301685 - Setting table's tHead or tFoot to wrong type of table section does weird things

Status: VERIFIED FIXED

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

Description

[Aiko]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b3) Gecko/20050712 Firefox/1.0+
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b3) Gecko/20050712 Firefox/1.0+

When trying to set the wrong kind of element as a table's caption, tHead or
tFoot the browser should throw HIERARCHY_REQUEST_ERR. It throws
XPC_BAD_CONVERT_JS instead.

Reproducible: Always

Steps to Reproduce:
1. See testcase

Comment 1

[Aiko]

Attachment: testcase

Comment 2

[Aiko]

On a related note, Mozilla allows setting tHead to a TBODY or TFOOT element and
setting tFoot to a TBODY or THEAD element. It shouldn't, the following should
throw an exception, too:

tbl = document.createElement("table")
tbl.tFoot = document.createElement("thead")

Instead, this sets tbl's tHead, tFoot is still null. Weird.

Comment 3

[Boris Zbarsky [:bzbarsky]]

This bug is invalid.  You're not passing the right type, so this code is never even making it into the DOM.  In C++ you'd get a compile-time error for this code, for example.

Comment 4

[Aiko]

(In reply to comment #3)
> This bug is invalid.  You're not passing the right type, so this code is never
> even making it into the DOM.  In C++ you'd get a compile-time error for this
> code, for example.

But this isn't C++. :-) Ok, internal these are probably two very different things, but from an API users point of view there isn't much difference between setting table.caption to a HTMLTableCellElement and setting document.body to a HTMLTableCellElement, which generates a HIERARCHY_REQUEST_ERR. Same error, same error message, no?

And what about the problem I mentioned in comment 2? Setting table.tHead to a TFOOT element leaves table.tHead untouched and sets table.tFoot instead. That's certainly not desirable behaviour.

I certainly don't claim these are devastating bugs that must be fixed ASAP but they are not invalid, IMHO.

Comment 5

[Aiko]

Attachment: testcase for the second problem

Comment 6

[Boris Zbarsky [:bzbarsky]]

> there isn't much difference between setting table.caption to a
> HTMLTableCellElement and setting document.body to a HTMLTableCellElement

document.body is an HTMLElement.  table.caption is an HTMLTableCaptionElement (per the DOM spec).  An HTMLTableCellElement is an HTMLElement, but not an HTMLTableCaptionElement.

So passing an HTMLTableCellElement to document.body is not a priori unreasonable (per the DOM API) and throws for reasons internal to the DOM.  Passing it to table.caption can't even make it into the DOM, like I said -- there's just no way to force that square peg into that round hole.  So no, it's not the same error -- in one case you're passing an X to a method that takes Y, while in athe other case the method takes X but you're not passing quite the right X.

> And what about the problem I mentioned in comment 2?

One bug per problem.  Please file a separate bug on that issue and cc me.

Comment 7

[Aiko]

Since you say the first problem isn't actually one there is just one problem left here, so I'll just fix the summary.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Mutating bugs is generally bad form, for what it's worth...

Comment 9

[Boris Zbarsky [:bzbarsky]]

Attachment: Fix

Comment 10

[Boris Zbarsky [:bzbarsky]]

Fixed.

Comment 11

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

You don't want to check the HTML-ness too?

Comment 12

[Boris Zbarsky [:bzbarsky]]

It's already an nsIDOMHTMLTableSectionElement.  There's not much more to check in the way of HTML-ness... ;)