Re: invalid (?) relative url resolve issue +patch
Hi, On Wed, Nov 29, 2006 at 12:47:39AM +0100, Schilter Roland wrote:
Hello, first of all, thanks for dillo :)
;)
I think I found a bug or at least, a misbehaviour that is handled in famous browsers differently. I looked through the code and finally figured out where to change code to get the expected behaviour. (Took some time though :p )
I need to mention that I didn't read the RFC2396 about URIs and I have the feeling that the implementation in dillo may be according to specifications but nevertheless, i couldn't browse several sites... :)
The problem is, if a `href' attribute of html-tag `a' starts only with a query (`?test') then the path of the baseurl is only taken until the last slash:
<a href="?test">test</a> while the baseurl being `www.example.com/example'
this link will resolve to www.example.com/?test although I expected it to go to www.example.com/example?test
Give a look to RFC-2396 C.1 Normal examples: base = http://a/b/c/d;p?q#f relative = ?y resolved => http://a/b/c/?y Dillo passes all those tests, including the abnormal section. Ah and we also test for these extra tests: http://www.ics.uci.edu/~fielding/url/test2.html http://www.ics.uci.edu/~fielding/url/test3.html http://www.ics.uci.edu/~fielding/url/test4.html http://www.ics.uci.edu/~fielding/url/test5.html HTH. -- Cheers Jorge.-
participants (1)
-
Jorge Arellano Cid