[Webkit-unassigned] [Bug 34529] New: [CSSOM] issues with cssText and selectorText

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=34529

           Summary: [CSSOM] issues with cssText and selectorText
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: All
               URL: http://glazman.org/forCSSWG/selectorTextBug.html
        OS/Version: Mac OS X 10.5
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: CSS
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: simon.fraser at apple.com


Resent-From:     www-style list <www-style at w3.org>
    From:     Daniel Glazman <daniel.glazman at disruptive-innovations.com>
    Subject:     [CSSOM] issues with cssText and selectorText
    Date:     February 3, 2010 7:58:15 am PST
    To:     www-style list <www-style at w3.org>
Message-Id:     <4B699D17.5060009 at disruptive-innovations.com>

Hi there,

I just discovered that Opera is the only "major" browser
that does not return a useless result for selectorText and
cssText in some cases that are not edge cases any more...

If the selector of a rule contains a ident itself containing
a colon, the colon must be escaped in the selector. But its
serialization in selectorText and cssText is NOT escaped. The
same issue happens with many escaped characters.

So the result of cssText and selectorText is useless because
unparseable and invalid...

FWIW, I have written a small test here

 http://glazman.org/forCSSWG/selectorTextBug.html

I said it's not an edge case any more because a small site called
Facebook gives its partners embeddable stylesheets with such selectors.
I discovered them validating a stylesheet in a famous newspaper's
website.
It's also super important to have a bijection here for editing
environments. If cssText and selectorText are unreliable, then building
a conforming style editor is almost impossible.

I think we should clarify the situation here, saying that the result
of selectorText and cssText must be parseable and represent really
what they are the serialization of...

FWIW, both Safari and Mozilla fail on the test above. MSIE of course
is not even able of running the three lines of script...

</Daniel>

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.