meataxe doctest failure on 8.2.beta8

[videlec]

Mainly because the constructor of matrices changed

      File "sage/matrix/matrix_gfpn_dense.pyx", line 304, in sage.matrix.matrix_gfpn_dense.Matrix_gfpn_dense.__init__ (build/cythonized/sage/matrix/matrix_gfpn_dense.c:4524)
        def __init__(self, parent, entries=None, *, bint copy=False, bint coerce=False, bint mutable=True):
    TypeError: __init__() takes at most 2 positional arguments (4 given)

See quasar short log

CC: @jdemeyer @simon-king-jena

Component: linear algebra

Author: Jeroen Demeyer

Branch/Commit: 447b4e5

Reviewer: Simon King

Issue created by migration from https://trac.sagemath.org/ticket/24947

[simon-king-jena]

comment:2

I guess that's the change in matrix initialisation introduced by Jeroen. But hasn't that been tested with optional matrix backends?

[videlec]

comment:3

Replying to @simon-king-jena:

I guess that's the change in matrix initialisation introduced by Jeroen. But hasn't that been tested with optional matrix backends?

I just did :-) One problem is that on the previous beta, there was a constant failure on patchbots (#24918). Hence the patchbot quasar (running with all optional packages) was off. Until everybody agrees that each beta must pass with all optional packages you will have to wait (see how much do we support optional packages and #23832).

[jdemeyer]

comment:4

Sorry for that...

This is actually fixed in #24742. Should I create a separate fix for just this issue or does anybody plan to review #24742?

[videlec]

comment:5

Replying to @jdemeyer:

Sorry for that...

This is actually fixed in #24742. Should I create a separate fix for just this issue or does anybody plan to review #24742?

Better to have a fix in an independent ticket (this one). #24742 is likely to take more time.

[jdemeyer]

Author: Jeroen Demeyer

[jdemeyer]

Branch: u/jdemeyer/meataxe_doctest_failure_on_8_2_beta8

[videlec]

comment:8

Replying to @videlec:

Replying to @simon-king-jena:

I guess that's the change in matrix initialisation introduced by Jeroen. But hasn't that been tested with optional matrix backends?

I just did :-) One problem is that on the previous beta, there was a constant failure on patchbots (#24918). Hence the patchbot quasar (running with all optional packages) was off. Until everybody agrees that each beta must pass with all optional packages you will have to wait (see how much do we support optional packages and #23832).

Actually #24918 is not merged in the new beta so that the patchbots are still useless.

---

New commits:

447b4e5

Properly fix signature of Matrix_gfpn_dense.__init__

[videlec]

Commit: 447b4e5

[jdemeyer]

comment:10

Interestingly, this failure is caused by the combination of two merged tickets, each of which individually passed the testsuite. This is just to say that even a perfect patchbot cannot catch everything.

[jdemeyer]

comment:11

ping

[simon-king-jena]

comment:12

If I understand correctly, you added your commit on 11th of March, but the patchbot gets errors on 12th of March.

However, both tests failures that I see are in test(interacts.statistics.coin), which apparently is unrelated with this ticket. And the commit looks good to me.

[simon-king-jena]

Reviewer: Simon King

[vbraun]

Changed branch from u/jdemeyer/meataxe_doctest_failure_on_8_2_beta8 to 447b4e5