I can't tell what exactly is leaking there, obviously it doesn't happen on
simple searches or it would have been caught earlier. It seems to me it has
to do with a partial reusableTokenStream implementation we indeed fixed in
the 2_3_2 branch. Try to see if implementing the following patches on 0.9.21
works for you.
 
http://clucene.git.sourceforge.net/git/gitweb.cgi?p=clucene/clucene;a=commit
diff;h=364c21b6c3f54fbb90df223621b660197366fb93
http://clucene.git.sourceforge.net/git/gitweb.cgi?p=clucene/clucene;a=commit
diff;h=c89f8a39fa1faa34374d8a6e92ae9c2467deeda7
 
In any case, no further releases are planned for 0.9.21, as all of our very
limited resources are focused at getting the 2_3_2 branch more mature.
 
HTH
 
Itamar.

  _____  

From: Niki Yoshiuchi [mailto:[email protected]] 
Sent: Friday, April 09, 2010 12:00 AM
To: [email protected]
Subject: [CLucene-dev] Memory leak in 0.9.21



Hey, I'm using CLucene 0.9.21 on a project and I think I've stumbled across
a memory leak.  Specifically it seems that the function:

 

Void IndexSearcher::_search(Query* query, Filter* filter, HitCollector
*results)

 

In IndexSearcher.cpp is leaking a TermQuery in the call to
query->weight(this) on line 310.  To create the weight the function:

 

Query* TermQuery::clone() const

 

Is being called which returns a new pointer, and it doesn't appear to be
getting deleted anywhere.

 

My question is, is there an easy way to fix this?  I'd rather not update to
2.3.2 right now which I'm assuming fixes this issue.

 

Thanks,

-Niki Yoshiuchi

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
CLucene-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/clucene-developers

Reply via email to