I, I'm sorry to say, gave up on PyLucene.PythonThread.
Too bad. So what do you do now?
Basically, the only thing I did (with cherrypy 2.1(svn 525)) was to pass a custom serverClass to cherrypy.server.start() (The WorkerThread in my derived _cherrypywsgiserver inherits from PyLucene.PythonThread in stead of threading.Thread).On Linux this sort of works, sometimes...but never for long.Un-scientifically speaking - most times when I start the server I get "GC Warning: Out of Memory! Returning NIL!", but sometimes the server starts without errors. When the server is started it will report "socket.error: (11, 'Resource temporarily unavailable')" on any attempt of executing a PyLucene enabled method.The funny thing is that the exact same code works 100% on Mac OS X (10.4.2). Unfortunately for me my servers are running Linux.
I basically did the same thing (but am using Win 2k). Not very effective. :(
Did some work yesterday night and have resorted to using an XMLRPC server which gives back search results to the cherrypy page. And am not getting this problem (No sight of those cursed threads I guess). Hopefully this will work solid.
But such a pity that there is no simpler approach.
Thanks to all.
Regards,
Deepak
_______________________________________________ pylucene-dev mailing list [email protected] http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
