ChrisHegarty commented on issue #12302:
URL: https://github.com/apache/lucene/issues/12302#issuecomment-1554223493
++ to all of what @uschindler and @rmuir said relating to which JDK versions
we add support for. Specifically, let's start with JDK 20 **only**. After which
we can prepare for,
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1198703579
##
gradle/testing/defaults-tests.gradle:
##
@@ -122,7 +122,7 @@ allprojects {
// Lucene needs to optional modules at runtime, which we want to enfo
ChrisHegarty commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554440332
I refactored the provider and impl's:
1. So as to separate them out from VectorUtil - this should improve
readability, etc, as we move beyond dotProduct.
2. I also moved them
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1198869676
##
gradle/testing/defaults-tests.gradle:
##
@@ -119,10 +119,13 @@ allprojects {
if (rootProject.runtimeJavaVersion < JavaVersion.VERSION_16) {
jvm
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1198879166
##
gradle/testing/defaults-tests.gradle:
##
@@ -119,10 +119,13 @@ allprojects {
if (rootProject.runtimeJavaVersion < JavaVersion.VERSION_16) {
jvm
contrebande-labs commented on issue #12302:
URL: https://github.com/apache/lucene/issues/12302#issuecomment-1554466920
Guys, relax. My intention is to _**HELP**_. I was trying to find something
that can be done in parallel to what @ChrisHegarty is doing so I don't step on
his toes. If you w
rmuir commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1198991426
##
gradle/testing/defaults-tests.gradle:
##
@@ -119,10 +119,13 @@ allprojects {
if (rootProject.runtimeJavaVersion < JavaVersion.VERSION_16) {
jvmArgs '-
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199003878
##
gradle/testing/defaults-tests.gradle:
##
@@ -119,10 +119,13 @@ allprojects {
if (rootProject.runtimeJavaVersion < JavaVersion.VERSION_16) {
jvmAr
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199008182
##
gradle/testing/defaults-tests.gradle:
##
@@ -122,7 +122,7 @@ allprojects {
// Lucene needs to optional modules at runtime, which we want to enforc
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199035703
##
gradle/testing/defaults-tests.gradle:
##
@@ -119,10 +119,13 @@ allprojects {
if (rootProject.runtimeJavaVersion < JavaVersion.VERSION_16) {
jvm
uschindler merged PR #12308:
URL: https://github.com/apache/lucene/pull/12308
--
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.
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199048935
##
lucene/core/src/java/org/apache/lucene/internal/vector/DefaultVectorUtilProvider.java:
##
@@ -0,0 +1,92 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199053205
##
lucene/core/src/java/org/apache/lucene/internal/vector/VectorUtilProvider.java:
##
@@ -76,4 +77,10 @@ static boolean vectorModulePresentAndReadable() {
}
tang-hi commented on PR #12255:
URL: https://github.com/apache/lucene/pull/12255#issuecomment-1554708934
@msokolov, thank you! I have successfully run the test and it confirms what
I mentioned earlier. I believe that #12303 by @jbellis could resolve this issue.
**baseline** (this pull
uschindler commented on code in PR #12290:
URL: https://github.com/apache/lucene/pull/12290#discussion_r1199073297
##
lucene/core/src/java/org/apache/lucene/store/ByteBufferGuard.java:
##
@@ -65,14 +62,8 @@ public ByteBufferGuard(String resourceDescription,
BufferCleaner cleane
uschindler commented on PR #12294:
URL: https://github.com/apache/lucene/pull/12294#issuecomment-1554737939
I noticed that the apijar files are quite large, because the extraction code
can't remove package private superclasses. Therefore all package private
classes stay alive as "empty" fra
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554738661
I noticed that the apijar files are quite large, because the extraction code
can't remove package private superclasses. Therefore all package private
classes stay alive as "empty" fra
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554742703
In addition, as we do not implement java 19 vector support yet, I would add
some code to don't extract it dependning on java version. So we can control
separately which of the 2 api a
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554743815
In addition at some point we should rename the files, but thats not urgent
because naming is not so important. We should then also rename the extraction
gradle script, as it will be u
alessandrobenedetti opened a new issue, #12313:
URL: https://github.com/apache/lucene/issues/12313
### Description
It would be nice to support multiple values in a Knn vector field.
This must be compatible with both the Exact and Approximate Nearest Neighbor
search.
There ar
alessandrobenedetti opened a new pull request, #12314:
URL: https://github.com/apache/lucene/pull/12314
### Description
This pull request aims to introduce support for multiple values in a single
Knn vector field.
The adopted solution relies on:
**Index time**
Sparse vector value
tang-hi commented on PR #12314:
URL: https://github.com/apache/lucene/pull/12314#issuecomment-1554789228
This concept is interesting, but I am curious about its practical uses.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub a
uschindler closed issue #12304: VirtualMethod does unprivileged reflection
access
URL: https://github.com/apache/lucene/issues/12304
--
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 com
rmuir commented on code in PR #12312:
URL: https://github.com/apache/lucene/pull/12312#discussion_r1199137356
##
lucene/core/src/java/org/apache/lucene/util/automaton/DaciukMihovAutomatonBuilder.java:
##
@@ -308,17 +290,83 @@ private void replaceOrRegister(State state) {
}
rmuir commented on code in PR #12312:
URL: https://github.com/apache/lucene/pull/12312#discussion_r1199139773
##
lucene/core/src/java/org/apache/lucene/util/automaton/DaciukMihovAutomatonBuilder.java:
##
@@ -308,17 +290,83 @@ private void replaceOrRegister(State state) {
}
rmuir commented on PR #12312:
URL: https://github.com/apache/lucene/pull/12312#issuecomment-1554807614
thanks for getting this started! Will be interested to see how the use of
`Terms.intersect` impacts the performance.
--
This is an automated message from the Apache Git Service.
To respo
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199144041
##
lucene/core/src/java/org/apache/lucene/internal/vector/DefaultVectorUtilProvider.java:
##
@@ -0,0 +1,92 @@
+/*
+ * Licensed to the Apache Software Foundation (
alessandrobenedetti commented on PR #12314:
URL: https://github.com/apache/lucene/pull/12314#issuecomment-1554825253
I'll follow up with many clean up and tidy up on my own in the next few
weeks.
I should have a bit of bandwidth from now till Berlin Buzzword (mid June).
Any feedback is
uschindler opened a new pull request, #12315:
URL: https://github.com/apache/lucene/pull/12315
this PR just fixes the apijar generator to reproduce the file exactly
(unfortunately java encodes the date using local timezone).
--
This is an automated message from the Apache Git Service.
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554866016
... I had to first the reason why my computer produced a different APIJAR
from beginning I HATE DEFAULT TIMEZONE, DIE; DIE; DIE
--
This is an automated message from the Apache G
uschindler merged PR #12315:
URL: https://github.com/apache/lucene/pull/12315
--
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.
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1554917991
I fixed it and merged the main branch into this one. Proceeding with fixing
API generator to exclude unreferenced, private classes
--
This is an automated message from the Apache Gi
donnerpeter opened a new pull request, #12316:
URL: https://github.com/apache/lucene/pull/12316
IntelliJ's allocation profiler shows some non-zero numbers from stream
allocation and traversal. While JIT might be able to eliminate this sometimes,
I'd prefer to avoid the doubt completely.
-
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199198237
##
lucene/core/src/java/org/apache/lucene/internal/vector/VectorUtilProvider.java:
##
@@ -76,4 +77,10 @@ static boolean vectorModulePresentAndReadable() {
}
benwtrent commented on PR #12314:
URL: https://github.com/apache/lucene/pull/12314#issuecomment-1554999300
@alessandrobenedetti thank you for kick starting this!
You are absolutely correct, this is a large, but pivotal and necessary
change for vector search in Lucene. I have not yet r
uschindler commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199048935
##
lucene/core/src/java/org/apache/lucene/internal/vector/DefaultVectorUtilProvider.java:
##
@@ -0,0 +1,92 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199236929
##
lucene/core/src/java/org/apache/lucene/internal/vector/VectorUtilProvider.java:
##
@@ -76,4 +77,10 @@ static boolean vectorModulePresentAndReadable() {
}
ChrisHegarty commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1555062236
> In addition at some point we should rename the files, but thats not urgent
because naming is not so important. We should then also rename the extraction
gradle script, as it will
ChrisHegarty commented on code in PR #12311:
URL: https://github.com/apache/lucene/pull/12311#discussion_r1199250652
##
lucene/core/src/java/org/apache/lucene/internal/vector/DefaultVectorUtilProvider.java:
##
@@ -0,0 +1,92 @@
+/*
+ * Licensed to the Apache Software Foundation (
dweiss commented on code in PR #12316:
URL: https://github.com/apache/lucene/pull/12316#discussion_r1199252099
##
lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Dictionary.java:
##
@@ -155,7 +155,7 @@ public class Dictionary {
boolean checkCompoundCase, c
donnerpeter commented on code in PR #12316:
URL: https://github.com/apache/lucene/pull/12316#discussion_r1199286526
##
lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Dictionary.java:
##
@@ -155,7 +155,7 @@ public class Dictionary {
boolean checkCompoundCa
dweiss commented on code in PR #12316:
URL: https://github.com/apache/lucene/pull/12316#discussion_r1199299390
##
lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/Dictionary.java:
##
@@ -155,7 +155,7 @@ public class Dictionary {
boolean checkCompoundCase, c
donnerpeter merged PR #12316:
URL: https://github.com/apache/lucene/pull/12316
--
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
RS146BIJAY commented on issue #12228:
URL: https://github.com/apache/lucene/issues/12228#issuecomment-1555145366
As of now is there is any way User can delete these unreferenced files on
their end?
--
This is an automated message from the Apache Git Service.
To respond to the message, ple
RS146BIJAY commented on issue #12228:
URL: https://github.com/apache/lucene/issues/12228#issuecomment-1555155442
Also Lucene provides a way to rollback to previous commit using
```rollback``` function call. But as of now, it also closes the IndexWriter as
well. I think ```close``` function
gsmiller commented on PR #12312:
URL: https://github.com/apache/lucene/pull/12312#issuecomment-1555164152
Here's what I'm seeing so far in benchmarking...
I took a custom benchmarking approach for this, similar to #12151 and other
related issues. I did this because, 1) we don't really
rmuir commented on PR #12312:
URL: https://github.com/apache/lucene/pull/12312#issuecomment-1555172675
hmm, disappointing. Was hoping to see gains on the terms dictionary since it
optimizes `intersect`. wonder what is going on.
Of course docvalues impl doesn't optimize `intersect` in
uschindler commented on PR #12311:
URL: https://github.com/apache/lucene/pull/12311#issuecomment-1555188725
> > In addition at some point we should rename the files, but thats not
urgent because naming is not so important. We should then also rename the
extraction gradle script, as it will
gsmiller commented on PR #12312:
URL: https://github.com/apache/lucene/pull/12312#issuecomment-1555244204
OK, here's a method profiler diff for the "High Cardinality PK" task,
comparing two postings approaches—one that's using the current MultiTermQuery
version, and one using AutomatonQuery
gsmiller commented on PR #12312:
URL: https://github.com/apache/lucene/pull/12312#issuecomment-1555337627
Hmm... not sure if I've got something setup incorrectly with my JFR
settings, but trying to dig into the other tasks, I can't even get the relevant
methods to show up in the profiled ca
msokolov commented on PR #12255:
URL: https://github.com/apache/lucene/pull/12255#issuecomment-1555366279
Thanks everyone for testing and fixing. I had reverted this yesterday and I
believe what we have on main now has recovered the performance we had before. I
also ran luceneutil a few tim
jbellis commented on PR #12255:
URL: https://github.com/apache/lucene/pull/12255#issuecomment-1555369584
The performance impact to building is more meaningful because that is where
you are allocating large queues for multiple levels
--
This is an automated message from the Apache Git Serv
jainankitk opened a new issue, #12317:
URL: https://github.com/apache/lucene/issues/12317
### Description
While working on a customer issue, I noticed that memory allocations for
recently added [term dictionary
compression](https://github.com/apache/lucene-solr/commit/33a7af9cbfb9f66
mikemccand commented on PR #12310:
URL: https://github.com/apache/lucene/pull/12310#issuecomment-1555414840
> LGTM. Apologies for the merge conflict I created for you (but thanks for
the review on that PR!).
No worries! I'll resolve and merge soon! Thanks for the review @gsmiller.
mikemccand commented on PR #12310:
URL: https://github.com/apache/lucene/pull/12310#issuecomment-1555415525
> Yeah we should explore a binary version. Even if it doesn't speedup
TermInSetQuery.
+1
> Pretty sure I added a comment along the lines of "we should not do this"
mikemccand commented on PR #12310:
URL: https://github.com/apache/lucene/pull/12310#issuecomment-1555416664
Actually, the terms must be sorted in Unicode code point order, and, we do
have the builder for `BytesRef` already: `public static Automaton
build(Collection input) {`. So I think we
hydrogen666 commented on PR #816:
URL: https://github.com/apache/lucene/pull/816#issuecomment-1555456315
Any progress of this PR?
@uschindler
--
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 t
57 matches
Mail list logo
