[
https://issues.apache.org/jira/browse/TIKA-4944?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18124764#comment-18124764
]
ASF GitHub Bot commented on TIKA-4944:
--------------------------------------
henry-lindeman-glean commented on PR #3278:
URL: https://github.com/apache/tika/pull/3278#issuecomment-6048651751
those all look reasonable.
ran more ai reviews on my own too. One thing it came up with I didn't take
since the fix would be in tika-core and seems like it's known and the failure
case is kinda pathological anyway:
```
Quadratic publish at the cap (security + performance, independently):
publish hands
100k values to Metadata.set(Property, String[]), whose limiter appends
one-by-one
with growing-array copies — ~5·10⁹ reference copies cumulative (~20 GB
churn) worst
case, executed inside finally where it can delay limit-exception
propagation. The cap
bounds it but doesn't make it cheap. Decide: lower cap, batch differently,
or accept
with a comment.
```
> Emit onenote page guids in properties
> -------------------------------------
>
> Key: TIKA-4944
> URL: https://issues.apache.org/jira/browse/TIKA-4944
> Project: Tika
> Issue Type: Improvement
> Components: parser
> Reporter: Henry Lindeman
> Priority: Minor
>
> Onenote section files contain page GUIDs which can be used to reconstruct a
> url pointing to a page in the MS SaaS Onenote application. Since MS hands you
> the documents section at a time the only other way to get to the page GUID is
> to ask MS to list the pages in the section, which we'd prefer not to do
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
