[
https://issues.apache.org/jira/browse/TIKA-3849?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17607658#comment-17607658
]
Tim Allison edited comment on TIKA-3849 at 9/21/22 10:15 AM:
-------------------------------------------------------------
Ugh. Wrong issue in commit message. This should have been: TIKA-3859
was (Author: [email protected]):
Ugh. Wrong issue in commit message. This should have been:
https://twitter.com/MickyLindlar/status/1572452576162992131?s=20&t=DUokmpBZt9vtXgLWCIM3mQTIKA-3859.
> Throw UnsupportedFormatException or similar for really old mdb files
> --------------------------------------------------------------------
>
> Key: TIKA-3849
> URL: https://issues.apache.org/jira/browse/TIKA-3849
> Project: Tika
> Issue Type: Wish
> Reporter: Tim Allison
> Priority: Minor
> Fix For: 2.5.0
>
>
> In processing some digipres2022 bake-off files, I noticed that a number of
> mdb files triggered the same exception:
> {noformat}
> java.io.IOException: Unrecognized map type: 75
> at
> com.healthmarketscience.jackcess.impl.UsageMap.initHandler(UsageMap.java:150)
> at com.healthmarketscience.jackcess.impl.UsageMap.read(UsageMap.java:136)
> at
> com.healthmarketscience.jackcess.impl.PageChannel.initialize(PageChannel.java:118)
> at
> com.healthmarketscience.jackcess.impl.DatabaseImpl.<init>(DatabaseImpl.java:579)
> at
> com.healthmarketscience.jackcess.impl.DatabaseImpl.open(DatabaseImpl.java:440)
> at
> com.healthmarketscience.jackcess.DatabaseBuilder.open(DatabaseBuilder.java:267)
> at
> org.apache.tika.parser.microsoft.JackcessParser.parse(JackcessParser.java:94)
> {noformat}
> Googling this confirmed suspicions that these are pre-97 versions of access
> databases. We should improve our exception type/labeling for this type of
> exception.
> ref: https://sourceforge.net/p/jackcess/bugs/101/
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
