Add opaque "layout item" to widgets for caller's use.

Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls, happening one object lower, pass
the same argument as which they are being inovked against.

Committed: https://pdfium.googlesource.com/pdfium/+/54ab2343c338f55735c37ed6751040791e1df1e2

[Tom Sepez, 2016-05-26 01:29:18]

Description was changed from

==========
Add opaque "layout item" to widges for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls. happening one object lower, pass
the same argument as which they are being inovked against.
==========

to

==========
Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls. happening one object lower, pass
the same argument as which they are being inovked against.
==========

[Tom Sepez, 2016-05-26 01:29:47]

Description was changed from

==========
Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls. happening one object lower, pass
the same argument as which they are being inovked against.
==========

to

==========
Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls, happening one object lower, pass
the same argument as which they are being inovked against.
==========

[Tom Sepez, 2016-05-26 17:09:32]

tsepez@chromium.org changed reviewers:
+ thestig@chromium.org

[Tom Sepez, 2016-05-26 17:09:44]

Lei, for review.

https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
File xfa/fwl/core/fwl_widgetimp.h (right):

https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
xfa/fwl/core/fwl_widgetimp.h:167: void* m_pLayoutItem;
Note: went with void* for now to avoid making parallel types in three different
widget layers.

[Lei Zhang, 2016-05-26 22:06:46]

https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
File xfa/fwl/core/fwl_widgetimp.h (right):

https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
xfa/fwl/core/fwl_widgetimp.h:167: void* m_pLayoutItem;
On 2016/05/26 17:09:44, Tom Sepez wrote:
> Note: went with void* for now to avoid making parallel types in three
different
> widget layers.

Does it have to be void* ? I played around with the code locally and came up
with https://codereview.chromium.org/2017773002/ (apply on top of this CL)

[Lei Zhang, 2016-05-26 22:43:18]

https://codereview.chromium.org/2010923002/diff/60001/xfa/fxfa/app/xfa_fwlada...
File xfa/fxfa/app/xfa_fwladapter.cpp (right):

https://codereview.chromium.org/2010923002/diff/60001/xfa/fxfa/app/xfa_fwlada...
xfa/fxfa/app/xfa_fwladapter.cpp:21: CXFA_FFField* pField =
static_cast<CXFA_FFField*>(pWidget->GetLayoutItem());
If you static_cast this to a CXFA_FFWidget instead, then all the GetLayoutItem()
calls cast to CXFA_FFWidget. And rename |pField| to match.

[Tom Sepez, 2016-05-26 23:56:50]

On 2016/05/26 22:06:46, Lei Zhang wrote:
>
https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
> File xfa/fwl/core/fwl_widgetimp.h (right):
> 
>
https://codereview.chromium.org/2010923002/diff/40001/xfa/fwl/core/fwl_widget...
> xfa/fwl/core/fwl_widgetimp.h:167: void* m_pLayoutItem;
> On 2016/05/26 17:09:44, Tom Sepez wrote:
> > Note: went with void* for now to avoid making parallel types in three
> different
> > widget layers.
> 
> Does it have to be void* ? I played around with the code locally and came up
> with https://codereview.chromium.org/2017773002/ (apply on top of this CL)

Yes, I want to keep it void for now, see comment on other CL.

[Tom Sepez, 2016-05-27 00:02:15]

https://codereview.chromium.org/2010923002/diff/60001/xfa/fxfa/app/xfa_fwlada...
File xfa/fxfa/app/xfa_fwladapter.cpp (right):

https://codereview.chromium.org/2010923002/diff/60001/xfa/fxfa/app/xfa_fwlada...
xfa/fxfa/app/xfa_fwladapter.cpp:21: CXFA_FFField* pField =
static_cast<CXFA_FFField*>(pWidget->GetLayoutItem());
On 2016/05/26 22:43:17, Lei Zhang wrote:
> If you static_cast this to a CXFA_FFWidget instead, then all the
GetLayoutItem()
> calls cast to CXFA_FFWidget. And rename |pField| to match.

Done.

[Lei Zhang, 2016-05-27 00:35:05]

lgtm

[Tom Sepez, 2016-05-27 00:35:29]

The CQ bit was checked by tsepez@chromium.org

[commit-bot: I haz the power, 2016-05-27 00:35:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2010923002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/2010923002/80001

[commit-bot: I haz the power, 2016-05-27 00:36:00]

Description was changed from

==========
Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls, happening one object lower, pass
the same argument as which they are being inovked against.
==========

to

==========
Add opaque "layout item" to widgets for caller's use.

This avoids another use of CFX_PrivateData.  Note that in the old
code, we'd be calling through a m_pImpl onto the same underlying
object as we passed as the "key" argument when setting the value,
which explains why the get calls, happening one object lower, pass
the same argument as which they are being inovked against.

Committed:
https://pdfium.googlesource.com/pdfium/+/54ab2343c338f55735c37ed6751040791e1d...
==========

[commit-bot: I haz the power, 2016-05-27 00:36:01]

Committed patchset #5 (id:80001) as
https://pdfium.googlesource.com/pdfium/+/54ab2343c338f55735c37ed6751040791e1d...