[jira] [Commented] (TIKA-1898) backslashes in mime-type : application/vnd.mif are wrong

[Nick Burch (JIRA)]

    [ 
https://issues.apache.org/jira/browse/TIKA-1898?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15209418#comment-15209418
 ] 

Nick Burch commented on TIKA-1898:
----------------------------------

Ah, ok, got it. We had a random impenetrable hex string magic too, which turned 
out to be "<MIFFile " which is what was working for the `test.mif` file. 
(Another reason to try to avoid hex-encoded magic when plain strings can be 
used)

I've fixed this in c94236a along with a unit test, based on your previously 
problematic file. Thanks for you help with this!

> backslashes in mime-type : application/vnd.mif are wrong 
> ---------------------------------------------------------
>
>                 Key: TIKA-1898
>                 URL: https://issues.apache.org/jira/browse/TIKA-1898
>             Project: Tika
>          Issue Type: Bug
>          Components: config, core
>         Environment: Win64, Eclipse
>            Reporter: Steffen Netz
>            Priority: Minor
>              Labels: easyfix, patch
>             Fix For: 1.13
>
>         Attachments: test.doc, test.fm, test.mif, tika-bug.log
>
>
> In    
> tika-core\src\main\resources\org\apache\tika\mime\tika-mimetypes.xml  
> there are the lines:
>     <match value="\&lt;MakerFile" type="string" offset="0" />
>       <match value="\&lt;MIFFile" type="string" offset="0" />
>       <match value="\&lt;MakerDictionary" type="string" offset="0" />
>       <match value="\&lt;MakerScreenFont" type="string" offset="0" />
>       <match value="\&lt;MML" type="string" offset="0" />
>       <match value="\&lt;Book" type="string" offset="0" />
>       <match value="\&lt;Maker" type="string" offset="0" />
>   wrong.
> the backslashes must be removed.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)