212891 - Hostname not url-decoded when DNS lookup is made on url-encoded a href value

Status: VERIFIED DUPLICATE of bug 43659

(Core :: Networking, defect)

Description

[Ian Batterbee]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624

If I create a link to a site and the string contains URL-encoded characters (for
example http://www.googl%2565.com), the URL is correctly decoded on the status
line as http://www.google.com, however if I click on the link, mozilla attempts
to do a DNS lookup on the un-decoded version, fails, and presents an alert
saying that site can not be found.

In addition, and probably because of this bug, the URL is not recognised as
being one that has already been visited, so shows up in the 'wrong' colour on
the page.


Reproducible: Always

Steps to Reproduce:
Create a document containing the following code:

I'm not sure if mozilla will automatically encode the appropriate characters, so
if this comes out looking really dumb, I'll post it properly in the next comment.

blah blah <a href='http://www.googl%65.com'>link</a>
<br>
blah blah <a href='http://www.google.com'>link</a>

Note that the status line displays the same value when you put your mouse over
either link, but only the second link will actually work when clicked on.

Actual Results:  
It appears that while the status line gets or at least displays the decoded
value, anything else that uses it looks at the raw url-encoded string.

Expected Results:  
Both links should act the same way

Comment 1

[Ian Batterbee]

bugzilla is smarter than I expected... please ignore the %25 in bug description.
The example URL should read as  'http://www.googl%65.com' 

Comment 2

[Jo Hermans]

*** This bug has been marked as a duplicate of 43659 ***

Comment 3

[benc]

VERIFIED/dupe.

Status unescapes the URL to make it more readable.

I suppose we could modify it to not unescape stuff that we won't accept...