Debian u1 patches

[infinity0]

This ticket lists some Debian patches that we're applying to Sage. They should be fairly non-controversial.

• Use system libGAP paths #22437
• docbuild: Use lexists when testing whether a symlink exists #22438
• Fix use of scipy rtol parameter #22439
• Work around Python's inadequate SIGPIPE handling #22440
• Don't fail matplotlib related doctests if DISPLAY is unset #22441
• Compile pynac interface with SINGULAR_CFLAGS to find its headers #22442
• Do not test the version of glpk #22443
• ipy install: Fix very incomplete logic in symlink() #22444
• doctest: Make sage.doctest.sources.get_basename work in more situations #22445

(We have other patches which need some more work and discussion; I'll be submitting those later as well.)

These were taken from our 7.5.1 branch, which is a WIP. They applied cleanly to 7.6, apart from a trivial merge conflict fix in the case of "Don't fail plot.py". However I haven't tested it on a non-Debian install yet (I don't have the build resources for that right this minute).

Component: misc

Author: Ximin Luo, Tobias Hansen, Jerome Benoit

Issue created by migration from https://trac.sagemath.org/ticket/22409

[infinity0]

Branch: u/infinity0/debian_u1_patches

[infinity0]

New commits:

084b452	Use system paths for libGAP
d1ce385	Use lexists instead of exists
317aef6	Fix Sage's use of scipy rtol parameter
b50511f	Work around python bug #1652
29962ec	Don't fail plot.py doctest if DISPLAY is unset
9dd9258	Compile pynac interface with SINGULAR_CFLAGS to find its headers
e9baf21	Do not test the version of glpk

[infinity0]

Commit: e9baf21

[sagetrac-git]

Changed commit from e9baf21 to 27d41fa

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

bd439fb	Use system paths for libGAP
b91b5a7	Use lexists instead of exists
49b5625	Fix Sage's use of scipy rtol parameter
9bbceda	Work around python bug #1652
6ac7288	Don't fail plot.py doctest if DISPLAY is unset
68152df	Compile pynac interface with SINGULAR_CFLAGS to find its headers
27d41fa	Do not test the version of glpk

[infinity0]

comment:4

Re-did the commits because the older ones leaked my local host name. Whoops...

[infinity0]

Upstream: None of the above - read trac for reasoning.

[infinity0]

Author: Ximin Luo, Tobias Hansen, Jerome Benoit

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

316c6c3	Fix atrocious logic in symlink()
399f911	Make sage.doctest.sources.get_basename work in more situations

[sagetrac-git]

Changed commit from 27d41fa to 399f911

[infinity0]

Changed upstream from None of the above - read trac for reasoning. to none

[jdemeyer]

comment:8

I would much prefer to have one ticket for each issue that you are fixing instead of grouping together a bunch of unrelated patches.

[infinity0]

comment:9

Some of these commits are very small and opening a new ticket just for a single-line commit seems like unnecessary overhead. The libgap and scipy ones are a bit larger. Shall I still split everything or would you be happy with me splitting out only these two?

[jdemeyer]

comment:10

One ticket per issue. If that means 10 tickets, then so be it.

[infinity0]

comment:11

Reopened as #22437, #22438, #22439, #22440, #22441, #22442, #22443, #22444, #22445.

You can close this ticket. I tried but it doesn't look like I have the permissions.

[videlec]

comment:12

Better to keep this ticket as a task ticket... changed accordingly

[videlec]

Changed commit from 399f911 to none

[videlec]

Changed branch from u/infinity0/debian_u1_patches to none

[videlec]

comment:13

(and thanks for the effort of proposing the patches to Sage!)

[infinity0]

comment:14

welcome! I've fixed the missing branches, I think it is something to do with the fact that trac logs me out every 10 minutes or so (because I connect through Tor).. will keep an eye out on that in the future