Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP

Loading…

gap is broken in subtle ways #50

Closed
kiwifb opened this Issue · 6 comments

2 participants

@kiwifb
Collaborator

A doctest introduced by a ticket merged in sage-4.6.2.alpha2 reveals that gap is broken:

sage -t -force_lib "devel/sage/sage/groups/perm_gps/permgroup.py"
**********************************************************************
File "/usr/share/sage/devel/sage/sage/groups/perm_gps/permgroup.py", line 1876:
    sage: G.isomorphism_to(H)
Expected:
    Permutation group morphism:
      From: Permutation Group with generators [(2,3), (1,2,3)]
      To:   Permutation Group with generators [(1,2,4), (1,4)]
      Defn: [(2,3), (1,2,3)] -> [(2,4), (1,2,4)]
Got:
    Permutation group morphism:
      From: Permutation Group with generators [(2,3), (1,2,3)]
      To:   Permutation Group with generators [(1,2,4), (1,4)]
      Defn: [(2,3), (1,2,3)] -> [(2,4), (1,4,2)]
**********************************************************************
1 items had failures:
   1 of   8 in __main__.example_62
***Test Failed*** 1 failures.

A vanilla build pass the test. Replacing gap from vanilla build with the gentoo
one leads to the failure. Applied all visible patch from sage without improvement.
Investigating.

@kiwifb
Collaborator

gap seems to be fragile. Re-merging gap may cause sage to hang and leave rogue python or gap processes when interrupted (only done so with the above test).

I somehow manage to break the sage-gap interaction completely by installing gap in a different location. Sage will spawn a gap process as usual but the test will complain that gap cannot be started.

@kiwifb
Collaborator

The gap stamp thing may be a source of problem. I sometimes have two gap processes started that are slightly different. I am guessing they are a source of trouble.

@cschwan
Owner

Check it again - I think the problem was fixed with Sage 4.7!

@kiwifb
Collaborator

For a value of fixed. What happened is http://trac.sagemath.org/sage_trac/ticket/10334 got pulled out of 4.7 before release because of some other problems. It is merged in 4.7.1.alpha1 so it reappears there.
The conclusion we had with Steve was that the results are probably equivalent but we would have to ask someone who knows what is done here.

@kiwifb
Collaborator

This could very well be related to this new ticket: http://trac.sagemath.org/sage_trac/ticket/11427

@kiwifb
Collaborator
@kiwifb kiwifb closed this
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.