Bug 44021 – <style> inside <noscript> is parsed when JavaScript is enabled.

• Component: DOM
• Status: RESOLVED FIXED

Summary:
When adding <style> definitions inside of a <noscript> tag the style declarations are applied even when JavaScript is enabled.

Steps to Reproduce:
<noscript>
  <style type="text/css">
    #myId { display: none; }
  </style>
</noscript>

Expected Results:
Style declarations inside of no script, attempting to gracefully degrade a form, should not be parsed. When JavaScript is in enabled #myDiv should remain visible, where during a disabled JavaScript page load the element should not be displayed.

Actual Results:
Style definitions are applied when JavaScript is enabled.
Firefox 3.6.8 correctly ignores <style> in <noscript>
IE 8 (Win 7) correctly ignores <style> in <noscript>

Chrome 5.0.375.99 parses <style> in <noscript>
Safari 5.0.1 (6533.17.8) parses <style> in <noscript>

[Adam Barth]

In all likelihood this is fixed on trunk.  Testing now.

[Adam Barth]

*** Bug 28195 has been marked as a duplicate of this bug. ***

[Adam Barth]

Created attachment 64668 [details]
Patch

[Alexey Proskuryakov]

Comment on attachment 64668 [details]
Patch

Thanks! It's a little surprising to see noscript without script, and maybe we should be also testing the more common case where they come together, but I don't think there is a strong technical reason to.

I'm a little confused by your comment in bug 28195 - is 13522 the same, after all?

[Adam Barth]

> Thanks! It's a little surprising to see noscript without script, and maybe we should be also testing the more common case where they come together, but I don't think there is a strong technical reason to.

Huh?  I don't understand this comment.  Why would script and noscript come together?

> I'm a little confused by your comment in bug 28195 - is 13522 the same, after all?

I haven't looked at Bug 13522 yet.  Bug 28195 is the same bug as this one.  The old parser didn't understand how to properly lex <noscript>.  The new parser lexs it as RAWTEXT and then makes it not display.

[Alexey Proskuryakov]

> Huh?  I don't understand this comment.  Why would script and noscript come together?

They usually did on Web pages (and sample code) where I saw <noscript>, with it logically being fallback for scripted behavior. Of course, they are not technically related, which is why I said there was no strong reason to test that case.

[WebKit Commit Bot]

Comment on attachment 64668 [details]
Patch

Clearing flags on attachment: 64668

Committed r65712: <http://trac.webkit.org/changeset/65712>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[WebKit Review Bot]

http://trac.webkit.org/changeset/65712 might have broken SnowLeopard Intel Release (Tests)
The following changes are on the blame list:
http://trac.webkit.org/changeset/65699
http://trac.webkit.org/changeset/65700
http://trac.webkit.org/changeset/65701
http://trac.webkit.org/changeset/65702
http://trac.webkit.org/changeset/65703
http://trac.webkit.org/changeset/65704
http://trac.webkit.org/changeset/65705
http://trac.webkit.org/changeset/65706
http://trac.webkit.org/changeset/65707
http://trac.webkit.org/changeset/65708
http://trac.webkit.org/changeset/65709
http://trac.webkit.org/changeset/65710
http://trac.webkit.org/changeset/65711
http://trac.webkit.org/changeset/65712
http://trac.webkit.org/changeset/65713