964212 - Node.insertBefore() method should allow the second parameter being omitted

Status: RESOLVED INVALID

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

Description

[Duan Yao]

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:27.0) Gecko/20100101 Firefox/27.0 (Beta/Release)
Build ID: 20140115144847

Steps to reproduce:

Omit the second parameter of Node.insertBefore() method, e.g.:
document.documentElement.insertBefore(document.createElement('div'));



Actual results:

An exception is thrown:
TypeError: Not enough arguments to Node.insertBefore.


Expected results:

No exception, child node is appended to the parant node.

Comment 1

[Duan Yao]

I known that the spec (http://www.w3.org/TR/2000/REC-DOM-Level-2-Core-20001113/core.html) does not say that insertBefore(newChild, refChild)'s param refChild is optional, only say that null is allowed.

However, chrome and IE do allow refChild being omitted or being `undefined`; firefox allow refChild being `undefined` but not omitted.

Chrome and IE's behavior looks more natural to me; while firefox's behavior is a little confusing: if a param can be undefined/null, why it can't be omitted?

So I suggest that firefox should follow chrome and IE's behavior. It even better if the spec could be revised to make parameter refChild optinal.

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

You're looking at an ancient spec. However, the current specification matches Firefox: <http://dom.spec.whatwg.org/#node>. If you want to append a child, use appendChild instead.

If you think the spec should be changed, there's a "Participate" section at the top that explains where you can make that argument. (It's not here.)

Comment 3

[Duan Yao]

Sorry, but I can't find a statement in the current spec says what insertBefore should do if the second param is omited.
Do you think chrome and IE are incorrect here and I should report bugs to them instead?
Regards.

Comment 4

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

If an argument is not explicitly marked as optional, it is required, and a TypeError is thrown when it is not passed. This is specified in IDL:

http://heycam.github.io/webidl/#es-operations
http://heycam.github.io/webidl/#dfn-overload-resolution-algorithm

Filing bugs on other browsers f they get this wrong is probably a good idea.

Comment 5

[Duan Yao]

I'll think about it. Thank you!