Opened 5 years ago
Closed 5 years ago
#24947 closed defect (fixed)
meataxe doctest failure on 8.2.beta8
Reported by:  Vincent Delecroix  Owned by:  

Priority:  blocker  Milestone:  sage8.2 
Component:  linear algebra  Keywords:  
Cc:  Jeroen Demeyer, Simon King  Merged in:  
Authors:  Jeroen Demeyer  Reviewers:  Simon King 
Report Upstream:  N/A  Work issues:  
Branch:  447b4e5 (Commits, GitHub, GitLab)  Commit:  447b4e567170715e234a13d5d4695c21292e034e 
Dependencies:  Stopgaps: 
Description
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)
Change History (13)
comment:1 Changed 5 years ago by
Summary:  meataxe failing on 8.2.beta8 → meataxe doctest failure on 8.2.beta8 

comment:2 followup: 3 Changed 5 years ago by
comment:3 followup: 8 Changed 5 years ago by
Replying to SimonKing:
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).
comment:4 followup: 5 Changed 5 years ago by
comment:5 Changed 5 years ago by
comment:6 Changed 5 years ago by
Authors:  → Jeroen Demeyer 

Component:  packages: optional → linear algebra 
comment:7 Changed 5 years ago by
Branch:  → u/jdemeyer/meataxe_doctest_failure_on_8_2_beta8 

comment:8 Changed 5 years ago by
Commit:  → 447b4e567170715e234a13d5d4695c21292e034e 

Replying to vdelecroix:
Replying to SimonKing:
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__

comment:9 Changed 5 years ago by
Status:  new → needs_review 

comment:10 Changed 5 years ago by
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.
comment:12 Changed 5 years ago by
Reviewers:  → Simon King 

Status:  needs_review → positive_review 
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.
comment:13 Changed 5 years ago by
Branch:  u/jdemeyer/meataxe_doctest_failure_on_8_2_beta8 → 447b4e567170715e234a13d5d4695c21292e034e 

Resolution:  → fixed 
Status:  positive_review → closed 
I guess that's the change in matrix initialisation introduced by Jeroen. But hasn't that been tested with optional matrix backends?