tika-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ray Gauss II (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (TIKA-1022) DWG Custom properties not extracted
Date Fri, 09 Nov 2012 22:15:12 GMT

    [ https://issues.apache.org/jira/browse/TIKA-1022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13494358#comment-13494358
] 

Ray Gauss II commented on TIKA-1022:
------------------------------------

Thanks Paolo.

I don't know a lot about these formats but I have a slightly different fix prepared which
includes a unit test and your test file.

As far as I can see from what Tika extracts, the file contains most of the phrases expected
by DWGParserTest (with the exception of the addition of "very lazy") but can you confirm that
there's no additional sensitive information contained in that file that should prevent it
from being part of Tika's source?
                
> DWG Custom properties not extracted
> -----------------------------------
>
>                 Key: TIKA-1022
>                 URL: https://issues.apache.org/jira/browse/TIKA-1022
>             Project: Tika
>          Issue Type: Bug
>          Components: metadata
>    Affects Versions: 1.0, 1.1, 1.2, 1.3
>            Reporter: Paolo Nacci
>            Assignee: Ray Gauss II
>              Labels: patch
>         Attachments: quick2010-tika-no-custom.dwg
>
>   Original Estimate: 1h
>  Remaining Estimate: 1h
>
> Based on some code I provided some time ago (Alfresco forum), Derek Hulley opened ALF-2262,
Nick Burch opened TIKA-413 issue and code has been committed to TIKA (0.8).
> With sample dwg provided TIKA (0.8 to 1.2) is correctly working but with attached file
returns no custom metadata (my original "C" returns correct custom metadata, dwg is "2010"
format).
> Tested tika-app.1.0.jar and tika-app.1.2.jar and tika 1.3 snapshot.
> All versions could be impacted by this bug. 
> I found failing code in skipToCustomProperties() of DWGParser.java, lines 320-321: 
> if(padding[0] == 0 && padding[1] == 0 &&
>   padding[2] == 0 && padding[3] == 0) {
> padding[0] byte is not always 0 (attached file has 0x2) and probably there is no need
to check those bytes.
> Index: DWGParser.java
> ===================================================================
> --- DWGParser.java	(revisione 1407024)
> +++ DWGParser.java	(copia locale)
> @@ -93,7 +93,7 @@
>       * How far to skip after the last standard property, before
>       *  we find any custom properties that might be there.
>       */
> -    private static final int CUSTOM_PROPERTIES_SKIP = 20; 
> +    private static final int CUSTOM_PROPERTIES_SKIP = 24; 
>  
>      public void parse(
>              InputStream stream, ContentHandler handler,
> @@ -317,13 +317,7 @@
>  
>      private int skipToCustomProperties(InputStream stream) 
>              throws IOException, TikaException {
> -       // There should be 4 zero bytes next
> -       byte[] padding = new byte[4];
> -       IOUtils.readFully(stream, padding);
> -       if(padding[0] == 0 && padding[1] == 0 &&
> -             padding[2] == 0 && padding[3] == 0) {
> -          // Looks hopeful, skip on
> -          padding = new byte[CUSTOM_PROPERTIES_SKIP];
> +          byte[] padding = new byte[CUSTOM_PROPERTIES_SKIP];
>            IOUtils.readFully(stream, padding);
>            
>            // We should now have the count
> @@ -337,10 +331,6 @@
>               // No properties / count is too high to trust
>               return 0;
>            }
> -       } else {
> -          // No padding. That probably means no custom props
> -          return 0;
> -       }
>      }
>  
>  }

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message