dweiss commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758985802
Let me take a look.
--
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 comme
zhaih opened a new pull request, #12660:
URL: https://github.com/apache/lucene/pull/12660
### Description
This PR is still a draft contains all kinds of magic number and names. It
contains changes of #12651 so I would like to merge that one first. But I do
think the logic is correct
gf2121 opened a new issue, #12659:
URL: https://github.com/apache/lucene/issues/12659
### Description
I'm seeing a regressions of term-dict queries on nightly benchmark:
https://home.apache.org/~mikemccand/lucenebench/2023.10.10.18.03.55.html
--
This is an automated message from
msokolov commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1356041142
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -40,31 +41,29 @@ public final class OnHeapHnswGraph extends HnswGraph
implements Acco
zhaih commented on PR #12651:
URL: https://github.com/apache/lucene/pull/12651#issuecomment-1758887879
Yes that's the idea, although I actually made some mistakes here so the
merging is not entirely pre-allocated, also something in searching might be
broken due to the size() behavior ch
easyice opened a new pull request, #12658:
URL: https://github.com/apache/lucene/pull/12658
The `BKDWriter#docsSeen` takes up a bit large proportion in the flame graph
sometimes in my production environment. but it is not always, maybe we can use
the docCount has computed in `PointValuesWr
msokolov commented on PR #12651:
URL: https://github.com/apache/lucene/pull/12651#issuecomment-1758885797
I like this! Actually I think when we are merging we can preallocate the
entire array so we don't need to resize at all which should greatly simplify
making this beast thread-safe (sinc
gf2121 closed pull request #12656: Larger default block size for block tree
index
URL: https://github.com/apache/lucene/pull/12656
--
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 commen
gf2121 commented on PR #12656:
URL: https://github.com/apache/lucene/pull/12656#issuecomment-1758866543
The improvement seem not stable, another run:
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct diff p-value
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758596094
It has to be the interaction with
```
// Any javac compilation tasks
tasks.withType(JavaCompile) {
dependsOn ":altJvmWarning"
options.fork = tru
uschindler commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758564752
Hrm
Let's see how this affects us us.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758554468
@dweiss already filed https://github.com/gradle/gradle/issues/22746 and it
looks like somehow the workaround in `gradle/hacks/turbocharge-jvm-opts.gradle`
is broken in Gradle 8.4 w/
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758551645
The cause is `gradle/hacks/turbocharge-jvm-opts.gradle`
If I comment out `apply from:
file('gradle/hacks/turbocharge-jvm-opts.gradle')` in top level `build.gradle`
things wo
zhaih commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1355784178
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,226 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF)
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758547290
snippet from:
```
RUNTIME_JAVA_HOME=/Library/Java/JavaVirtualMachines/openjdk.jdk/Contents/Home
./gradlew check -x test -Pvalidation.git.failOnModified=false --info
```
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758537719
I locally reverted gradlew changes and that didn't help fix anything. So it
looks like its something in the newer gradle version and how
`RUNTIME_JAVA_HOME` is being handled :/
--
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758523181
`RUNTIME_JAVA_HOME=/Library/Java/JavaVirtualMachines/openjdk.jdk/Contents/Home
./gradlew check -x test` reproduced the same exception for me locally.
but for
```
>
uschindler commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758519245
I have no idea why those are injected only for compilation of those tests.
Core test compilation works fine. So there seems to be some difference on the
distribution tests.
uschindler commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758517444
Jenkins uses: RUNTIME_JAVA_HOME and TEST_ARGS environment variables.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Git
stefanvodita commented on code in PR #12625:
URL: https://github.com/apache/lucene/pull/12625#discussion_r1355758154
##
lucene/core/src/java/org/apache/lucene/index/TermsHashPerField.java:
##
@@ -255,6 +255,81 @@ final void writeBytes(int stream, byte[] b, int offset,
int len)
uschindler commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758514194
See this build. Gradle command line and env cars can be found there:
https://jenkins.thetaphi.de/job/Lucene-main-Linux/44942/console
--
This is an automated message from the Ap
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758513642
Hmmm I checked this with `./gradlew check` on both JDK 17 and JDK 21
locally.
So it must be something w/ Gradle property being sent in :( - do you have a
Jenkins job url? I
uschindler commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758510093
There seems to be an issue with compiling the distribution tests:
```
> Task :lucene:core.tests:compileJava FAILED
error: invalid flag: -XX:+UseParallelGC
Usage: ja
risdenk closed issue #12655: Upgrade to Gradle 8.4
URL: https://github.com/apache/lucene/issues/12655
--
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: is
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758484315
Merged to branch_9x as:
* d1f73214a3c0fbc0f9cb783c8a9ab11ffa73ea36
* b3a9375f2793f3d414bbcfdd659a8a6da485d9a8
* e59c607daf98bd0e0faf259ac9c9cf2e3cff5807
* aa968f96d6c9d80d
uschindler commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1758478781
Thanks! 👍
--
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 uns
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1758470856
Merged to main as:
* 30d3eba93314fbaa014e885a21d2aa1588433df5
* 2c42b8941aa3247b8e24451af1553993f6a95079
* de3b294be44c2cea51a5e909db05043aa8d15150
* a8fba38f1691199e0813d
risdenk closed pull request #12650: Gradle 8.4
URL: https://github.com/apache/lucene/pull/12650
--
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-u
risdenk commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1758470625
Merged to main as:
* 30d3eba93314fbaa014e885a21d2aa1588433df5
* 2c42b8941aa3247b8e24451af1553993f6a95079
* de3b294be44c2cea51a5e909db05043aa8d15150
* a8fba38f1691199e0813d71783
benwtrent commented on PR #12582:
URL: https://github.com/apache/lucene/pull/12582#issuecomment-1758391298
I realize this is a large PR. I am happy to go into more details in anything
that doesn't make sense (probably indicates the code isn't very readable or
there should be better comments
benwtrent opened a new pull request, #12657:
URL: https://github.com/apache/lucene/pull/12657
While working on the quantization codec & thinking about how merging will
evolve, it became clearer that having merging attached directly to the vector
writer is weird.
I extracted it out to
gf2121 opened a new pull request, #12656:
URL: https://github.com/apache/lucene/pull/12656
I tried to adjust the block size to [min:35, max:68] and run perf tasks on
`wikimediumall`. Surprisingly, I see almost every term-related task get faster,
including PKLookup, which expected to be slow
risdenk commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757986422
Agreed @uschindler I created https://github.com/apache/lucene/issues/12655
so I can prefix each commit w/ the issue identifier.
--
This is an automated message from the Apache Git Ser
risdenk commented on issue #12655:
URL: https://github.com/apache/lucene/issues/12655#issuecomment-1757985173
Handled by https://github.com/apache/lucene/pull/12650
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
U
benwtrent commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1355224614
##
lucene/core/src/java/org/apache/lucene/util/ScalarQuantizer.java:
##
@@ -0,0 +1,251 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
benwtrent commented on issue #12654:
URL: https://github.com/apache/lucene/issues/12654#issuecomment-1757926532
```
final int numIterations = TEST_NIGHTLY ? atLeast(100) : atLeast(5);
```
Is one such place where `nightly` matters from what I can tell.
I verified the `139`
benwtrent commented on issue #12654:
URL: https://github.com/apache/lucene/issues/12654#issuecomment-1757910972
Added some println statements to
```
public void eval(MockDirectoryWrapper dir) throws IOException {
if (r.nextInt(3000) == 0) {
if (callStackC
benwtrent opened a new issue, #12654:
URL: https://github.com/apache/lucene/issues/12654
### Description
CI indicated the test suite timed out. So, I ran the reproduction line
locally and had to kill the test running after 5 minutes.
I seriously doubt this test should take long
benwtrent commented on issue #12648:
URL: https://github.com/apache/lucene/issues/12648#issuecomment-1757820922
The replication has a `min` segment size of `695` in bytes. Running the test
multiple times shows that it generally has a higher segment size of at least
`699` when the test passe
risdenk merged PR #2680:
URL: https://github.com/apache/lucene-solr/pull/2680
--
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.
easyice commented on PR #12557:
URL: https://github.com/apache/lucene/pull/12557#issuecomment-1757655210
Thanks @jpountz , it isn't perfect, there seems to be no better way to avoid
probing data and know early on that all values are the same, I am just trying
to improve, but i haven't thou
risdenk commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757647698
I pushed changes based on @dweiss suggestion of
> separate all build-related changes from the google format upgrade. Then
apply build changes (one patch), apply google format upgra
risdenk commented on code in PR #12650:
URL: https://github.com/apache/lucene/pull/12650#discussion_r1354938574
##
gradlew:
##
@@ -1,7 +1,7 @@
-#!/usr/bin/env sh
Review Comment:
addressed this in the reformatted changes -
dcb33e6b1842d22ca3d10b0fb6ed269a66c67828
--
This
benwtrent commented on issue #12418:
URL: https://github.com/apache/lucene/issues/12418#issuecomment-1757640051
I verified on all three of the failing seeds, setting
`.setMergePolicy(NoMergePolicy.INSTANCE);` passes the test.
I don't know what that means and if we should consider that
risdenk commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757618130
Agree with cleaning up the commit history. I'll do that shortly.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use
benwtrent commented on issue #12601:
URL: https://github.com/apache/lucene/issues/12601#issuecomment-1757601139
closing as duplicate to https://github.com/apache/lucene/issues/12418
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Git
benwtrent closed issue #12601: Reproducible TestDrillSideways failure
URL: https://github.com/apache/lucene/issues/12601
--
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 unsu
benwtrent commented on issue #12418:
URL: https://github.com/apache/lucene/issues/12418#issuecomment-1757600283
Another seed to reproduce:
```
./gradlew test --tests TestDrillSideways.testRandom
-Dtests.seed=668BF9B25DA9EE8A
```
If I change the indexwriting config to:
uschindler commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757465742
> Oh, impressive work. I'm surprised it's worked because I know there are
incompatibilities between palantir's consistent version plugin and newer
spotless releases.
>
> Overal
dungba88 commented on code in PR #12624:
URL: https://github.com/apache/lucene/pull/12624#discussion_r1354640682
##
lucene/core/src/java/org/apache/lucene/util/fst/BytesStore.java:
##
@@ -21,19 +21,18 @@
import java.util.List;
import org.apache.lucene.store.DataInput;
import
dweiss commented on code in PR #12650:
URL: https://github.com/apache/lucene/pull/12650#discussion_r1354635205
##
gradlew:
##
@@ -1,7 +1,7 @@
-#!/usr/bin/env sh
Review Comment:
this was intentional (/usr/bin/env).
--
This is an automated message from the Apache Git Servi
dweiss commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757346355
Just to be clear - the above is not _strictly_ necessary but I think it'd
make for a much saner commit history and ability to review what you did to the
build files, otherwise it gets obs
dweiss commented on PR #12650:
URL: https://github.com/apache/lucene/pull/12650#issuecomment-1757307273
Oh, impressive work. I'm surprised it's worked because I know there are
incompatibilities between palantir's consistent version plugin and newer
spotless releases.
Overall, it look
shubhamvishu commented on code in PR #12653:
URL: https://github.com/apache/lucene/pull/12653#discussion_r1354595316
##
lucene/core/src/java/org/apache/lucene/codecs/MultiLevelSkipListWriter.java:
##
@@ -63,6 +63,9 @@ public abstract class MultiLevelSkipListWriter {
/** for e
gf2121 opened a new pull request, #12652:
URL: https://github.com/apache/lucene/pull/12652
In #12623, we introduced a `MergeSorter` to take advantage of extra memory
to speed up sorting.
This PR enables `StringStableSorter` to also benefit from this optimization.
--
This is an automate
gf2121 commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1757126242
Get it ! :)
[profile.log](https://github.com/apache/lucene/files/12866842/profile.log)
--
This is an automated message from the Apache Git Service.
To respond to the message,
vsop-479 commented on PR #12528:
URL: https://github.com/apache/lucene/pull/12528#issuecomment-1757016549
@iverase
Here is a performance data of geo cases. There are some slowdown due to the
extra check.
query | metric | baseline | candidate | Diff
-- | -- | -- | -- | --
pol
57 matches
Mail list logo
