[
https://issues.apache.org/jira/browse/TIKA-3699?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17510828#comment-17510828
]
Tim Allison commented on TIKA-3699:
-----------------------------------
All looks good. We've improved tika-ooxml detection on broken files so that
some of these are now parsed as docx or xlsx so we're getting more content.
There are a few new exceptions that I'll look into (likely on the Tika side or
on malformed files that we were skipping before), but nothing problematic.
> Upgrade to POI 5.2.2
> --------------------
>
> Key: TIKA-3699
> URL: https://issues.apache.org/jira/browse/TIKA-3699
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Blocker
> Fix For: 1.28.2, 2.4.0
>
>
> [~ahubold] reported a serious memory bug in POI 5.2.1 over on TIKA-3690. We
> should revert to 5.2.0 or wait for 5.2.2 before our next release in both the
> 2.x and 1.x branches.
> An issue has not been opened, but the initial email is here:
> https://lists.apache.org/thread/fmb746gypgfpj8k0lmcvtn89zppwb95p
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
