leif wrote:
> leif wrote:
>> Ralf Stephan wrote:
>>> Does anyone else get these segfaults in the kernel log?
>>
>> Yup. But they seem to originate from builds, not testing.
>
> Or not. I can produce them (a single one) with
>
> ./sage -t --long src/sage/graphs/digraph_generators.py
>
> (Without '--long' "works" as well.)
Rebuilt nauty with debug info:
Starting program: ${SAGE_ROOT}/local/bin/gentourng -c 2
>A ${SAGE_ROOT}/local/bin/gentourng -cd1D0 n=2
Program received signal SIGSEGV, Segmentation fault.
makeleveldata () at gentourng.c:488
488 d->xstart[xcard[nxsets-1]+1] = nxsets;
(gdb) print nxsets
$1 = 0
So apparently array index out of bounds (xcard[-1] and/or d->xstart[...]).
Note that already the parameters get "echoed" wrong (same if one
explicitly passes '-d0 -D1', which Sage does).
Incidentally(?), that's exactly the case where no tournaments exist at
all (i.e., strongly connected ones on two vertices, which is a
contradiction by itself).
The corresponding doctest lists the counts of tournaments on up to eight
vertices (silently but also not explicitly ignoring the error):
sage: tournaments = digraphs.tournaments_nauty
[...]
sage: [len(list(tournaments(x, strongly_connected = True)))
for x in range(1,9)]
[1, 0, 1, 1, 6, 35, 353, 6008]
tournaments_nauty() is implemented as a generator using Popen() and
PIPE. I so far failed to get any returncode from the Popen subprocess;
poll() keeps returning None.
(Haven't opened a ticket [yet], nor reported anything upstream.)
-leif
>>> Aug 29 17:28:04 ark kernel: gentourng[1549]: segfault at 40065a02c ip
>>> 00000000004018
>>> 1f sp 00007ffce9d03520 error 4 in gentourng[400000+19000]
>>> Aug 30 17:03:54 ark kernel: gentourng[20347]: segfault at 401d4402c ip
>>> 0000000000401
>>> 81f sp 00007ffe5bca7d60 error 4 in gentourng[400000+19000]
>>> Aug 31 11:44:22 ark kernel: gentourng[1779]: segfault at 401f1502c ip
>>> 00000000004018
>>> 1f sp 00007fff9a7f4720 error 4 in gentourng[400000+19000]
>>> Aug 31 15:09:49 ark kernel: gentourng[10162]: segfault at 401d8502c ip
>>> 0000000000401
>>> 81f sp 00007fff310eeea0 error 4 in gentourng[400000+19000]
>>>
>>> I can associate them with patchbot or ptestlong runs.
>>> They go as far back as Apr 20 (limited by my kernel log).
--
You received this message because you are subscribed to the Google Groups
"sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-release.
For more options, visit https://groups.google.com/d/optout.
