i noticed that starting a thread for each request is in fact slower
than using no thread at all.

thread pool should be a better solution but for now since the search
requests aren't pretty often we can live with it.

also compared to Java + Tomcat + Lucene solution it seems that using
pylucene + twisted (in my case) is slower, so migrating "back" to java
might be something.

this is my first mini project where i use entirely only pylucene.

the others have the "search" part in Java (Tomcat) and the "indexer"
part in pylucene (- command line tool), this way i have the easyest
way to build the index on a cron basis.

have a nice evening !

Friday, October 28, 2005, 7:32:43 PM, Andi Vajda wrote:
> Does processRequest start a thread for every request ?
> In Chandler, where we use Twisted and PyLucene together, we created a thread
> pool whose threads are Python.PyLucene thread instances and it seems to work
> well.


-- 
Catalin Constantin
Bounce Software
http://www.bounce-software.com
http://www.cabanova.ro

_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev

Reply via email to