Gentoo Team:
In-between, dev-lang/python-2.7.15:2.7 has been stabilized on all arches.
So this bug could be considered as 'RESOLVED FIXED' ?