At ApacheCon EU I roped one of the most productive developers (Thomas
Dudziak) I know (who also has SWIG experience =) into the Ruby/Lucene
thing. Anyway, he's had little success with gcj and lucene4c thus far
(lucene4c isn't quite complete enough, and as Garrett knows (and said
he's working on) kind of tough to build.
Anyway, Thomas went and in an afternoon put SWIG bindings around
CLucene =)
Now, the more fun part, Ben (whose email I don't have) of CLucene
would like to move the project to Apache =)
Thoughts?
-Brian
On Aug 8, 2005, at 6:34 AM, Thomas Dudziak wrote:
Hi,
after much tinkering and installing/reinstalling gcc/gcj (3.4.3 and
4.0.1) I finally got a combo of ruby+swig+gcj to compile, only to be
stopped dead by a internal compiler error of GCJ. I honestly don't
know why this works for PyLucene (which btw. I didnt' get to compile
because the mac version of Python is 2.3 whereas PyLucene seems to
require 2.4).
And then yesterday by chance I spotted a mail by Ben van Klinken on
the SWIG mailing list who is the lead developer of the CLucene project
(http://clucene.sourceforge.net/), a full C++ port of Lucene. So I
fired up an email to him and he told me that they've rewritten CLucene
to be easily usable with SWIG (currently he's doing a C# and COM
wrapper for CLucene) and they already have more or less the
functionality as Lucene 1.4.3.
So I decided to give it a try, and after about half an hour I not only
had CLucene compiled and linked, but also a basic SWIG ruby wrapper
around one of the helper classes of CLucene (compared to about a week
for the same using gcj).
The interesting thing now is that they'd like to move to Apache, they
even proposed incubation
(http://clucene.sourceforge.net/incubatorproposal.htm) though they
seem to be missing a sponsor (Erik didn't answer as far as I could see
on the Lucene dev mailing list).
I'd very much like to use CLucene as the basis for the ruby binding
(and Ben is quite willing to help with any SWIG wrappers and C++
issues), so my question is: could you talk to Erik as to whether it
would be possible to accept the incubation proposal (via sponsoring by
the Lucene PMC) ? From what I saw so far of CLucene, I might be able
manage to create a ruby binding of the querying in August, which would
be a good start for the RubyLucene repository.