lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael McCandless <luc...@mikemccandless.com>
Subject Re: Lucene 2.9.3 ? ( blocking Solr 1.4.1 ? ? ? )
Date Tue, 01 Jun 2010 19:34:43 GMT
On Tue, Jun 1, 2010 at 2:48 PM, Chris Hostetter
<hossman_lucene@fucit.org> wrote:
>
> : I'd really like to get most of those back-ported, for 2.9.3.  I'm
> : actively working on this today.  I was the one who went and marked
> : them as 2.9.3 in the first place on Sat :).
>
> Ok ... i mainly just wanted ot make sure we were all on the same page
> about what issues were being targed for 2.9.3.
>
> My main confusion is that if these issues are UNRESOLVED, then presumably
> they either don't have fixes yet, *OR* people aren't yet comfortable
> commiting hte patches we do have to trunk -- in the first case: there's
> nothing to backport, in hte second case do we really want to backport to a
> bug fix branch if we aren't even confident enough to commit to trunk?
>
> ...looking at the list in detail, i realize now that a few of them were
> commited to trunk (and resolved) you just reopend (presumably to better
> track the backport) ...

Well, most of them were this case (reopened for backport).  But I've
been whittling those ones down, today...

> but the majority don't have any patch at all.

Hmm only 3 right?  You can see them as the OPEN ones (not REOPENED),
here:

https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&mode=hide&sorter/order=DESC&sorter/field=priority&resolution=-1&pid=12310110&fixfor=12314799

> it would be nice to fix every known bug immediately, but personally i'm
> more concerned with having a release "soon" that includes the bug fixes we
> already have then i am with having a release "at some point in the
> undetermined future" with a fix for every known bug.

In general I agree (we should not hold up a point release if the "dev"
is not done yet)... there were many other bugs in my sweep that I
didn't mark for 2.9.3.

But the two that I marked for 2.9.3 I think are trivial & low risk &
highish importance to fix.

> So the real question is: what kind of timeline do folks think we're
> looking at for hte 2.9.3 issues that don't even have patches yet? ...

I plan to finish my 2 by eod tomorrow...

> https://issues.apache.org/jira/browse/LUCENE-2348

This one I agree we should push out -- I didn't mark it as 2.9.3, and
there's no patch and on looking at it I don't understand where the
problem is.  I'll go push it out...

> https://issues.apache.org/jira/browse/LUCENE-2356

This one makes near real-time search unusable by certain apps.  It's a
trivial fix (setter/getter, forward), to have IW specify the terms
index divisor when it opens internal segment readers.

> https://issues.apache.org/jira/browse/LUCENE-2311

This one is also quite bad -- for apps that need to warm readers
before deploying, this also makes near real-time search unusable since
you can't actually do "real" warming (eg you can't run a Query).  I
think this is also a simple fix, but I'll go dig into it next to make
sure...

> (2356 & 2311 don't seem like "bug fixes" at all -- unless i'm completley
> missunderstanding hte issue descriptions -- so they *really* don't seem
> suitable for 2.9.3)

I consider them bugs, with trivial fixes, in NRT.

Mike

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message