W3C home > Mailing lists > Public > public-script-coord@w3.org > April to June 2014

Re: removeEventListener with only one passed parameter...

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Mon, 9 Jun 2014 13:22:03 -0700
Message-ID: <CAAWBYDB=0JG5W8jcJUZds0NMo3M4pe_z5U7ck71f_t70mcZzmg@mail.gmail.com>
To: Travis Leithead <travis.leithead@microsoft.com>
Cc: Gary Kačmarčík (Кошмарчик) <garykac@google.com>, "Anne van Kesteren (annevk@annevk.nl)" <annevk@annevk.nl>, "Boris Zbarsky (bzbarsky@mozilla.com)" <bzbarsky@mozilla.com>, Cameron McCormack <cmccormack@mozilla.com>, "public-script-coord@w3.org" <public-script-coord@w3.org>
On Mon, Jun 9, 2014 at 1:07 PM, Travis Leithead
<travis.leithead@microsoft.com> wrote:
> Hey folks,
>
> I was reviewing a site-compatibility issue with IE on the phone, and ran
> across what I think to be a Chrome implementation bug, but I wanted to post
> my findings in case I messed up the interpretation—it has to do with
> undefined handling in WebIDL I think.
>
> The API is removeEventListener defined in DOM4 [1] as:
>
> void removeEventListener(DOMString type, EventListener? callback, optional
> boolean capture = false);
>
> The site-compatibility problem occurs because the site is [unknowingly?]
> calling this API with only one param:
>
> ob.removeEventListener(‘click’);
>
> and no exception is thrown. In IE and Firefox, this throws because 2 params
> are expected to this function. In Chrome, the call silently fails (no
> exception). Note that IE and Firefox do not throw when passing undefined
> explicitly: ob.removeEventListener(‘click’, undefined).
>
> Per WebIDL AFAICT, this method would be invoked with the following values:
> (‘click’, <missing>, false), since the optional 3rd arg with a default gets
> substituted when no value is passed (and would also be substituted if
> undefined as a value were specifically passed). I’m pretty sure the nullable
> 2nd arg is not considered “optional” per the “optionality” determination in
> WebIDL. If it is, then undefined would type-convert to null, and Chrome’s
> behavior of not-throwing is right, but I can’t quite see how that would
> work.
>
> Would love your clarification. Is this a Chrome bug or a spec bug?
>
> [1] http://w3c.github.io/dom/#eventtarget

Chrome bug.  We've got a lot of places where we use standard JS
behavior for missing values (passing in undefined) rather than WebIDL
behavior (throwing).

~TJ
Received on Monday, 9 June 2014 20:22:50 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:21 UTC