[jira] [Commented] (LUCENE-10579) fix smoketester backwards-check to not parse stdout

2022-05-20 Thread ASF subversion and git services (Jira)


[ 
https://issues.apache.org/jira/browse/LUCENE-10579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17540069#comment-17540069
 ] 

ASF subversion and git services commented on LUCENE-10579:
--

Commit 2090ac431843f689ca6f124723ef37c3b12a2826 in lucene's branch 
refs/heads/main from Robert Muir
[ https://gitbox.apache.org/repos/asf?p=lucene.git;h=2090ac43184 ]

LUCENE-10579: fix smoketester backwards-check to not parse stdout (#903)

This is very noisy, can contain gradle status updates, various other 
tests.verbose prints from other threads, you name it.

It causes the check to be flaky, and randomly "miss" seeing a test that 
executed.

Instead, let's look at the zip files. We can still preserve the essence of what 
the test wants to do, but without any flakiness.

> fix smoketester backwards-check to not parse stdout
> ---
>
> Key: LUCENE-10579
> URL: https://issues.apache.org/jira/browse/LUCENE-10579
> Project: Lucene - Core
>  Issue Type: Bug
>Reporter: Robert Muir
>Priority: Major
> Fix For: 9.3
>
> Attachments: backwards.log.gz
>
>  Time Spent: 0.5h
>  Remaining Estimate: 0h
>
> The smoketester parses the output of TestBackwardsCompatibility -verbose 
> looking for certain prints for each index release. 
> But I think this is a noisier channel than you might expect. I added a hack 
> to log the stuff its trying to parse... it is legit crazy. See attachment
> Let's rethink, maybe we should just examine the zip files?



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

-
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org



[jira] [Commented] (LUCENE-10579) fix smoketester backwards-check to not parse stdout

2022-05-20 Thread ASF subversion and git services (Jira)


[ 
https://issues.apache.org/jira/browse/LUCENE-10579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17540070#comment-17540070
 ] 

ASF subversion and git services commented on LUCENE-10579:
--

Commit 5a5c8bca4e691d91801d04d881aa25560413835c in lucene's branch 
refs/heads/branch_9x from Robert Muir
[ https://gitbox.apache.org/repos/asf?p=lucene.git;h=5a5c8bca4e6 ]

LUCENE-10579: fix smoketester backwards-check to not parse stdout (#903)

This is very noisy, can contain gradle status updates, various other 
tests.verbose prints from other threads, you name it.

It causes the check to be flaky, and randomly "miss" seeing a test that 
executed.

Instead, let's look at the zip files. We can still preserve the essence of what 
the test wants to do, but without any flakiness.

> fix smoketester backwards-check to not parse stdout
> ---
>
> Key: LUCENE-10579
> URL: https://issues.apache.org/jira/browse/LUCENE-10579
> Project: Lucene - Core
>  Issue Type: Bug
>Reporter: Robert Muir
>Priority: Major
> Fix For: 9.3
>
> Attachments: backwards.log.gz
>
>  Time Spent: 0.5h
>  Remaining Estimate: 0h
>
> The smoketester parses the output of TestBackwardsCompatibility -verbose 
> looking for certain prints for each index release. 
> But I think this is a noisier channel than you might expect. I added a hack 
> to log the stuff its trying to parse... it is legit crazy. See attachment
> Let's rethink, maybe we should just examine the zip files?



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

-
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org



[jira] [Commented] (LUCENE-10579) fix smoketester backwards-check to not parse stdout

2022-05-18 Thread Robert Muir (Jira)


[ 
https://issues.apache.org/jira/browse/LUCENE-10579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17539050#comment-17539050
 ] 

Robert Muir commented on LUCENE-10579:
--

or even maybe a gradle status update with its escape characters and so on (it 
has a progress bar and such), seems like that could be enough to break the 
check.

> fix smoketester backwards-check to not parse stdout
> ---
>
> Key: LUCENE-10579
> URL: https://issues.apache.org/jira/browse/LUCENE-10579
> Project: Lucene - Core
>  Issue Type: Bug
>Reporter: Robert Muir
>Priority: Major
> Attachments: backwards.log.gz
>
>
> The smoketester parses the output of TestBackwardsCompatibility -verbose 
> looking for certain prints for each index release. 
> But I think this is a noisier channel than you might expect. I added a hack 
> to log the stuff its trying to parse... it is legit crazy. See attachment
> Let's rethink, maybe we should just examine the zip files?



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

-
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org



[jira] [Commented] (LUCENE-10579) fix smoketester backwards-check to not parse stdout

2022-05-18 Thread Robert Muir (Jira)


[ 
https://issues.apache.org/jira/browse/LUCENE-10579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17539042#comment-17539042
 ] 

Robert Muir commented on LUCENE-10579:
--

There's all kinds of stuff being printed, but this gives you an idea of what 
the 28.4MB looks like.

So I'm not surprised if this smoketester check fails here and there, its such a 
noisy channel. All it takes is something like MockRandomMergePolicy or some 
other component logging from another thread to prevent that multiline regexp 
from doing the right thing?

{noformat}
ESC[2AESC[1m 92% EXECUTING 
[26s]ESC[mESC[35DESC[1BESC[1m> 
:lucene:backward-codecs:testESC[mESC[30DESC[1BESC[2AESC[1m
 92% EXECUTING [27s]ESESC[mESC[35DESC[2BESC[1AESC[1m> 
:lucene:backward-codecs:test > 0 tests completedESC[mESC[50DESC[1B
ESC[3AESC[35CESC[0KESC[35DESC[2BESC[1m> :lucene:backward-codecs:test > 
Executing test 
org.apache.lucene.backward_indeESC[mESC[79DESC[1BESC[3AESC[1m
 92% EXECUTING [28s]ESC[mESC[35DESC[3BESC[3AESC[0K
ESC[1m> Task :lucene:backward-codecs:testESC[mESC[0K
  1> filesystem: 
ExtrasFS(HandleLimitFS(LeakFS(ShuffleFS(DisableFsyncFS(VerboseFS(sun.nio.fs.LinuxFileSystemProvider@7764d0d3))ESC[0K
  1> FS 0 [2022-05-18T19:37:29.645632Z; 
SUITE-TestBackwardsCompatibility-seed#[2EBBD700BDB7349D]-worker]: 
createDirectory: ../../../../../../../../lucene_gradle (FAILED: 
java.nio.file.FileAlreadyExistsException: /tmp/lucene_gradle)
  1> Loaded codecs: [Lucene92, Asserting, CheapBastard, 
DeflateWithPresetCompressingStoredFieldsData, FastCompressingStoredFieldsData, 
FastDecompressionCompressingStoredFieldsData, 
HighCompressionCompressingStoredFieldsData, 
LZ4WithPresetCompressingStoredFieldsData, DummyCompressingStoredFieldsData, 
SimpleText, Lucene80, Lucene84, Lucene86, Lucene87, Lucene70, Lucene90, 
Lucene91]
  1> Loaded postingsFormats: [Lucene90, MockRandom, RAMOnly, LuceneFixedGap, 
LuceneVarGapFixedInterval, LuceneVarGapDocFreqInterval, 
TestBloomFilteredLucenePostings, Asserting, UniformSplitRot13, 
STUniformSplitRot13, BlockTreeOrds, BloomFilter, Direct, FST50, UniformSplit, 
SharedTermsUniformSplit, Lucene50, Lucene84]
  1> FS 0 [2022-05-18T19:37:29.780830Z; 
SUITE-TestBackwardsCompatibility-seed#[2EBBD700BDB7349D]-worker]: 
createDirectory: 
../../../../../../../../lucene_gradle/lucene.backward_index.TestBackwardsCompatibility_2EBBD700BDB7349D-001
  1> FS 0 [2022-05-18T19:37:29.783274Z; 
SUITE-TestBackwardsCompatibility-seed#[2EBBD700BDB7349D]-worker]: 
createDirectory: 
../../../../../../../../lucene_gradle/lucene.backward_index.TestBackwardsCompatibility_2EBBD700BDB7349D-001/8.0.0-cfs-001
  1> FS 0 [2022-05-18T19:37:29.785704Z; 
SUITE-TestBackwardsCompatibility-seed#[2EBBD700BDB7349D]-worker]: 
createDirectory: 
../../../../../../../../lucene_gradle/lucene.backward_index.TestBackwardsCompatibility_2EBBD700BDB7349D-001/8.0.0-cfs-001
 (FAILED: java.nio.file.FileAlreadyExistsException: 
/tmp/lucene_gradle/lucene.backward_index.TestBackwardsCompatibility_2EBBD700BDB7349D-001/8.0.0-cfs-001)
  1> FS 0 [2022-05-18T19:37:29.789291Z; 
SUITE-TestBackwardsCompatibility-seed#[2EBBD700BDB7349D]-worker]: 
newOutputStream[]: 
../../../../../../../../lucene_gradle/lucene.backward_index.TestBackwardsCompatibility_2EBBD700BDB7349D-001/8.0.0-cfs-001/_0.cfe
{noformat}

> fix smoketester backwards-check to not parse stdout
> ---
>
> Key: LUCENE-10579
> URL: https://issues.apache.org/jira/browse/LUCENE-10579
> Project: Lucene - Core
>  Issue Type: Bug
>Reporter: Robert Muir
>Priority: Major
> Attachments: backwards.log.gz
>
>
> The smoketester parses the output of TestBackwardsCompatibility -verbose 
> looking for certain prints for each index release. 
> But I think this is a noisier channel than you might expect. I added a hack 
> to log the stuff its trying to parse... it is legit crazy. See attachment
> Let's rethink, maybe we should just examine the zip files?



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

-
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org



[jira] [Commented] (LUCENE-10579) fix smoketester backwards-check to not parse stdout

2022-05-18 Thread Robert Muir (Jira)


[ 
https://issues.apache.org/jira/browse/LUCENE-10579?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17539040#comment-17539040
 ] 

Robert Muir commented on LUCENE-10579:
--

I attached compressed file of what the smoketester is parsing with regexps 
today. I guarantee it is wilder than you would imagine looking at the code.

I simply added this patch to log it:
{noformat}
   stdout = stdout.decode('utf-8',errors='replace').replace('\r\n','\n')
+  with open('%s/backwards.log' % unpackPath, 'w') as logfile:
+logfile.write(stdout)
{noformat}

And now you can look at the 28.4MB of output that it parses.

> fix smoketester backwards-check to not parse stdout
> ---
>
> Key: LUCENE-10579
> URL: https://issues.apache.org/jira/browse/LUCENE-10579
> Project: Lucene - Core
>  Issue Type: Bug
>Reporter: Robert Muir
>Priority: Major
> Attachments: backwards.log.gz
>
>
> The smoketester parses the output of TestBackwardsCompatibility -verbose 
> looking for certain prints for each index release. 
> But I think this is a noisier channel than you might expect. I added a hack 
> to log the stuff its trying to parse... it is legit crazy. See attachment
> Let's rethink, maybe we should just examine the zip files?



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

-
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org