[jira] [Commented] (TIKA-1305) New list processing changes appear to be causing RTFParser exception

["Tim Allison (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TIKA-1305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14013630#comment-14013630 ] 

Tim Allison commented on TIKA-1305:
-----------------------------------

The cause of the problem with the attached file is that there is a listoverride, but then there is no listlevel in the override table.  Note, too, that the listOverrideCount is 0 (the spec [http://msdn.microsoft.com/en-us/library/office/aa140301(v=office.10).aspx#rtfspec_14] says that it has to be 1 or 9).

When the parser hits listoverride, it resets listTableLevel to -1.  When the parser then hits levelnfc within the override table, it tries to set that value and the AIOOBE happens.

Some options:
1) Simple solution, add a range check before setting the param
{code}
if (listTableLevel > -1 && listTableLevel < currentList.numberType.length) {
    currentList.numberType[listTableLevel] = param;
}
{code}
2) More complex: try to keep track of whether or not this is a valid override table and/or if it actually contains any listlevel items.

[~mikemccand] and [~dkincaid], any preference or other recommendations?

> New list processing changes appear to be causing RTFParser exception
> --------------------------------------------------------------------
>
>                 Key: TIKA-1305
>                 URL: https://issues.apache.org/jira/browse/TIKA-1305
>             Project: Tika
>          Issue Type: Bug
>          Components: parser
>    Affects Versions: 1.6
>         Environment: Mac OSX 10.7.5
> Tika 1.6-SNAPSHOT
>            Reporter: Chris Bamford
>            Priority: Minor
>              Labels: newbie
>         Attachments: rtfparsererror_2.rtf
>
>
> Some RTFs cause RTFParser to throw a RuntimeException:
> Unexpected RuntimeException from org.apache.tika.parser.rtf.RTFParser@425e60f2
> When tracing in the debugger (surfaces in CompositeParser.parse() where it catches the RuntimeException, line 244 in my copy), the exception (e) is:
> java.lang.ArrayIndexOutOfBoundsException: -1
> A committer (Tim Allison) believes that it is being caused by recent list processing changes.



--
This message was sent by Atlassian JIRA
(v6.2#6252)