henryrneh commented on issue #12165:
URL: https://github.com/apache/lucene/issues/12165#issuecomment-1674365549
Now we have started to do some bug triaging of bugs from OSS-Fuzz. There are
multiple issues discovered with the fuzzer, for example OutOfMemory or
StackOverflow, that we can disc
uschindler commented on issue #12165:
URL: https://github.com/apache/lucene/issues/12165#issuecomment-1674532419
Just open public issues.
Actually not all of those errors would be fixed, because Apache Lucene does
not always do all possible checks, as performance is more important th
benwtrent merged PR #12500:
URL: https://github.com/apache/lucene/pull/12500
--
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.a
sabi0 opened a new issue, #12501:
URL: https://github.com/apache/lucene/issues/12501
### Description
`Lucene70Codec` had:
```
private final PostingsFormat defaultFormat =
PostingsFormat.forName("Lucene50");
```
In the `Lucene80Codec` PostingsFormat instantiation was mo
uschindler commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674943197
Hi,
the SPI should only be used when READING indexes. When you create a codec
for IndexWriter the codec version hardcodes its postings formats and other
subtypes. As you see i
sabi0 commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674950075
I see. Thank you for the explanation.
The commits that change this behavior did not say anything about this.
So it looked like this SPI extension point loss was an unwitting side-
uschindler commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674955262
> So it looked like this SPI extension point loss was an unwitting
side-effect of a sequence of refactorings.
No, the SPI fromName does not allow you to change the implemen
uschindler commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674957426
For the FST loading mode mentioned above, the codec does not need to be
changed, you can tell DirectoryReader to use FST load modes using attributes.
--
This is an automated me
uschindler closed issue #12501: Default PostingsFormat lost the SPI extension
point in the Codec class
URL: https://github.com/apache/lucene/issues/12501
--
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
uschindler commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674970444
The general rule is: If you want to change the index postings format (but
nothing else like codec itsself) when writing a new index, you need to subclass
default codec. By that i
sabi0 commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674983089
The codec classes are `final`.
Besides having two implementations (`Lucene84PostingsFormat` in lucene-core
and `MyLucene84PostingsFormat`) with the same "Lucene84" name will likely
uschindler commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1674996502
> The postings format classes are `final`. Besides having two
implementations (`Lucene84PostingsFormat` in lucene-core and
`MyLucene84PostingsFormat`) with the same "Lucene84" na
ashvardanian opened a new issue, #12502:
URL: https://github.com/apache/lucene/issues/12502
### Description
I was recently approached by Lucene and Elastic users, facing low
performance and high memory consumption issues, running Vector Search tasks on
JVM. Some have also been using
henryrneh opened a new issue, #12503:
URL: https://github.com/apache/lucene/issues/12503
### Description
Dear Apache Lucene maintainers,
The OutOfMemory is triggered in this
[line](https://github.com/apache/lucene/blob/main/lucene/core/src/java/org/apache/lucene/util/ArrayUtil.
benwtrent commented on issue #12502:
URL: https://github.com/apache/lucene/issues/12502#issuecomment-1675079917
I don't think we need a native implementation. JNI stuff can be dangerous. I
honestly don't know the history around Lucene and if there have ever been
considerations in the area b
jbellis commented on issue #12502:
URL: https://github.com/apache/lucene/issues/12502#issuecomment-1675081860
Hi Ash,
(1) Have you compared usearch directly with Lucene? This could be a useful
starting point: https://github.com/jbellis/hnswrecall
(2) My understanding is that i
uschindler commented on issue #12502:
URL: https://github.com/apache/lucene/issues/12502#issuecomment-1675084211
Yes:
- no external libraries for Lucene Core
- no native code
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitH
benwtrent opened a new pull request, #12504:
URL: https://github.com/apache/lucene/pull/12504
This is a follow up to: https://github.com/apache/lucene/pull/12434
Adds a test for when parents are missing in the index and verifies we return
no hits. Previously this would have thrown an
benwtrent opened a new issue, #12505:
URL: https://github.com/apache/lucene/issues/12505
### Description
Lucene always does an approximate nearest neighbors search when no filter is
provided.
This seems like unnecessary work. Some benchmarks would have to be done, but
some id
jbellis commented on code in PR #12421:
URL: https://github.com/apache/lucene/pull/12421#discussion_r1291603910
##
lucene/core/src/java/org/apache/lucene/util/hnsw/ConcurrentNeighborSet.java:
##
@@ -0,0 +1,292 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
sabi0 commented on issue #12501:
URL: https://github.com/apache/lucene/issues/12501#issuecomment-1675193472
> > Besides having two implementations ... with the same "Lucene84" name
will likely result in a lookup error?
> Exactly and because of that its final.
I just do not understa
ashvardanian commented on issue #12502:
URL: https://github.com/apache/lucene/issues/12502#issuecomment-1675201748
Thank you, @benwtrent, @jbellis, and @uschindler! It's very insightful!
[Nmslib.java](https://github.com/opensearch-project/k-NN/blob/main/src/main/java/org/opensearch/knn/index
jpountz merged PR #12415:
URL: https://github.com/apache/lucene/pull/12415
--
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
reta commented on issue #12498:
URL: https://github.com/apache/lucene/issues/12498#issuecomment-1675403043
> It makes sense to me to push the responsibility of figuring out how to
execute tasks to the executor. Also pinging @reta.
Thanks @jpountz , I second that
> Additionally,
squirmy closed pull request #1681: SOLR-10804: Allow same version updates in
DocBasedVersionConstraintsProcessor
URL: https://github.com/apache/lucene-solr/pull/1681
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL
searchivarius commented on issue #12342:
URL: https://github.com/apache/lucene/issues/12342#issuecomment-1675721505
Looking great, many thanks! Could you remind me what is ordered and
reversed? This is something related to insertion order?
--
This is an automated message from the Apache G
26 matches
Mail list logo
