| 1 |
Hi Sage-on-Gentoo-amd64 users, |
| 2 |
|
| 3 |
yesterday I have tested the recent version of Sage to check if the problem |
| 4 |
Steven first reported remain - which unfortunately is the case. I tried and |
| 5 |
noticed the following things: |
| 6 |
|
| 7 |
- When running Sage's tests (I do that with ". local/bin/sage-env && sage- |
| 8 |
starts && sage-maketest"), I notice that this problem seems to be randomly |
| 9 |
triggered - I tried to run failing tests manually, which then succeeded. Also, |
| 10 |
another testrun showed failures with different testcases. |
| 11 |
- This problem showed up first with Sage 4.4 (Steven reported this) and remains |
| 12 |
in 4.4.1, upgrading to python-2.6.5-r99 did not change anything (I upgraded |
| 13 |
python and ran revdep-rebuild, but I did not recompile sage-core/sage-clib). |
| 14 |
- As long as I did not try to exit Sage everything seems to be ok - I opened |
| 15 |
Sage with "sage -gdb" and typed several simple commands (2+2, maxima('2+2'), |
| 16 |
integral(sin(x),x,0,pi)) which worked fine - when I type quit there is the |
| 17 |
possibilty that Sage exits with failure. GDB (I typed "bt" for backtrace) |
| 18 |
reports now the known error: |
| 19 |
|
| 20 |
*** glibc detected *** /usr/bin/python2.6: corrupted double-linked list |
| 21 |
|
| 22 |
The backtrace does not give very much information just function calls in libc |
| 23 |
and python itself. |
| 24 |
- I removed the patch for sage-clib which replaces gmp with mpir - this did |
| 25 |
not change anything noticable. |
| 26 |
- I just found this bugreport from numpy which could trigger this problem: |
| 27 |
|
| 28 |
http://projects.scipy.org/numpy/ticket/1464 |
| 29 |
|
| 30 |
Cheers, |
| 31 |
|
| 32 |
Christopher |
Archives