Move Element.dataset to HTMLElement.dataset

Move Element.dataset to HTMLElement.dataset

BUG=498485
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197259

[ramya.v, 2015-06-16 08:58:05]

ramya.v@samsung.com changed reviewers:
+ philipj@opera.com

[ramya.v, 2015-06-16 08:58:06]

PTAL!

Thanks

[philipj_slow, 2015-06-16 15:42:18]

Can it be moved in the implementation as well, or are there internal uses of
this API? I haven't looked and am not sure what would be the ideal state, so
what do you think?

[ramya.v, 2015-06-17 06:28:29]

On 2015/06/16 15:42:18, philipj wrote:
> Can it be moved in the implementation as well, or are there internal uses of
> this API? I haven't looked and am not sure what would be the ideal state, so
> what do you think?

There are no internal uses of this API.
Since we can't directly use ElementRareData from HTMLElement,
we need to move the internal details of dataset member from ElementRareData to
HTMLElement.
Updated patchset2. PTAL!


Thanks

[philipj_slow, 2015-06-17 11:39:35]

I think moving m_dataset to HTMLElement is probably not a good idea, it can't be
common enough to justify that, so ElementRareData seems better. Since the bits
on Element.h only forward to ElementRareData I think leaving the implementation
side alone is actually fine. Making the ElementRareData bits protected instead
of private would also work, but that would expose it to random things like
HTMLTableElement which probably shouldn't be poking at it.

In other words, let's go with PS1.

[ramya.v, 2015-06-17 11:44:52]

On 2015/06/17 11:39:35, philipj wrote:
> I think moving m_dataset to HTMLElement is probably not a good idea, it can't
be
> common enough to justify that, so ElementRareData seems better. Since the bits
> on Element.h only forward to ElementRareData I think leaving the
implementation
> side alone is actually fine. Making the ElementRareData bits protected instead
> of private would also work, but that would expose it to random things like
> HTMLTableElement which probably shouldn't be poking at it.
> 
> In other words, let's go with PS1.

Thanks for the comments philip. I also feel the same.
Shall I delete the patchset2 or override it for history?

[philipj_slow, 2015-06-17 11:46:22]

On 2015/06/17 11:44:52, ramya.v wrote:
> On 2015/06/17 11:39:35, philipj wrote:
> > I think moving m_dataset to HTMLElement is probably not a good idea, it
can't
> be
> > common enough to justify that, so ElementRareData seems better. Since the
bits
> > on Element.h only forward to ElementRareData I think leaving the
> implementation
> > side alone is actually fine. Making the ElementRareData bits protected
instead
> > of private would also work, but that would expose it to random things like
> > HTMLTableElement which probably shouldn't be poking at it.
> > 
> > In other words, let's go with PS1.
> 
> Thanks for the comments philip. I also feel the same.
> Shall I delete the patchset2 or override it for history?

Just upload a PS3 that's identical to PS1, I don't know if it's even possible to
delete PS2 :)

[ramya.v, 2015-06-17 11:51:42]

On 2015/06/17 11:46:22, philipj wrote:
> On 2015/06/17 11:44:52, ramya.v wrote:
> > On 2015/06/17 11:39:35, philipj wrote:
> > > I think moving m_dataset to HTMLElement is probably not a good idea, it
> can't
> > be
> > > common enough to justify that, so ElementRareData seems better. Since the
> bits
> > > on Element.h only forward to ElementRareData I think leaving the
> > implementation
> > > side alone is actually fine. Making the ElementRareData bits protected
> instead
> > > of private would also work, but that would expose it to random things like
> > > HTMLTableElement which probably shouldn't be poking at it.
> > > 
> > > In other words, let's go with PS1.
> > 
> > Thanks for the comments philip. I also feel the same.
> > Shall I delete the patchset2 or override it for history?
> 
> Just upload a PS3 that's identical to PS1, I don't know if it's even possible
to
> delete PS2 :)


Uploaded same patch as PS1. PTAL! (There is a delete option beside download
option of every patchset :) )

[philipj_slow, 2015-06-17 12:01:21]

On 2015/06/17 11:51:42, ramya.v wrote:
> On 2015/06/17 11:46:22, philipj wrote:
> > On 2015/06/17 11:44:52, ramya.v wrote:
> > > On 2015/06/17 11:39:35, philipj wrote:
> > > > I think moving m_dataset to HTMLElement is probably not a good idea, it
> > can't
> > > be
> > > > common enough to justify that, so ElementRareData seems better. Since
the
> > bits
> > > > on Element.h only forward to ElementRareData I think leaving the
> > > implementation
> > > > side alone is actually fine. Making the ElementRareData bits protected
> > instead
> > > > of private would also work, but that would expose it to random things
like
> > > > HTMLTableElement which probably shouldn't be poking at it.
> > > > 
> > > > In other words, let's go with PS1.
> > > 
> > > Thanks for the comments philip. I also feel the same.
> > > Shall I delete the patchset2 or override it for history?
> > 
> > Just upload a PS3 that's identical to PS1, I don't know if it's even
possible
> to
> > delete PS2 :)
> 
> 
> Uploaded same patch as PS1. PTAL! (There is a delete option beside download
> option of every patchset :) )

Oh, I'd actually never noticed that :)

[philipj_slow, 2015-06-17 12:01:24]

The CQ bit was checked by philipj@opera.com

[philipj_slow, 2015-06-17 12:01:25]

lgtm

[commit-bot: I haz the power, 2015-06-17 12:01:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1189663004/40001

[commit-bot: I haz the power, 2015-06-17 13:25:01]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197259