[
https://issues.apache.org/jira/browse/TIKA-1436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15214082#comment-15214082
]
Stefano Fornari commented on TIKA-1436:
---------------------------------------
sorry, it took much more than I expected... however, here is a new patch. I
realized I did not create the previous one correctly, and I have added some
more testing code.
I have done it on the master HEAD from git.
Regarding your concern here:
"I'm looking at the raw patch now (not applied), and I'm a bit concerned that
there is special handling for catching and swallowing a WriteLimitReached
within the PDFParser. I may be misunderstanding your proposal, but the nice
thing about the exception was that it put the burden/opportunity on the client
to handle it, and we didn't have to add catch blocks to every parser (this
point was already made by Jukka)."
There are two main reasons:
1. the limit is in the ContentHandler and the Parser is the client of such
functionality, which therefore should handle the condition
2. the condition is handled because expected: we want the parsing to be
successful in the case the limit is reached so that the so far read content can
be handled; but I am open to explore a different approach if anyone thinks a
better way.
> improvement to PDFParser
> ------------------------
>
> Key: TIKA-1436
> URL: https://issues.apache.org/jira/browse/TIKA-1436
> Project: Tika
> Issue Type: Improvement
> Components: parser
> Affects Versions: 1.6
> Reporter: Stefano Fornari
> Labels: parser, pdf
> Fix For: 1.13
>
> Attachments: ste-20140927.patch
>
>
> with regards to the thread "[PDFParser] - read limited number of characters"
> on Mar 29, I would like to propose the attached patch. I noticed that in Tika
> 1.6 there have been some work around a better handling of the
> WriteLimitReachedException condition, but I believe it could be even
> improved.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
