[jira] [Commented] (LUCENE-8191) merge TestRandomChains "brokenConstructors" list with "brokenOffsetsConstructors"

["Robert Muir (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-8191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16385202#comment-16385202 ] 

Robert Muir commented on LUCENE-8191:
-------------------------------------

Attached is a patch: I tested it with {{ant beast -Dtestcase=TestRandomChains -Dbeasts.iters=100}}, I plan to commit it soon.

It makes the test easier to understand, there is only one broken list. If the test finds a bug in a tokenstream, add it to the that list until the tokenstream is fixed.

It will also silence recent failures about offsets coming from indexwriter with the output "offsetsAreCorrect=false", again that list makes no sense anymore post- LUCENE-7626.

> merge TestRandomChains "brokenConstructors" list with "brokenOffsetsConstructors"
> ---------------------------------------------------------------------------------
>
>                 Key: LUCENE-8191
>                 URL: https://issues.apache.org/jira/browse/LUCENE-8191
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Robert Muir
>            Priority: Major
>         Attachments: LUCENE-8191.patch
>
>
> Now that indexwriter checks offsets (LUCENE-7626), there is no difference between the two: A tokenstream that has brokenoffsets will fail regardless, only in a harder-to-debug way (e.g. some low level exception from indexwriter).
> So I think we should just merge the two lists to reflect that: if it produces brokenOffsets, its broken. 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org