Hello,
porting Sage to Gentoo progresses well, we now have a "split" version
where all all dependencies of Sage are installed with portage.
Unfortunately, we experience serious problems on amd64 when exiting
Sage. When doing so, there is a slight chance of getting a SIGABRT
error, which is always looking like this one:
sage -t -force_lib "devel/sage/doc/en/constructions/rings.rst"
*** glibc detected *** /usr/bin/python2.6: corrupted double-linked
list: 0x0000000004afc6e0 ***
======= Backtrace: =========
/lib/libc.so.6[0x7f8fa7a7f808]
/lib/libc.so.6[0x7f8fa7a7fb00]
/lib/libc.so.6[0x7f8fa7a81078]
/lib/libc.so.6(cfree+0x6c)[0x7f8fa7a8424c]
/lib/libc.so.6[0x7f8fa7a40ea5]
/lib/libc.so.6[0x7f8fa7a40ef5]
/usr/lib/libpython2.6.so.1.0[0x7f8fa80bc981]
/usr/lib/libpython2.6.so.1.0[0x7f8fa80bb1e5]
/usr/lib/libpython2.6.so.1.0(PyErr_PrintEx+0x23)[0x7f8fa80bb20a]
/usr/lib/libpython2.6.so.1.0(PyErr_Print+0xe)[0x7f8fa80baf41]
/usr/lib/libpython2.6.so.1.0(PyRun_SimpleFileExFlags+0x248)
[0x7f8fa80baa9d]
/usr/lib/libpython2.6.so.1.0(PyRun_AnyFileExFlags+0x79)
[0x7f8fa80ba335]
/usr/lib/libpython2.6.so.1.0(Py_Main+0xad9)[0x7f8fa80cbafa]
/usr/bin/python2.6(main+0x1b)[0x400847]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f8fa7a2aa26]
/usr/bin/python2.6[0x400769]
followed by a memory dump which is aborted and the error message: The
doctested process was killed by signal 6
The problem is also that these failures seem to be completely random -
running the failing tests manually again does not show this error
(with a high probability).
I already tried to debug this error by starting Sage with -gdb and
typed several times "2+2" and "maxima('2+2')". Again, this worked in
most cases but failed in some. After triggering this error I was able
to find out that PyRun_SimpleFileExFlags (see backtrace above,
declared in Python/pythonrun.c:892) invokes PyRun_FileExFlags which
returns NULL - PyRun_SimpleFileExFlags then tries to print out an
error with PyErr_Print (see backtrace) which leads to SIGABRT.
I have struggled with this error a long time and now I am running out
of ideas ... So these are my questions:
- Is this type of error known ? Has it already been fixed in Sage (I
know that the recent gcc-4.5 error look alike, but we are still using
gcc-4.3.4) ?
- If not, are there any ways to find out more about this error ?
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
