537948 - [HTML5] The procedure to fix a <a> around a <tr> element is acting weird.

Status: RESOLVED INVALID

(Core :: DOM: Core & HTML, defect, P3)

Description

[vjeuxx]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; fr; rv:1.9.2b5) Gecko/20091204 Firefox/3.6b5 (.NET CLR 3.5.30729)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; fr; rv:1.9.2b5) Gecko/20091204 Firefox/3.6b5 (.NET CLR 3.5.30729)

	 -> It moves the <a> before the <table> (good) but closes it (bad)
	 -> If there is a <a> around the <table>, it is being closed (bad)
	 -> If there is a <div> between the outer <a> and <table>, the outer <a>
			is being copied and therefore appear twice, and is being closed. (bad)
	 -> If there is a <span> between the outer <a> and <table>, the outer <a>
			is not being closed (good)


Reproducible: Always

Steps to Reproduce:
Load the page
Actual Results:  
<a class="outer" href=""></a>
<a class="inner"></a>
<table>
	<tr>
		<td>Table</td>
	</tr>
</table>

Expected Results:  
<a class="outer" href="">
	<a class="inner">
		<table>
			<tr>
				<td>Table</td>
			</tr>
		</table>
	</a>
</a>


See the HTML comments for a detailed bug report

Comment 1

[Olli Pettay [:smaug][[e-mail]]]

Henri can probably say what the current HTML5 draft says about this case.

(and HTML5 parser seems to create something strange, many inner a elements or something like that. HTML5 parser isn't enabled by default.)

Comment 2

[vjeuxx]

Comparing to the other browsers, it seems that the inner being closed is the norm but the outer being closed is indeed a bug. None of them have problem with adding <span> and <div> inside the outer <a>

Chrome Beta 4.0.266.0
<a href="" class="outer">
	<a class="inner"></a>
	<table>...</table>
</a>

Opera 10.00 Build 1750

<a href="" class="outer">
	<table>
		<a class="inner"></a>
		<tbody>...</tbody>
	</table>
</a>

Comment 3

[Henri Sivonen (:hsivonen) (not reading bugmail until 2025-02-06)]

With the HTML5 parser, the result is less reasonable than in WebKit. Not sure yet if it's an implementation bug or a spec bug.

Morphing this into an HTML5 parser bug, since other weirdness in the old parser is also being left unfixed.

Comment 4

[Henri Sivonen (:hsivonen) (not reading bugmail until 2025-02-06)]

Is this causing compatibility problems with a real-world Web site?

Comment 5

[Henri Sivonen (:hsivonen) (not reading bugmail until 2025-02-06)]

The behavior here is as specced. Not an implementation bug.

Comment 6

[Henri Sivonen (:hsivonen) (not reading bugmail until 2025-02-06)]

IE8 (both IE8 standards mode and quirks mode) acts mostly like the outer <a> were closed. (The DOM IE8 creates is not a tree, though.)

Is this causing a problem on a real site that works in IE8?

Comment 7

[Henri Sivonen (:hsivonen) (not reading bugmail until 2025-02-06)]

Since evidence of a real-world site breakage hasn't been presented, I'm resolving this INVALID per spec.