-Push for a Solr 1.4.1 Bug Fix Release?
Chris Hostetter 2010-05-28, 22:23
I remember thinking several months back, when some Jira fix caught my eye,
"Man, that's a heinous bug, we should really put out a 1.4.1 bug fix
release for that once it's committed." Then i evidently got distracted,
and there was a lot going on with the trunk merge, and then the 3x branch
creation, etc... and i evidently forgot all about it.
Auditing all of the "post 1.4" Jira issues yesterday, Nine (*9*) Fixed
bugs stood out to me as being really, REALLY bad -- bugs that made me
think "Man, that was a really heinous bug, why didn't we put out a 1.4.1
bug fix release for that???". In addition, there where nine other bugs
that seemed anoyingly trivial that if we *do* have a 1.4.1 bug fix
release, we should try to include them. (See below for these two bug
I get the sense that people are itching to do a 3.1 release of Lucene &
Solr "soon" -- but realizsticly speaking "soon" seems like it will be at
least a few months (if nothing else, since this will be the first "merged"
release, there are probably some release process logistics that will need
to be shaken out).
So my questions is: How many other people agree that it's worthwhile to do
a 1.4.1 release?
This would be a "bug fix" release, so there is no need for "oh wait, i
want better flesh out this API/Feature" hesitation. And we would release
from the svn:lucene/solr/branches/branch-1.4/ so none of the trunk merge
build system changes would really affect things. It should be relatively
painless, all we need to do is merge the existing fixes for these issues
onto that branch -- and that *should* be straight forward: all of these
issues were already fixed on the 1.5-dev branch, so theres almost no
chance they depend on much new Lucene 3x stuff (and if i'm wrong, we just
skip that issue)
I'm willing to step up and do most of the backporting work -- but i don't
want to waste that time if there aren't a few other people ready to help
with the testing and the rest of the release process. In particular i'd
rather not be the release manager (i don't even think i can be since i
don't have a PGP Key) on top of doing the back porting. So i'm at least
looking for a commitment from one other person whose willingto help ou
So what do people think?
## 9 Bugs w/fixes on the 1.5 branch that seem serious enough
## that they warrant a 1.4.1 bug-fix release...
FunctionQueries aren't correctly weighted
FastLRUCache doesn't correctly implement LRU
StreamingUpdateSolrServer doesn't specify UTF-8
capitalizationfilter crashes using maxWordCountOption
BufferedTokenStream incorrect cloning
Race condition in StreamingUpdateSolrServer
fields with sortMissingLast don't sort correctly
Memory leak in FastLRUCache
## 9 Other bugs that already have fixes on the 1.5 branch
## that seem simple enough to backport if we are going
## to have a 1.4.1 bug-fix release...