Hi Team,
since we released Lucene 2.1 in February there have been powerful new
features
like "point-in-time searching" (LUCENE-710), "Payloads" (LUCENE-755) and
"API for pre-analyzed fields" (LUCENE-580), good performance
improvements like
"multi-level skipping" (LUCENE-866), "improved buffering" (LUCENE-888) and
"improved RAMDirectory performance" (LUCENE-431). In addition there have
been
various other optimizations and bug fixes.
Considering all these improvements I think it's time for a new release,
especially since many of you voted in February to have releases more
frequently.
Another good thing is that since 2.1 we improved the build files
significantly.
In 2.1 we had problems with failing contrib tests and wrong build files
in the
binary package. Thanks to Hoss and Doron (and other contrib owners) we fixed
already most of those contrib problems, and I'm working on fixing the binary
build problem (LUCENE-894).
I would volunteer to do the release work this time. In case of no
negative votes
we could try to aim for a release date in about 2 weeks?
If we can decided that we want to make a 2.2 release as I suggest I will go
ahead and create a staging area in my home directory like Yonik did with
2.1.
I really liked that approach. Next week I will then upload binary and
source
packages built from the current trunk (after I committed 894). This will
*not*
be a release candidate yet, but is supposed to give us the chance to
test the
packages on different platforms to ensure that all build problems we had
in 2.1
are solved.
There are currently 9 issues in Jira targeted for 2.2:
- LUCENE-510: "IndexOutput.writeString() should write length in bytes",
Grant Ingersoll
Not much progress has been made here lately, so I think we should move
this
to 2.3?
- LUCENE-446: "search.function - (1) score based on field value, (2) simple
score customizability", Doron Cohen
This looks ready to commit, Doron?
- LUCENE-894: "Custom build.xml for binary distributions", Michael Busch
I'm planning to commit this soon.
- LUCENE-854: "Create merge policy that doesn't periodically inadvertently
optimize", Michael McCandless
Do you want to get this into 2.2, Mike?
- LUCENE-887: "Interruptible segment merges", Unassigned
I will move this to a later release.
The following 4 issue are about contrib packages and build warnings, I think
they'd be nice to get in but shouldn't hold up the release.
- LUCENE-875: "javadocs creation has too many warnings/errors", Doron Cohen
- LUCENE-876: "contrib/gdata has many javadoc warnings", Unassigned
- LUCENE-759: "Add n-gram tokenizers to contrib/analyzers", Otis Gospodnetic
- LUCENE-848: "Add supported for Wikipedia English as a corpus in the
benchmarker
stuff", Grant Ingersoll
Any others that should go in?
- Michael
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
