jpountz commented on PR #11997:
URL: https://github.com/apache/lucene/pull/11997#issuecomment-1359057180
Thanks for having a look @fcofdez. I plan on merging this PR soon if there
are no objections.
--
This is an automated message from the Apache Git Service.
To respond to the message, pl
javanna commented on code in PR #12019:
URL: https://github.com/apache/lucene/pull/12019#discussion_r1053216655
##
lucene/backward-codecs/src/test/org/apache/lucene/backward_codecs/lucene94/Lucene94RWCodec.java:
##
@@ -0,0 +1,46 @@
+/*
+ * Licensed to the Apache Software Foundat
javanna commented on PR #12019:
URL: https://github.com/apache/lucene/pull/12019#issuecomment-1359227916
I am not sure how to label this change, and whether it needs an entry in the
changelog or not.
Also, once I merged it I should backport it to 9x, correct?
--
This is an automate
rmuir commented on PR #12019:
URL: https://github.com/apache/lucene/pull/12019#issuecomment-1359307445
You can always use "Other" category. +1 to backport
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above t
javanna merged PR #12019:
URL: https://github.com/apache/lucene/pull/12019
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apa
javanna commented on PR #12019:
URL: https://github.com/apache/lucene/pull/12019#issuecomment-1359322771
Thanks @benwtrent !
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
jpountz merged PR #11997:
URL: https://github.com/apache/lucene/pull/11997
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apa
jpountz closed issue #11199: Add IntField, LongField, FloatField and
DoubleField classes to index both points and doc values [LUCENE-10162]
URL: https://github.com/apache/lucene/issues/11199
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
stevenschlansker opened a new issue, #12026:
URL: https://github.com/apache/lucene/issues/12026
### Description
I am doing some performance profiling with Java Flight Recorder (if you
haven't tried it yet, you should, it's awesome!) and am getting diagnostic
warnings about thrown exc
rmuir commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360035214
It is harmless, it isn't a performance issue really, it only happens at most
once per segment. Exception is the correct mechanism as it is an exceptional
case to bail on processing he
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360056272
Here's an example of what I am seeing:
--
This is
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360058121
This is during an intensive indexing operation, and it is using a
near-real-time reader created from an IndexWriter. I wonder if we are somehow
in an unusual spot where we
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360058923
It's also worth noting that we are firing off many small count queries,
rather than executing one large query. So maybe we are visiting all segments
many more times than us
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360063018
Would a change that at least pre-allocates a single exception instance,
rather than making a new one every time, be welcome / desired?
--
This is an automated message fro
rmuir commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360081092
You are doing count queries that get implemented with Weight.count().
Separately, that's really strange, i dont know what you are doing, but that
isn't very typical. e.g. if you are j
rmuir commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360097390
Also, when using this flightrecorder tool, to help reduce noise you need to
pass a few flags (e.g. `DebugNonSafepoints`) or you WILL see a ton of ghosts,
this could easily be one of t
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360140114
Our current use case involves doing a number of existence-tests of boolean
queries of term queries, like `count(docType = A AND docId = B) == 0`, which is
why we end up doi
stevenschlansker closed issue #12026: CollectionTerminationException is used
for flow control
URL: https://github.com/apache/lucene/issues/12026
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the s
rmuir commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360146258
ok, thanks for the explanation. `BooleanWeight.count` can only help in "rare
cases" IMO here, but it isn't rare enough in your measurements (e.g. maybe for
some segments, all document
stevenschlansker commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360183432
Another note: we have `-XX:-OmitStackTraceInFastThrow` on our command line,
which inhibits JVM optimization for exception stack traces. So that probably
makes it yet more e
rmuir commented on issue #12026:
URL: https://github.com/apache/lucene/issues/12026#issuecomment-1360373248
that one isn't relevant, it doesn't apply to our own custom exception here.
I would just rerun with DebugNonSafePoints before even worrying about it.
Otherwise profiler can waste imme
zacharymorn opened a new pull request, #12027:
URL: https://github.com/apache/lucene/pull/12027
### Description
Remove IOContext from `Directory#openChecksumInput` as suggested in
https://github.com/apache/lucene/issues/11933
--
This is an automated message from the Apache Git
22 matches
Mail list logo
