[jira] Commented: (TIKA-422) Wrong charset conversion in some RTF documents.

["Leszek Piotrowicz (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TIKA-422?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12874025#action_12874025 ] 

Leszek Piotrowicz commented on TIKA-422:
----------------------------------------

Look RTFParser.jj at http://www.koders.com/noncode/fid69741D7970C9D0BA22243871D7E5682434A2A19E.aspx - seems to be Apache licensed. Also RTFPlainTextExtractor.java is Apache licensed but RTFParserDelegate.java is LGPL (a bit strange)??? However it is rather a simple interface which may be rewritten.

After some hacking I was able to try this parser. It correctly extracts Polish diactric characters. It also inserts extra spaces into text in table cells (see https://issues.apache.org/jira/browse/TIKA-392)
But it failed on some big rtf file claiming that it is syntactically incorrect. Anyway it seems much better for non-english texts.


> Wrong charset conversion in some RTF documents.
> -----------------------------------------------
>
>                 Key: TIKA-422
>                 URL: https://issues.apache.org/jira/browse/TIKA-422
>             Project: Tika
>          Issue Type: Bug
>          Components: parser
>    Affects Versions: 0.7
>            Reporter: Piotr B.
>         Attachments: test-windows-1250.rtf
>
>
> RTF parser uses javax.swing.text.rtf, but it sucks.
> It doesn't support '\ansicpg' tag (cite from RTF file format specification:
> "This keyword represents the default ANSI code page used to perform the Unicode to ANSI conversion when writing RTF text").
> Unfortunately Windows WordPad saves nonascii characters using \ansicpg instead of supported by javax.swing.text.rtf unicode characters.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.