Mailing-List: contact dev-help@felix.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@felix.apache.org
Date: Thu, 19 Nov 2015 11:35:10 +0000 (UTC)
From: "Rob Walker (JIRA)" <jira@apache.org>
To: dev@felix.apache.org
Message-ID: <JIRA.12914425.1447930360000.121761.1447932910990@Atlassian.JIRA>
In-Reply-To: <JIRA.12914425.1447930360000@Atlassian.JIRA>
References: <JIRA.12914425.1447930360000@Atlassian.JIRA>
 <JIRA.12914425.1447930360243@arcas>
Subject: [jira] [Comment Edited] (FELIX-5109) Wrong path matching for http
 and http whiteboard service
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/FELIX-5109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15013387#comment-15013387 ] 

Rob Walker edited comment on FELIX-5109 at 11/19/15 11:34 AM:
--------------------------------------------------------------

Just downloaded the latest OSGi v6 compendium spec in case this area changed.

According to table 102.1 on p50 the described behavior, whilst apparently at odds with servlet matching, is correct for an OSGi HttpService.

/foo should match /foo/something (or /fudd if you prefer the examples used in the spec).

There is no concept of standard /foo/* matching within OSGi as far as I can see. Maybe I missed it, but it's also unnecessary because /foo is considered a root path.

There's no issue with us having non-compliant extension services or bundles I guess, but requiring /foo/* would actually mean we are not conforming to the published OSGi HttpService. IMO that would not only be wrong, but would break existing code.

Unless I'm totally misunderstanding this report!


was (Author: walkerr):
Just downloaded the latest OSGi v6 compendium spec in case this area changed.

According to table 102.1 on p50 the described behavior, whilst apparently at odds with servlet matching, is correct for an OSGi HttpService.

/foo should match /foo/something (or /fudd if you prefer the examples used in the spec).

There's no issue with us having non-compliant extension services or bundles I guess, but requiring /foo/* would actually mean we are not conforming to the published OSGi HttpService. IMO that would not only be wrong, but would break existing code.

Unless I'm totally misunderstanding this report!

> Wrong path matching for http and http whiteboard service
> --------------------------------------------------------
>
>                 Key: FELIX-5109
>                 URL: https://issues.apache.org/jira/browse/FELIX-5109
>             Project: Felix
>          Issue Type: Bug
>          Components: HTTP Service
>    Affects Versions: http.base-3.0.0, http.base-3.0.2, http.jetty-3.1.0, http.jetty-3.1.2
>            Reporter: Carsten Ziegeler
>            Assignee: Carsten Ziegeler
>             Fix For: http.base-3.0.4, http.jetty-3.1.4
>
>
> It seems we have some bugs in the path matching which result from different handling between the http service and the servlet spec. 
> For the http service, an alias like "/foo" matches "/foo" and "/foo/*" - while for the servlet spec "/foo" only matches "/foo" - exact match.
> Right now, in both cases "/foo/*" matches.
> For the servlet spec a pattern like "/foo/*" matches "/foo/something" but also "/foo". Right now we don't handle the second case.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)