You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tika.apache.org by "Berry van Ginkel (JIRA)" <ji...@apache.org> on 2011/07/31 11:51:09 UTC
[jira] [Updated] (TIKA-687) Temporary file not removed after detection

     [ https://issues.apache.org/jira/browse/TIKA-687?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Berry van Ginkel updated TIKA-687:
----------------------------------

    Description: 
Temporary files created by Tika are not removed in the case the TikaInputStream has been created using a byte array or BufferedInputStream and using the ZipContainerDetector (in our case for Office 2007 documents).
The fix for bug TIKA-654 solves part of the problem (when using file as input) but when the byte array is being used, TikaInputStream will create a temp file (when getFile() is called). This file will be removed when close() is called, but in the ZipDetector a ZipFile is instantiated which also opens a stream to the same temp file. This stream is not closed and therefor the file can not be deleted when TikaInputStream.close() is called.

See attached patch for unittest and solution.


  was:
Temporary files created by Tika are not removed in the case the TikaInputStream has been created using a byte array or BufferedInputStream and using the ZipContainerDetector (in our case for Office 2007 documents).
The fix for bug TIKA-654 solves part of the problem (when using file as input) but when the byte array is being used, TikaInputStream will create a temp file (when getFile() is called). This file will be removed when close() is called, but in the ZipDetector a ZipFile is instantiated which also opens a stream to the same temp file. This stream is not closed and therefor the file can not be deleted when TikaInputStream.close() is called.



> Temporary file not removed after detection
> ------------------------------------------
>
>                 Key: TIKA-687
>                 URL: https://issues.apache.org/jira/browse/TIKA-687
>             Project: Tika
>          Issue Type: Bug
>          Components: parser
>    Affects Versions: 1.0
>         Environment: Windows
>            Reporter: Berry van Ginkel
>         Attachments: tika-temp-files.patch
>
>
> Temporary files created by Tika are not removed in the case the TikaInputStream has been created using a byte array or BufferedInputStream and using the ZipContainerDetector (in our case for Office 2007 documents).
> The fix for bug TIKA-654 solves part of the problem (when using file as input) but when the byte array is being used, TikaInputStream will create a temp file (when getFile() is called). This file will be removed when close() is called, but in the ZipDetector a ZipFile is instantiated which also opens a stream to the same temp file. This stream is not closed and therefor the file can not be deleted when TikaInputStream.close() is called.
> See attached patch for unittest and solution.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira