I should add that &lta> style links seem to be working correctly ... so there
are two different code paths being executed to determine the relative path.

-Doug


Doug Donohoe wrote:
> 
> I think I've discovered a bug with how wicket decides to output "../.." in
> front of images and css files when working with mounted pages.  In this
> case, I'm using mixed param encoding.
> 
> A working URL is :
> 
> http://localhost:8080/poker/search/doug
> 
> My CSS gets generated as:
> 
> <link href="../styles.css" rel="stylesheet" type="text/css">
> 
> But I search for something that needs to be URL-encoded, like a +, the URL
> becomes:
> 
> http://localhost:8080/poker/search/%2B
> 
> and the CSS link gets generated as:
> 
> <link href="../../../styles.css" rel="stylesheet" type="text/css">
> 
> It seems to happen with most non-ascii encodings.
> 
> I saw a similar error when there were null/empty param values and the URL
> generated was like "foo//bar"
> (http://issues.apache.org/jira/browse/WICKET-1597).
> 
> Note that all relative paths seem to be effected:  images, shortcut icon,
> etc.
> 
> MY QUESTION:  Can someone give me a pointer where to look for the code
> that calculates the relative paths?  I'd like to try and track this down.
> 
> Thanks,
> 
> -Doug
> 
> 
> 
> 
> 
> 

-- 
View this message in context: 
http://www.nabble.com/trying-to-track-down-a-relative-path-bug-tp17187699p17187755.html
Sent from the Wicket - Dev mailing list archive at Nabble.com.

Reply via email to