[pylucene-dev] fatal error in GC : too many heap sections.

[Bill Janssen]

> Today we are considering a two part approach. Eventually, we will evaluate
> whether to port to JCC pyLucene, or to  bite the bullet and go the full Java
> Lucene route. The GCJ version has served us well in a single thread app, and
> we believe a JCC version will work great, but sorting out the JCC/windows
> build process may be more risky for us than simply implementing Java Lucene,
> with all its horrors. [Simply? Did I just say that?]

The JCC build process is so much simpler than the GCJ build process.
But of course everything is hard (for developers) on Windows.

Re-writing in Java is an interesting option.  Last week I took an
indexing program I had in Java, and re-wrote it for JCC PyLucene.  It
shrunk to 1/6 the size of the Java version.  I also ported a search
program I'd written in GCJ PyLucene to JCC PyLucene.  It took two
days, but mainly that was poking at a bug in JCC (which got fixed
between the time I went to bed and the time I got up again in the
morning :-).  The code transferred over without any significant
changes.

Bill