As the title says, I am experiencing segmentation faults for some
queries, if I access my index from Perl. Unfortunately the index in
question is too large too share, but I will be attempting to cut it down
to manageable levels.
Here some more details about my setup and further observations I made
that might help to pin down the problem.
We are talking about Debian Wheezy 64bit (3.2.0-4-amd64 #1 SMP Debian
3.2.46-1+deb7u1 x86_64 GNU/Linux) with the swish-e Debian package
installed. The index is generated with the swish-e binary and a Perl
script is used to access the index to produce search results.
Normally this works fine but recently I stumbled on a particular query
which leads to a segmentation fault when running the Execute() method on
the Perl search object.
In case it helps, the particular query is
(domain="www.biologie.uni-hamburg.de") AND (structure=start)
If I remove single characters from the end of the url, there continue to
be segmentation faults until I reach the query
(domain="www.biologie.un") AND (structure=start)
I compiled different SWISH-E Versions (2.4.3 to 2.4.7) with the API and
recreated the index, but got the same segmentation fault for all versions.
Only when I switch to compiling 32bit versions of SWISH-E on this
machine the problem goes away, so it might be a 64bit architecture
On a true 32bit machine the Debian packages work fine with the problem
Oh, and the swish-e binary has no problems with the query on the command
line, it's just the Perl API that acts up.
I hope you have some ideas on this.
Technik Blinde Kuh (__)
Die Suchmaschine fuer Kinder (__)\ )\/\
Tel.: ++49 (0)40 32034076 ||----w | \
Ort : D-20359 Hamburg - Pinnasberg 29-33 || ||
Mehr Infos dazu unter http://www.blinde-kuh.de/impressum.html
Users mailing list
Received on Fri Oct 25 2013 - 09:32:01 GMT