tests fail on armhf

Bug #1882228 reported by Dimitri John Ledkov
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
regina-normal (Ubuntu)
Fix Released
Undecided
Logan Rosen

Bug Description

tests fail on armhf

https://github.com/regina-normal/regina/issues/58

LinkTest::jones... ok.
LinkTest::homfly...

50% tests passed, 1 tests failed out of 2

Total Test time (real) = 97.84 sec

The following tests FAILED:
          2 - cpp-test (SEGFAULT)

Tags: ftbfs armhf
Revision history for this message
Ben Burton (bab) wrote :

This is fixed in regina-normal 5.96, which I uploaded to debian/sid a few days ago.

Note that 5.96 is a real upstream release, and is both newer and more polished than the 6.0~alpha1 package currently in ubuntu (which was pulled from a branch of the upstream git repository).

I would recommend, instead of trying to patch this issue on an unsupported git pull, to simply migrate the real 5.96 release from Debian in a way that does not break versioning. I would suggest using an "upstream" version of 6.0~pr5.96 or something similar (which is what I've done in my upstream unofficial ubuntu packages). It would be great to not bump the epoch here, since this would (I presume) permanently break the automatic debian-to-ubuntu migration even once a 6.0 release comes out.

- Ben.

Revision history for this message
Ben Burton (bab) wrote :

Oh: also I just noticed that the ubuntu packagers have temporarily disabled the tests on armhf.

Please re-enable these tests.

It turns out this was a real mathematical bug that was picked up by the test suite, which was doing exactly what it was designed to do. The bug was not picked up in debian because it never appeared in a real upstream release, but it *did* appear in ubuntu because ubuntu is currently shipping an unofficial git pull.

I do appreciate that the test failure was reported upstream, but nevertheless the test suite is meant to be a first line of defence against bugs. I have never released a version of regina that failed any of its tests on any of the many platforms and architectures that I build it for, and this is precisely so that users can be confident in the reliability of this mathematical software.

If you are shipping an unsupported git pull that fails one or more tests, I would strongly request (as the lead upstream developer) that ubuntu should not ship a knowingly broken git pull at all, and should instead wait for an official release (as opposed to just disabling the tests, which of course does not actually fix the underlying problem at all).

Thanks :) - Ben.

Revision history for this message
Ben Burton (bab) wrote :

An update: I have now uploaded regina-normal 6.0-1 to debian sid (and it has built on armhf).

The correct fix for this should now be to just migrate the 6.0-1 package into hirsute.

The ubuntu-specific patches should (I believe) be able to go also, since my understanding is that they were for the python3 migration, but I have now done the python3 migration in the debian packages also.

Thanks - Ben.

Revision history for this message
Logan Rosen (logan) wrote :

This bug was fixed in the package regina-normal - 6.0-1

---------------
regina-normal (6.0-1) unstable; urgency=low

  * New upstream release.
  * Bumped standards-version to 4.5.1 (no changes required).

 -- Ben Burton <email address hidden> Sat, 9 Jan 2021 12:12:06 +1000

Changed in regina-normal (Ubuntu):
assignee: nobody → Logan Rosen (logan)
status: New → In Progress
status: In Progress → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.