On Mon, 26 Mar 2007, Ofer Nave wrote:
So far I agree with you. Though I'm not ready to give up trying.
Excellent. The real solution lies with gcj/libgcj. I believe, gcj 4.2.0 (not yet released ?) actually has fixes that allow threads to be registered with libgcj's GC after the fact via the JvAttachCurrentThread() API and these fixes are probably (?) limited to Linux (Red Hat only ?) at the moment. I've had no time to investigate this much but I believe I saw mail about this on [EMAIL PROTECTED] If you don't want to give up, this is the most promising route as it would let you use PyLucene without the current stringent threading constraints.
Been dying to use subprocess since I started writing Python almost a month ago. Waiting on the admins to have time to upgrade from Python 2.3 on my dev box.
Ohhh, you're using Python 2.3 ? Python 2.3.x with x < 5 has some threading deadlock problem that was fixed in later versions.
Andi.. _______________________________________________ pylucene-dev mailing list [email protected] http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
