[
https://issues.apache.org/jira/browse/TIKA-581?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12983711#action_12983711
]
Jukka Zitting commented on TIKA-581:
------------------------------------
The HTML parsing problem is caused by double close tags incorrectly returned by
the tagsoup library. See
https://groups.google.com/forum/#!topic/tagsoup-friends/_fCiasXtB68 for
details. In revision 1060813 I added a workaround that allows Tika to continue
processing the document even in such cases.
The PDF parsing problem is caused by PDFBOX-578 and is already fixed in PDFBox
trunk. We'll get the fix in Tika as soon as PDFBox 1.5.0 is out.
> Parser fails on files that parsed with v0.7
> -------------------------------------------
>
> Key: TIKA-581
> URL: https://issues.apache.org/jira/browse/TIKA-581
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 0.8
> Environment: Win Pro 7, x64, jdk1.6.0_22, jre 6.0.220.4
> Reporter: Dennis Adler
> Attachments: Tika-581 files.zip
>
>
> I am including 3 files. Two of these (Employee Application.pdf; pertoc.html)
> parsed without error under Tika 0.7. The third (I-9.pdf) fails under both
> versions and is included more as another test case for PDF parse errors.
> Perltoc.html fails with a Null Pointer exception in
> org.apache.tika.parser.html.HtmlParser@53786b79. Employee Application.pdf
> fails in org.apache.tika.parser.pdf.PDFParser@39fb9fb3, also with a Null
> Pointer exception. Both cases repro with tikaapp.jar built from the 0.8
> release.
> I-9 (which also fails under 0.7, but for different reasons) also fails with a
> Null Pointer exception at the same location as Employee Application.pdf.
> Looks like I have to create the bug before I can attach files...
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
