[
https://issues.apache.org/jira/browse/JENA-662?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13946794#comment-13946794
]
Jeremy Carroll commented on JENA-662:
-------------------------------------
Re-reading the original issue, I may have misunderstood.
getDirectoryName should indeed do a % decode, and that is relatively easy
compared with the % encode.
The method is:
- replace every %[0-9A-Fa-f][0-9A-Fa-f] by the corresponding byte and do the
utf-8 thing
If it goes wrong, throw an exception.
It would be fairly straightforward to write test cases for this behavior. The
problem of what to encode when, lies elsewhere, and is harder
> Wrong space handling in FileModelAssembler
> ------------------------------------------
>
> Key: JENA-662
> URL: https://issues.apache.org/jira/browse/JENA-662
> Project: Apache Jena
> Issue Type: Bug
> Components: Jena
> Affects Versions: Jena 2.11.1
> Reporter: Natanael Arndt
> Priority: Minor
>
> The method {{getDirectoryName}} in
> {{com.hp.hpl.jena.assembler.assemblers.FileModelAssembler}} should also make
> sure that at some chars, like space should be urldecoded again. E.g. If the
> URI is {{file:///tmp/file%20name%20with%20space}} the corresponding file can
> not be found under {{/tmp/file%20name%20with%20space}} but under {{/tmp/file
> name with space}}.
> There might be some other chars which have to be taken into account, e.g.
> "%25" -> "%", but I don't think a urldecode over the whole string is correct.
> Sorry I can't provide an example yet, because I couldn't isolate this part of
> the code so far and I didn't try to write a test for this.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
