W3C home > Mailing lists > Public > public-webapps@w3.org > April to June 2011

Re: [Bug 12111] spec for Storage object getItem(key) method does not match implementation behavior

From: Ian Hickson <ian@hixie.ch>
Date: Thu, 2 Jun 2011 18:59:24 +0000 (UTC)
To: Arthur Barstow <art.barstow@nokia.com>
cc: Philippe Le Hegaret <plh@w3.org>, Aryeh Gregor <Simetrical+w3c@gmail.com>, public-webapps <public-webapps@w3.org>
Message-ID: <Pine.LNX.4.64.1106021855300.26539@ps20323.dreamhostps.com>
On Thu, 2 Jun 2011, Arthur Barstow wrote:
> On Jun/2/2011 2:51 PM, ext Ian Hickson wrote:
> > On Thu, 2 Jun 2011, Philippe Le Hegaret wrote:
> > > On Thu, 2011-06-02 at 18:38 +0000, Ian Hickson wrote:
> > > > On Thu, 2 Jun 2011, Arthur Barstow wrote:
> > > > > Hixie, All - PLH proposed a fix for this bug in comment #5 (use
> > > > > DOMString instead of any in {get,set}Item).
> > > > > 
> > > > > AFAIU, PLH's proposal matches what has been widely implemented. As
> > > > > such, it seems like the spec should be updated accordingly.
> > > > This isn't a bug, it's a new feature that just hasn't been implemented
> > > > yet.
> > > I don't believe that this new feature will get implemented. It's going
> > > to break too many pages on the Web,
> > That's the kind of thing implementation feedback will determine.
> 
> Or flip this around and spec what is widely implemented today and then 
> during the Candidate phase, implementors can submit feedback (e.g. if 
> DOMString doesn't cut it)?

It's always the candidate phase, as far as I'm concerned.

We've already had feedback that just DOMString was insufficent, e.g. it 
doesn't let you store Blobs. That's actually why we added cloning in the 
first place here. We've also received lots of author feedback to the same 
effect. There's plenty of ways we can provide deep-cloning behaviour here 
while still being compatible with legacy sites that don't expect it, e.g. 
we can stringify primitive (non-reference) types, while still cloning 
objects (reference types). However, exactly how this should be done 
depends on what the exact constraints are. There's no rush in finding out 
what the constraints are; implementors are busy doing lots of other things 
at the moment. Eventually, this will bubble to the top and we'll find out, 
and then we can update the spec accordingly.

The platform evolves at its own speed. There's nothing gained by rushing it.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 2 June 2011 18:59:47 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:49:45 GMT