Bug 17421 – Lack of end tag for SELECT element causes the rest of page to not be rendered

• Component: DOM
• Status: VERIFIED FIXED

[Robert Blaut]

The problem was found on the above page. Lack of </selection> tag causes blank a rest of the page rendered. The problem is only visible in Webkit. Other tested browsers: Minefield and Opera has no problem with it. <input> element in these browsers automatically closes <select> element in DOM. Webkit ignores the <input> element.

Check DOM tree in all three browsers: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!doctype%20html%3E%3Chtml%3E%3Cbody%3E%3Cp%3EThere%20should%20be%20GREEN%20bar%20below%3C%2Fp%3E%3Cselect%3E%3Cinput%3E%3Cp%20style%3D%22background-color%3Alime%3B%22%3EPASS%3C%2Fp%3E%3C%2Fbody%3E%3C%2Fhtml%3E

[Robert Blaut]

Created attachment 19184 [details]
minimal test case

[Robert Blaut]

<rdar://problem/5762895>

[Robert Blaut]

I removed URL since the </select> error on a reported page was corrected. 

I wonder if we should fix the bug since HTML5lib renders the test case identically as Webkit today? I don't know if HTML5 parsing compatibility has absolute priority despite other browsers current behavior.

[Alexey Proskuryakov]

I do not know whether this is a bug in the spec, or intentional change. Asking on [e-mail] is probably the way to go, unless Dave or Ian choose to comment here.

[Ian 'Hixie' Hickson]

It's the <input> that's the issue. It should cause a </select> to be implied. I'll fix it in the spec. Ping me if it's not fixed by next week.

[Robert Blaut]

(In reply to comment #5)
> It's the <input> that's the issue. It should cause a </select> to be implied.
> I'll fix it in the spec. Ping me if it's not fixed by next week.
> 

Ian as far as I tested <textarea> also closes <select> element in Gecko, Presto and IE 6.

Test case: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3EThere%20should%20be%20GREEN%20bar%20below%3C%2Fp%3E%3Cselect%3E%3Ctextarea%3E%3C%2Ftextarea%3E%3Cp%20style%3D%22background-color%3Alime%3B%22%3EPASS%3C%2Fp%3E

[Ian 'Hixie' Hickson]

I fixed the spec.

[Robert Blaut]

Created attachment 21456 [details]
Fix for the bug reflecting the latest HTML5 spec changes

Thank you Ian for fixing the specification. Here is my fix proposition with test case attached.

[Darin Adler]

Comment on attachment 21456 [details]
Fix for the bug reflecting the latest HTML5 spec changes

This patch is good.

What would make it better would be to add a call to reportError here so the web inspector would explain the error in the page. I think the call would be:

    reportError(MisplacedContentRetryError, &localName, &currentTagName)

I also think that it would be better to have a test with a text result rather than a Mac-specific render tree result.

r=me, as-is

[Robert Blaut]

Comment on attachment 21456 [details]
Fix for the bug reflecting the latest HTML5 spec changes

Thank you Darin for usefull review. I'll improve the patch soon. So for now I clear the review flag.

[Robert Blaut]

Created attachment 21591 [details]
Improved patch for unclosed SELECT element

I've made changes based on Darin suggestions.

[Darin Adler]

Comment on attachment 21591 [details]
Improved patch for unclosed SELECT element

r=me

[Mark Rowe (bdash)]

Landed in r34734.