665851 - Cookies are not sent to a path of which characters are encoded

Status: RESOLVED DUPLICATE of bug 527043

(Core :: Networking: Cookies, defect)

Description

[Onno Marsman]

User-Agent:       Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0.1) Gecko/20100101 Firefox/4.0.1 FirePHP/0.5
Build Identifier: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.0.1) Gecko/20100101 Firefox/4.0.1 FirePHP/0.5

I submit this bug as is suggested in the discussion at the following bug: https://bugzilla.mozilla.org/show_bug.cgi?id=664398

Firefox automatically encodes urls in some cases when html is manipulated. It is suggested that the paths of cookies should match whether urls are encoded exactly like the path or not. In my example a cookie is set to path /~onno/ but not sent with a request made to /%7Eonno/empty.gif

I do not know whether this should be considered correct behavior or not, but in combination with the html manipulation as described in the original bug it can cause problems of which an example can be found at the url I posted there: http://88.159.207.48/~onno/firefoxbugreport.php
This problem can be solved by either fixing this bug or the original one, but if both are fixed that is of course fine by me too ;)

Reproducible: Always

Steps to Reproduce:
1. Browse to http://88.159.207.48/~onno/firefoxbugreport2.php
2. Look at the results in the NET-tab of firebug

Actual Results:  
- A cookie is set on the initial request with path /~onno/
- A request is made to /~onno/empty.gif with the cookie
- A request is made to /%7Eonno/empty.gif WITHOUT the cookie

Expected Results:  
- A cookie is set on the initial request with path /~onno/
- A request is made to /~onno/empty.gif with the cookie
- A request is made to /%7Eonno/empty.gif WITH the cookie

Comment 1

[Boris Zbarsky [:bzbarsky]]

Thank you for filing this.

Dwitte, any idea what's up here?  Those paths are technically equivalent, though our URI code doesn't canonicalize them....  Is this last all that's going on here from the cookie pov?

Comment 2

[Onno Marsman]

Thought I'd check this again, and the latest version of Firefox still behaves like this. Note that the example url no longer works, I hope that is not a problem.

Comment 4

[Markus.N]

There is a circular "duplicate" mark between this bug and 527043.

Comment 5

[Boris Zbarsky [:bzbarsky]]

No, there is not.  One is "duplicate of bug: N" and the other is "duplicate of this bug: N".  It's just two-way references indicating that this bug is a duplicate of that one.