[
https://issues.apache.org/jira/browse/TIKA-2099?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15535154#comment-15535154
]
Robin Schimpf commented on TIKA-2099:
-------------------------------------
It seems like the ZipContainerDetector gets never called in
{{tika-parsers/src/test/java/org/apache/tika/mime/TestMimeTypes.java}}. I will
place the test in
{{tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java}}.
I felt save to remove the special handling because of the changes made in
COMPRESS-331.
> Tar files without magic bytes are sporadically detected as text
> ---------------------------------------------------------------
>
> Key: TIKA-2099
> URL: https://issues.apache.org/jira/browse/TIKA-2099
> Project: Tika
> Issue Type: Bug
> Affects Versions: 1.11
> Reporter: Robin Schimpf
>
> When a tar is created with 7 Zip 9.20 the magic bytes "ustar" are not added.
> Everything seems to work file if the tar contains Microsoft Office files. But
> when only text files are contained Tika sporadically recognices it as
> text/plain. It also seems to depend on the size of the first file in the tar.
> This has to be several KB big.
> The problem was found in version 1.11 and also exists in the latest
> 1.14-SNAPSHOT.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
