Issue 65: recoll 1.16.0 ubuntu lucid crash on indexing from gui

Gui repeatedly crashes when trying to index after upgrade. Same issue whether trying to update exsinging xapiandb or creating a new index after I deleted the old xapiandb.

I then started recol from commandline like this (several times) trying to gather some more information. My cautious guess is that it chokes on some some character encoding issue, that is handled differently in 1.16 from 1.15.9. I can provide further information if needed.

name@server:_{$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-2.txt Segmentation fault name@server:}$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-3.txt * glibc detected * recoll: free(): corrupted unsorted chunks: 0x0000000003078f80 * Segmentation fault name@server:_{$ name@server:}$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-4.txt Segmentation fault name@server:_{$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-5.txt Aborted name@server:}$ name@server:_{$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-6.txt Segmentation fault name@server:}$ name@server:_{$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-7.txt} * glibc detected * recoll: free(): corrupted unsorted chunks: 0x00007ffa24016f30 * Segmentation fault name@server:$ recoll -c "/path/to/config" 1 > recoll-idx-errors-1.txt 2 >recoll-idx-erros-8.txt Segmentation fault

I’d be very interested by the trace files and possibly a stack trace or core dump. I’ve tested on lucid and seen nothing special.

If you are unsure how to proceed, there are some instructions here: https://bitbucket.org/medoc/recoll/wiki/GettingAStackTrace

You can reach me directly at jfd@recoll.org if you wish, this may make for a more convenient exchange than multiple comments on the bitbucket bug report.

Cheers, jf

Hello, And I forgot, the nicest gift possible would be the file that causes the crash of course !

jf

Thanks for your response I will provide that as soon as possible. However, I have reverted to 1.15.9 and don’t have time to experiment before about 2 weeks from. By then I will return here with more information. fsando

threading cleanup: add mutex protection around moronic change to transcode. Add mutex to equiv issue in unac. Rename const strings everywhere to cstr_xx to ease future detection of potentially problematic static variables. Most probably close issue #65

→ <<cset 0d7b30d12f26 > >