[jira] [Comment Edited] (TIKA-3730) New ExternalParser doesn't work on Windows

["Tim Allison (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TIKA-3730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17528315#comment-17528315 ] 

Tim Allison edited comment on TIKA-3730 at 4/26/22 5:21 PM:
------------------------------------------------------------

Thank you, [~tilman]!  I've gotten mostly clean builds on my Windows laptop.  Confirmation is welcomed!

What I can't figure out is that I got one failed build because maven's resources plugin was complaining that droste.zip was a malicious file, but then I -rf from the pkg module, and it worked.  Who knows...


was (Author: tallison@mitre.org):
Thank you, [~tilman]!  I've gotten mostly clean builds on my Windows laptop.  Confirmation is welcomed!

What I can't figure out is that I got a failed build because maven's resources plugin was complaining that droste.zip was a malicious file, but then I -rf from the pkg module, and it worked.  Who knows...

> New ExternalParser doesn't work on Windows
> ------------------------------------------
>
>                 Key: TIKA-3730
>                 URL: https://issues.apache.org/jira/browse/TIKA-3730
>             Project: Tika
>          Issue Type: Task
>            Reporter: Tim Allison
>            Priority: Trivial
>             Fix For: 2.4.0
>
>
> [~tilman] noted that the external2.ExternalParser uses "replaceAll" on a regex where the replacement is a file path does not work on Windows.  The replaceAll strips the file separators.  I admit that I cannot figure out why this is is happening.  I've tried a couple of combinations of backslashing etc, but nothing is working.  I even tried Pattern.quote() and that doesn't work on Windows. 
> If we back off to use "replace" with a string, everything seems to work.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)