On Mon, 7 Jan 2008, mario alejandro wrote:
Yes, I read the docs. I took the GCJ route because the jcc route look far
more complicated, demand install a lot more and demanding a java start for
each request sound very bad. I need run under a 256 RAM limit, and I'm not
sure if this could scale well (I must fit the webserver, postgress, django,
python, nginx).
Think again :)
The JCC route is much easier to install and the runtime size much easier to
control. By default, depending on the OS, the Java VM uses around 32-64m.
There is no reason to believe (and much harder to measure) that libgcj,
gcj's VM, is using any less. Also, getting gcj to produce a sane PyLucene on
Solaris might be tricky.
In the other hand, why is so hard to install pylucene? I'm on google reading
about it a look like everyone have problems with this.
Yes, gcj is tricky to get right. On the other hand, jcc-PyLucene is almost
trivial to install in comparison.
Andi..
_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
