[
https://issues.apache.org/jira/browse/TIKA-2527?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16337531#comment-16337531
]
Andreas Meier commented on TIKA-2527:
-------------------------------------
I attached a patch to address the mentioned problems.
Furthermore I added three new MIMEType sections for application/x-lz4,
Image/x-tga and audio/x-caf.
The Image/x-tga part had to be placed in front of the application/x-123
mime-type recognition, because the starting bytes overlap in some cases.
The important part of the Image/x-tga recognition is the inner match that
searches for the trailing part
54 52 55 45 56 49 53 49 TRUEVISI
4F 4E 2D 58 46 49 4C 45 ON-XFILE
2E 00 ..
Is there an easier possibility to search for trailing magic-strings than using
a regex?
I thought that a simple regex might be to expensive to recognize Image/x-tga,
therefore I combined the recognition with the basic tga-recognition of the
linux magic file.
While testing tika.mimetypes.xml I noticed that I often thought that the
matching string already was correct, when the actual recognition was done by
the file-extension. Therefore I had to remove the fileextensions of my
testfiles to validate the matching parts.
To avoid this I suggest to create either a testcase that only takes care of the
matches without taking file-extensions into account or to delete the
fileextensions of testfiles to validate the matchings.
Is there a testcase that does this already?
If you have any questions or suggestions I would be glad to hear them.
> Typos in tika-mimetypes.xml
> ---------------------------
>
> Key: TIKA-2527
> URL: https://issues.apache.org/jira/browse/TIKA-2527
> Project: Tika
> Issue Type: Bug
> Components: core
> Affects Versions: 2.0, 1.16, 1.17, 1.18
> Environment: ALL
> Reporter: Andreas Meier
> Priority: Minor
> Attachments:
> fix-for-TIKA2527-contributed-by-AMeier-Fixed-adpcmmi.patch
>
>
> Are these mimetypes in tika-mimetypes.xml
> audio/x-adbcm instead audio/x-adpcm
> {code:xml} <mime-type type="audio/x-adbcm">{code}
> and
> audio/x-dec-adbcm instead audio/x-dec-adpcm
> {code:xml} <mime-type type="audio/x-dec-adbcm">{code}
> intended?
> Couldn't find these mimetypes.
> Regards
> Andreas
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
