[
https://issues.apache.org/jira/browse/TIKA-4065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17731706#comment-17731706
]
Tilman Hausherr commented on TIKA-4065:
---------------------------------------
Thanks! If I understand correctly, they're using the same write-only buffer
accross threads but to write the read() data into it, instead of calling
skip(). I wonder why this brings this effect.
> Apparent multithreading bug in HWP5 parser (2)
> ----------------------------------------------
>
> Key: TIKA-4065
> URL: https://issues.apache.org/jira/browse/TIKA-4065
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 2.8.0
> Reporter: Tilman Hausherr
> Priority: Major
> Labels: commons-io
> Attachments: a.txt, b.txt
>
>
> TIKA-3092 is happening again? {{testMultiThreadedSkipFully}} fails when using
> commons-io 2.12.0 or 2.13.0 (we still use 2.11.0 for some reason). I have
> attached the two extracts of the failing test. The only difference is that
> "(TIKA-2776)" from A appears as "(TIKA-2776 " in B.
> It happens with java 8, 11 and 20.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
