[
https://issues.apache.org/jira/browse/TIKA-391?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12878579#action_12878579
]
Jean-Philippe Ricard commented on TIKA-391:
-------------------------------------------
I had a similar problem when using the detector with a Microsoft Word document
(.doc). Using the detector multiple times on the same file returns different
results. Sometimes it is application/msword and sometimes
application/x-tika-msoffice.
The problem lies in the compareTo() method of the class Magic. The compareTo()
method is broken and therefore the order of Magic instances in the SortedSet of
the class MimeTypes is also broken which leads to inconsistent mime type
detection.
The comparison algorithm is using the toString() method of a Magic in the case
that the priority and the size are the same. Actually the toString() of a Magic
returns something like "50/org.apache.tika.detect.magicdetec...@732b3d53".
Since the toString() of MagicDetector in this case is not redefined, the
comparison of Magic returns random results. Instead of relying on the
toString() in the compareTo() of a Magic, using the type (getType()) of a Magic
would ensure more consistent results.
Jean-Philippe
> Intermittent errors detecting xls files
> ---------------------------------------
>
> Key: TIKA-391
> URL: https://issues.apache.org/jira/browse/TIKA-391
> Project: Tika
> Issue Type: Bug
> Components: mime
> Affects Versions: 0.6
> Reporter: Simon Tyler
> Assignee: Chris A. Mattmann
> Fix For: 0.8
>
> Attachments: MimeTypes.java
>
>
> I am doing some testing of Tika 0.6 and noticed some odd results for the
> testEXCEL.xls file included in the test suite.
> 100 calls to the following code:
>
> is = new BufferedInputStream(new FileInputStream(filename));
>
> Metadata metadata = new Metadata();
> metadata.set(Metadata.RESOURCE_NAME_KEY, filename);
>
> String type = tika.detect(is, metadata);
>
> Results in different matches as application/msword or
> application/vnd.ms-excel seemingly at random.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
