Catch and fix the segmentation fault in dlx_solver

[jdemeyer]

On a 64-bit Gentoo Linux system running sage-4.7.2.alpha2:

sage: from sage.combinat.matrices.dancing_links import dlx_solver
sage: rows = []
sage: x = dlx_solver(rows)
sage: x.search()
/usr/local/src/sage-4.7.2.alpha2/local/lib/libcsage.so(print_backtrace+0x31)[0x7f6c37015e02]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libcsage.so(sigdie+0x14)[0x7f6c37015e34]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libcsage.so(sage_signal_handler+0x20c)[0x7f6c37015a82]
/lib/libpthread.so.0(+0xf400)[0x7f6c3c26e400]
/usr/local/src/sage-4.7.2.alpha2/local/lib/python2.6/site-packages/sage/combinat/matrices/dancing_links.so(+0x738b)[0x7f6c1612f38b]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x4c32)[0x7f6c3c55d842]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCode+0x32)[0x7f6c3c55f912]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x4ccb)[0x7f6c3c55d8db]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5364)[0x7f6c3c55df74]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5364)[0x7f6c3c55df74]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5a8a)[0x7f6c3c55e69a]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5364)[0x7f6c3c55df74]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5364)[0x7f6c3c55df74]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x5364)[0x7f6c3c55df74]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCodeEx+0x879)[0x7f6c3c55f819]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalCode+0x32)[0x7f6c3c55f912]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyRun_FileExFlags+0xb0)[0x7f6c3c581d20]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyRun_SimpleFileExFlags+0xdf)[0x7f6c3c58275f]
/usr/local/src/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(Py_Main+0xb23)[0x7f6c3c58fa23]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f6c3b897ba6]
python[0x4006e9]

------------------------------------------------------------------------
Unhandled SIGSEGV: A segmentation fault occurred in Sage.
This probably occurred because a *compiled* component of Sage has a bug
in it and is not properly wrapped with sig_on(), sig_off(). You might
want to run Sage under gdb with 'sage -gdb' to debug this.
Sage will now terminate.
------------------------------------------------------------------------

Note that this behaviour is actually documented in sage/combinat/tiling.py:

    def is_suitable(self):
        r"""
        Return whether the volume of the box is equal to sum of the volume
        of the polyominoes and the number of rows sent to the DLX solver is
        larger than zero.

        If these conditions are not verified, then the problem is not suitable
        in the sense that there are no solution.

        .. NOTE::

            The DLX solver throws a Segmentation Fault when the
            number of rows is zero::

                sage: from sage.combinat.matrices.dancing_links import dlx_solver
                sage: rows = []
                sage: x = dlx_solver(rows)
                sage: x.search()        # not tested
                BOOM !!!

For sage-on-gentoo, there is an assertion failure:

sage:  x = dlx_solver(rows)
python2.7: sage/combinat/matrices/dancing_links_c.h:217: void dancing_links::setup_columns(): Assertion `nr_columns > 0' failed.
/usr/lib64/libcsage.so(print_backtrace+0x24)[0x7fea73a945f7]
/usr/lib64/libcsage.so(sigdie+0x1d)[0x7fea73a94687]
/usr/lib64/libcsage.so(sage_signal_handler+0x157)[0x7fea73a94822]
/lib64/libpthread.so.0(+0xfee0)[0x7fea79206ee0]
/lib64/libc.so.6(gsignal+0x35)[0x7fea78ea5ee5]
/lib64/libc.so.6(abort+0x186)[0x7fea78ea7896]
/lib64/libc.so.6(__assert_fail+0xf5)[0x7fea78e9e7a5]
/usr/lib64/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x855d)[0x7fea4e73255d]

Additionally, we can segfault with a different kind of trivial problem,

    sage: from sage.combinat.matrices.dancing_links import dlx_solver
    sage: d = dlx([[]])
    sage: d.search()
    *boom!*

And even adding rows can be a problem... this doesn't always go boom

    sage: d = dlx_solver([[0,1],[2]])
    sage: d.add_rows([[0,1],[1]])
    *boom!*

Actually... *why the hell* would we ever support the following:

    sage: d = dlx_solver(anything)
    sage: d.search()
    sage: d.add_rows(anything else)

[leif]

A bit more information on the platform, or did you try this on different systems?

On Ubuntu 10.04.3 x86_64 I get:

...
sage: sage: from sage.combinat.matrices.dancing_links import dlx_solver
==27537== Invalid read of size 4
==27537==    at 0x4EB5E83: PyObject_Free (obmalloc.c:931)
==27537==    by 0x4EDBE00: unicode_dealloc (unicodeobject.c:369)
==27537==    by 0x4E9C8E0: frame_dealloc (frameobject.c:420)
==27537==    by 0x4F1562E: PyEval_EvalCodeEx (ceval.c:2979)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==  Address 0x6156020 is 2,256 bytes inside a block of size 4,400 free'd
==27537==    at 0x4C27058: realloc (vg_replace_malloc.c:525)
==27537==    by 0x9F13E6A: xrealloc (xmalloc.c:74)
==27537==    by 0x9F1442E: add_history (history.c:302)
==27537==    by 0x9F1761F: read_history_range (histfile.c:270)
==27537==    by 0x9CE3089: read_history_file (readline.c:102)
==27537==    by 0x4F13E19: PyEval_EvalFrameEx (ceval.c:3706)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4E9DC2A: function_call (funcobject.c:524)
==27537== 
==27537== Invalid read of size 4
==27537==    at 0x4EB5E83: PyObject_Free (obmalloc.c:931)
==27537==    by 0x4E68C66: freechildren (node.c:135)
==27537==    by 0x4E68FAD: PyNode_Free (node.c:123)
==27537==    by 0x4F370C8: PyParser_ASTFromString (pythonrun.c:1439)
==27537==    by 0x4F37221: Py_CompileStringFlags (pythonrun.c:1382)
==27537==    by 0x4F0CE6B: builtin_compile (bltinmodule.c:546)
==27537==    by 0x4F13E19: PyEval_EvalFrameEx (ceval.c:3706)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4E9DB2E: function_call (funcobject.c:524)
==27537==    by 0x4E741A1: PyObject_Call (abstract.c:2492)
==27537==    by 0x4E868BC: instancemethod_call (classobject.c:2579)
==27537==    by 0x4E741A1: PyObject_Call (abstract.c:2492)
==27537==  Address 0xa73d020 is not stack'd, malloc'd or (recently) free'd
==27537== 
sage: sage: rows = []
sage: sage: x = dlx_solver(rows)
sage: sage: x.search()
==27537== Conditional jump or move depends on uninitialised value(s)
==27537==    at 0x2F2C3281: __pyx_pf_4sage_8combinat_8matrices_13dancing_links_20dancing_linksWrapper_9search(_object*, _object*) (dancing_links_c.h:75)
==27537==    by 0x4F13B46: PyEval_EvalFrameEx (ceval.c:3690)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F15A21: PyEval_EvalCode (ceval.c:522)
==27537==    by 0x4F14510: PyEval_EvalFrameEx (ceval.c:4401)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537== 
==27537== Use of uninitialised value of size 8
==27537==    at 0x2F2C32AA: __pyx_pf_4sage_8combinat_8matrices_13dancing_links_20dancing_linksWrapper_9search(_object*, _object*) (dancing_links_c.h:76)
==27537==    by 0x4F13B46: PyEval_EvalFrameEx (ceval.c:3690)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F15A21: PyEval_EvalCode (ceval.c:522)
==27537==    by 0x4F14510: PyEval_EvalFrameEx (ceval.c:4401)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537== 
==27537== Conditional jump or move depends on uninitialised value(s)
==27537==    at 0x2F2C32A8: __pyx_pf_4sage_8combinat_8matrices_13dancing_links_20dancing_linksWrapper_9search(_object*, _object*) (dancing_links_c.h:76)
==27537==    by 0x4F13B46: PyEval_EvalFrameEx (ceval.c:3690)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F15A21: PyEval_EvalCode (ceval.c:522)
==27537==    by 0x4F14510: PyEval_EvalFrameEx (ceval.c:4401)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537== 
==27537== Conditional jump or move depends on uninitialised value(s)
==27537==    at 0x2F2C3291: __pyx_pf_4sage_8combinat_8matrices_13dancing_links_20dancing_linksWrapper_9search(_object*, _object*) (dancing_links_c.h:76)
==27537==    by 0x4F13B46: PyEval_EvalFrameEx (ceval.c:3690)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F15A21: PyEval_EvalCode (ceval.c:522)
==27537==    by 0x4F14510: PyEval_EvalFrameEx (ceval.c:4401)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537== 
==27537== Invalid read of size 4
==27537==    at 0x2F2C328C: __pyx_pf_4sage_8combinat_8matrices_13dancing_links_20dancing_linksWrapper_9search(_object*, _object*) (dancing_links_c.h:76)
==27537==    by 0x4F13B46: PyEval_EvalFrameEx (ceval.c:3690)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F15A21: PyEval_EvalCode (ceval.c:522)
==27537==    by 0x4F14510: PyEval_EvalFrameEx (ceval.c:4401)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==    by 0x4F14CEE: PyEval_EvalFrameEx (ceval.c:3792)
==27537==    by 0x4F15947: PyEval_EvalCodeEx (ceval.c:2968)
==27537==    by 0x4F13BF3: PyEval_EvalFrameEx (ceval.c:3802)
==27537==  Address 0x30 is not stack'd, malloc'd or (recently) free'd
==27537== 
/tmp/Sage/sage-4.7.2.alpha2/local/lib/libcsage.so(print_backtrace+0x31)[0xbfa6817]
/tmp/Sage/sage-4.7.2.alpha2/local/lib/libcsage.so(sigdie+0x14)[0xbfa6849]
/tmp/Sage/sage-4.7.2.alpha2/local/lib/libcsage.so(sage_signal_handler+0x20e)[0xbfa6474]
/lib/libpthread.so.0(+0xf8f0)[0x52018f0]
/tmp/Sage/sage-4.7.2.alpha2/local/lib/python2.6/site-packages/sage/combinat/matrices/dancing_links.so(+0x828c)[0x2f2c328c]
/tmp/Sage/sage-4.7.2.alpha2/local/lib/libpython2.6.so.1.0(PyEval_EvalFrameEx+0x4af7)[0x4f13b47]
... (same backtrace)

[boothby]

Amusingly I found this in add_rows

        The following example would crash in Sage's debug version
        from :trac:`13864` prior to the fix from :trac:`13822`::

            sage: from sage.combinat.matrices.dancing_links import dlx_solver
            sage: x = dlx_solver([])          # indirect doctest
            sage: x.get_solution()
            []

[slabbe]

I just pushed a branch (based on 6.6.rc2) where the segmentation fault is catch using sig_on and sig_off.

Do we want to also solve the segmentation fault itself in this ticket?

---

New commits:

​8ee8dd8

Trac #11814: Catch the segmentation fault in dancing links

[git]

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

​25741df

Trac #11814: Merge with 6.8.rc1

[jdemeyer]

Replying to slabbe:

I just pushed a branch (based on 6.6.rc2) where the segmentation fault is catch using sig_on and sig_off.

Do we want to also solve the segmentation fault itself in this ticket?

First of all, I would absolutely never doctest segfaults unless you know that it's safe to do so.

I don't mind just adding the sig_on/sig_off in this ticket provided that you create a new ticket for the segmentation fault itself.

[jdemeyer]

Also: add your author name please.

[git]

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

​35f074c

Trac #11814: Added not tested flag for segmentation fault

[slabbe]

Replying to jdemeyer:

I don't mind just adding the sig_on/sig_off in this ticket provided that you create a new ticket for the segmentation fault itself.

I just created #18950 for the segmentation fault itself.

[jdemeyer]

Why needs_work?

[git]

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

​5495759

Trac #11814: Fix segmentation Fault raised by an assert

[slabbe]

Replying to jdemeyer:

Why needs_work?

Because I found the problem and was about to push the solution here. #18950 will not be necessary.

[slabbe]

I also removed the dumps method which was broken and useless:

sage: x = dlx_solver([])                                                
sage: x.dumps()                                          
Traceback (most ...
...              
AttributeError: 'list' object has no attribute 'dumps'   

I was about to remove make_dlxwrapper from the sage namespace, but decided it belongs to another ticket.

[jdemeyer]

The initialization of self.rows is a pure Python-level initialization, so I recommend to move it to __init__ instead of __cinit__.

[jdemeyer]

I took the opportunity to fix the initialization even more and really clean up the whole file.

---

New commits:

​392b62c

Simplify dancing links code

[slabbe]

Great! I was feeling those Py_INCREF were not necessary and old fashion. You were the good guy to clean/update that part. I removed the cimport Integer which was unused.

There is one doctest failing in combinat/tiling.py because of the __str__ -> __repr__ change. I did not fix the doctest because I wonder if showing a list of list is really what we want for the __repr__? Maybe a repr saying Dancing links of 34 rows would be better together with a method rows returning the rows?

---

New commits:

​3a3dd8b

Removed cimport Integer from dancing links

[slabbe]

Note: I started to fix the repr thing. Will push something soon.

[jdemeyer]

I would prefer to see the rows in __repr__, so maybe Dancing links of 2 rows [0,1,2], [0,2].

Note: the only reason I changed it is because it looked strange to me that the class had a __str__ method but not a __repr__.

[git]

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

​56c74ed

Better __repr__ for dancing links, added rows method

[slabbe]

With the last commit we now have:

sage: from sage.games.quantumino import QuantuminoSolver
sage: q = QuantuminoSolver(0)                           
sage: T = q.tiling_solver()                             
sage: T.dlx_solver()                                    
Dancing links solver for 96 columns and 5484 rows

I think it is better like this instead of printing the whole list of rows.

[jdemeyer]

There is a doctest failure in src/sage/games/quantumino.py related to the printing.

[git]

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

​e2b82a4

Fix doctest in games/quantumino.py

[slabbe]

Just to say that I agree with the changes made by Jeroen in this ticket.

[vbraun]

When building with SAGE_DEBUG=yes:

sage: from sage.combinat.matrices.dancing_links import dlx_solver ## line 163 ##
sage: x = dlx_solver([])          # indirect doctest ## line 164 ##
python: /mnt/disk/home/buildslave-sage/slave/sage_git/build/src/build/cythonized/sage/combinat/matrices/dancing_links_c.h:217: void dancing_links::setup_columns(): Assertion `nr_columns > 0' failed.
------------------------------------------------------------------------
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/ext/interrupt/interrupt.so(+0x3747)[0x7f012e49a747]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/ext/interrupt/interrupt.so(+0x3902)[0x7f012e49a902]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/ext/interrupt/interrupt.so(+0x3257)[0x7f012e49a257]
/lib64/libpthread.so.0(+0x10430)[0x7f013a8b8430]
/lib64/libc.so.6(gsignal+0x38)[0x7f0139e0d9c8]
/lib64/libc.so.6(abort+0x16a)[0x7f0139e0f65a]
/lib64/libc.so.6(+0x2d187)[0x7f0139e06187]
/lib64/libc.so.6(+0x2d232)[0x7f0139e06232]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(_ZN13dancing_links13setup_columnsEv+0x36)[0x7f00f9ec3fe8]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(_ZN13dancing_links8add_rowsESt6vectorIS0_IiSaIiEESaIS2_EE+0x153)[0x7f00f9ec4335]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x1301a)[0x7f00f9ebf01a]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x121c9)[0x7f00f9ebe1c9]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x161b6)[0x7f00f9ec21b6]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x16378)[0x7f00f9ec2378]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x1091e)[0x7f00f9ebc91e]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x106dd)[0x7f00f9ebc6dd]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(+0xdf3ed)[0x7f013aba33ed]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x160bb)[0x7f00f9ec20bb]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x13c13)[0x7f00f9ebfc13]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/python2.7/site-packages/sage/combinat/matrices/dancing_links.so(+0x13ac4)[0x7f00f9ebfac4]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(+0x15a60b)[0x7f013ac1e60b]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x7c85)[0x7f013ac19290]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x113f)[0x7f013ac1bd59]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(PyEval_EvalCode+0x41)[0x7f013ac115e6]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(+0x15d369)[0x7f013ac21369]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(PyEval_EvalFrameEx+0x432d)[0x7f013ac15938]
/mnt/disk/home/buildslave-sage/slave/sage_git/build/local/lib/libpython2.7.so.1.0(PyEval_EvalCodeEx+0x113f)[0x7f013ac1bd59]

[jdemeyer]

Sébastien, I won't have time in the coming month to look at this. If other people want to fix and to review, go ahead...

[slabbe]

No problem Jeroen!

So how can I reproduce the problem? Like this?

make distclean
export SAGE_DEBUG="yes"
make
sage -t src/sage/combinat/matrices/dancing_links.pyx

or is there an easier way? Can I just recompile some packages (python and cython and gcc?) packages with SAGE_DEBUG=yes? Does sage -f <package> considers the SAGE_DEBUG flag?

[vbraun]

You need to compile Python and everything that links to it, the only sure way is starting with make distclean imho.

[git]

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

​dcdfd06

Trac #11814: Fix failing Assertion nr_columns > 0 in setup_columns()