rmuir opened a new pull request, #14869:
URL: https://github.com/apache/lucene/pull/14869
* EmptyBlockTag: ecj linter detects missing descriptions for all these.
* StringCharset: ast-grep rule
* SunApi: forbidden-apis takes care
* ThreeLetterTimeZoneID: ast-grep rule
--
Th
rmuir commented on code in PR #14867:
URL: https://github.com/apache/lucene/pull/14867#discussion_r2174050512
##
gradlew:
##
@@ -289,4 +291,12 @@ eval "set -- $(
tr '\n' ' '
)" '"$@"'
-exec "$JAVACMD" "$@"
+if [[ "$buildarg" == "eclipse" ]]; then
+ "$JAVACMD" "$
rmuir opened a new issue, #14868:
URL: https://github.com/apache/lucene/issues/14868
### Description
Cause of a lot of my confusion over #14865. I don't think this gradle logic
is working properly, I think the change is very recent.
No `.settings/ folder` is created, which is t
vsop-479 commented on PR #14777:
URL: https://github.com/apache/lucene/pull/14777#issuecomment-3017423555
> what motivated this change? Was it performance related or general code
hygiene?
Thanks @vigyasharma , Since we use grouping function (self-developed base on
`FirstPassGroupingC
msokolov commented on issue #14864:
URL: https://github.com/apache/lucene/issues/14864#issuecomment-3017293810
Something I don't understand with the test framework here ... i think when
we re-run with the same seed, we should get repeatable results from `random()`
but I'm seeing different o
vigyasharma commented on issue #13883:
URL: https://github.com/apache/lucene/issues/13883#issuecomment-3017158053
I think both of them are viable options with pros and cons.
With the `MultiTenantCMSManager` approach, you can probably avoid making
deep changes to existing CMS logic. T
vigyasharma merged PR #14860:
URL: https://github.com/apache/lucene/pull/14860
--
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
rmuir commented on PR #14867:
URL: https://github.com/apache/lucene/pull/14867#issuecomment-3016916900
I added a thermonuclear hack on top of the other hacks, at least so it has
correct behavior.
All I want here is, when you run `./gradlew eclipse`, that it does exactly
what it does
rmuir commented on PR #14867:
URL: https://github.com/apache/lucene/pull/14867#issuecomment-3016848711
Thanks for looking, no rush. It's still the same approach. High level I just
want `gradle eclipse` to download source jars and attach the location as XML
attribute.
Applying the pl
dweiss commented on code in PR #14862:
URL: https://github.com/apache/lucene/pull/14862#discussion_r2173824500
##
build-tools/build-infra/src/main/java/org/apache/lucene/gradle/plugins/java/JavaFolderLayoutPlugin.java:
##
@@ -0,0 +1,82 @@
+/*
+ * Licensed to the Apache Software
dweiss commented on PR #14867:
URL: https://github.com/apache/lucene/pull/14867#issuecomment-3016837978
Can't do it today, Rob - away from my computer. I'll try to see what it does
tomorrow though. If I remember correctly, we used a single top-level Eclipse
project, without separating subpr
lukewilner commented on issue #13883:
URL: https://github.com/apache/lucene/issues/13883#issuecomment-3016833540
We're currently exploring two possible designs for implementing multi-tenant
coordination of merge threads across multiple IndexWriter instances in the same
JVM. One option is to
rmuir opened a new pull request, #14867:
URL: https://github.com/apache/lucene/pull/14867
hackishly applies "eclipse" plugin to all subprojects. this forces it to
download the source jars, which doesn't happen today.
pokes around in the gradle cache to find corresponding source jars a
rmuir commented on issue #14865:
URL: https://github.com/apache/lucene/issues/14865#issuecomment-3016743040
First step is to get eclipse plugin to download the correct stuff: this
isn't happening today because it doesn't actually run on the subprojects.
I add this hack and the `-sourc
github-actions[bot] commented on PR #14866:
URL: https://github.com/apache/lucene/pull/14866#issuecomment-3016696862
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
rmuir opened a new pull request, #14866:
URL: https://github.com/apache/lucene/pull/14866
the typechecker will automatically do this, for the linter we need to set an
environment variable.
![Screen_Shot_2025-06-29_at_09 26
45](https://github.com/user-attachments/assets/34d689cd-2bca-
rmuir opened a new issue, #14865:
URL: https://github.com/apache/lucene/issues/14865
### Description
Currently, gradlew eclipse creates a `.classpath` file, but it only contains
jars of `kind="lib"`.
This means there's e.g. no documentation for third-party libs. Not much of a
rmuir commented on code in PR #14862:
URL: https://github.com/apache/lucene/pull/14862#discussion_r2173755259
##
build-tools/build-infra/src/main/java/org/apache/lucene/gradle/plugins/java/JavaFolderLayoutPlugin.java:
##
@@ -0,0 +1,82 @@
+/*
+ * Licensed to the Apache Software F
msokolov commented on issue #14864:
URL: https://github.com/apache/lucene/issues/14864#issuecomment-3016639175
./gradlew -p lucene/core test --tests
TestLucene99ScalarQuantizedVectorsFormat.testReadQuantizedVectorWithEmptyRawVectors
-Dtests.seed=E75EAB09A838DE99 -Dtests.locale=en-GD
-D
msokolov opened a new issue, #14864:
URL: https://github.com/apache/lucene/issues/14864
### Description
Build: https://jenkins.thetaphi.de/job/Lucene-main-macOS/13150/
Java: 64bit/hotspot/jdk-24.0.1 -XX:+UseCompressedOops -XX:+UseParallelGC
2 tests failed.
FAILED:
org.ap
rmuir merged PR #14853:
URL: https://github.com/apache/lucene/pull/14853
--
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.apach
kaivalnp commented on PR #13497:
URL: https://github.com/apache/lucene/pull/13497#issuecomment-3016585891
Thanks @benwtrent! I opened #14863
--
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 sp
kaivalnp commented on issue #13515:
URL: https://github.com/apache/lucene/issues/13515#issuecomment-3016586648
FYI I opened #14863 for off-heap quantized scoring, would appreciate reviews!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
kaivalnp commented on PR #14863:
URL: https://github.com/apache/lucene/pull/14863#issuecomment-3016581312
FYI I observed a strange phenomenon where if the query vector is on heap
like:
```java
this.query = MemorySegment.ofArray(targetBytes);
```
instead of the current off-hea
kaivalnp commented on PR #14863:
URL: https://github.com/apache/lucene/pull/14863#issuecomment-3016573861
I ran some benchmarks on Cohere vectors (768d) for 7-bit and 4-bit
(compressed) quantization..
`main` without `jdk.incubator.vector`:
```
recall latency(ms) netCPU avgCpu
github-actions[bot] commented on PR #14863:
URL: https://github.com/apache/lucene/pull/14863#issuecomment-3016556401
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
kaivalnp opened a new pull request, #14863:
URL: https://github.com/apache/lucene/pull/14863
### Description
Off-heap scoring for quantized vectors! Related to #13515
This scorer is in-line with
[`Lucene99MemorySegmentFlatVectorsScorer`](https://github.com/apache/lucene/blob/7
dungba88 commented on PR #14009:
URL: https://github.com/apache/lucene/pull/14009#issuecomment-3016409630
I put a backport PR to 10.3 here: https://github.com/apache/lucene/pull/14860
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to G
28 matches
Mail list logo
