[Webkit-unassigned] [Bug 56703] window.localStorage should throw SECURITY_ERR when localStorage is disabled

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=56703





--- Comment #15 from Brady Eidson <beidson at apple.com>  2011-12-02 12:16:19 PST ---
(In reply to comment #14)
> As background, if that helps, I discovered the difference with the spec while working on this bug:
> 
> https://bugs.webkit.org/show_bug.cgi?id=56354 Web Inspector: empty, non-functional window
> 
> Although it's true that there's a "may" in step 1, there's a "must" prefacing the steps, and if you go through all of the steps, it seems as far as I can tell that the only options are to throw an exception or return a local storage object. 

I think you're referring to "User agents must have a set of local storage areas, one for each origin" and yes - to be a *full* HTML5 compliant user agent it would have to have the local storage areas.

But WebKit is not *only* a 100% HTML5 engine.  It supports plenty of modes where various HTML5 features are missing entirely.  This is one of those modes we need to continue supporting.

>Returning null doesn't seem to be a valid option according to the spec.

This made me realize that yes, returning null for a missing feature is incorrect behavior.  We should be returning undefined here, which is the appropriate behavior for any user agent and any missing feature.

> I should note that in my experience of running Safari with local storage disabled, which I continue to do to this day, the choice is usually not between throwing an exception or having a web site work correctly. The web sites just assume they can use local storage and don't even check for null, so the choice is between throwing an exception or getting a javascript error.

How is it you run Safari with LocalStorage disabled?  AFAIK, Safari doesn't support doing this.  Twiddling the hidden WebKit default yourself gets you in to unsupported power-user land.

And this may be true for web sites where you notice it, but I'm not suggesting there *might be* web content that properly and wisely do feature testing and change their behavior to match.  I'm saying there *is* such content that this change would break.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.