cc: Removed DCHECK that asserted MapPoint returns clipped as true

cc: Removed DCHECK that asserted MapPoint returns clipped as true

When we apply a perspective to a div, we could end up with a
situation where the mapped points end up behind the view.  When
this happened we would DCHECK, but since this a valid scenario
and the rendered results look correct (as far as I can tell),
I'm removing the DCHECK.

Please see http://jsfiddle.net/kkryf2v4/ for an example where
we animate from non-clipped to clipped div.

BUG=416367
Committed: https://crrev.com/06aebb2f4b89f7bb0a511fa12881e919a860287a
Cr-Commit-Position: refs/heads/master@{#297082}

[hendrikw, 2014-09-26 20:55:15]

hendrikw@chromium.org changed reviewers:
+ reveman@chromium.org

[hendrikw, 2014-09-26 20:55:53]

PTAL, thanks!

[danakj, 2014-09-26 21:09:19]

danakj@chromium.org changed reviewers:
+ danakj@chromium.org

[danakj, 2014-09-26 21:09:19]

https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
File cc/output/gl_renderer.cc (left):

https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
If it's clipped then the coordinates returned by MapPoint are not valid:
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....

[hendrikw, 2014-09-26 21:11:31]

On 2014/09/26 21:09:19, danakj wrote:
> https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
> File cc/output/gl_renderer.cc (left):
> 
>
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
> cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
> If it's clipped then the coordinates returned by MapPoint are not valid:
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....

Actually, they are.

Well, unless w happens to be zero, but I've never seen that happen.

[hendrikw, 2014-09-26 21:15:02]

On 2014/09/26 21:11:31, Hendrik wrote:
> On 2014/09/26 21:09:19, danakj wrote:
> > https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
> > File cc/output/gl_renderer.cc (left):
> > 
> >
>
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
> > cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
> > If it's clipped then the coordinates returned by MapPoint are not valid:
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....
> 
> Actually, they are.
> 
> Well, unless w happens to be zero, but I've never seen that happen.

Well, I should be more clear on that.  The value they return produces results
that you would expect in that situation.

[danakj, 2014-09-26 21:21:58]

On 2014/09/26 21:15:02, Hendrik wrote:
> On 2014/09/26 21:11:31, Hendrik wrote:
> > On 2014/09/26 21:09:19, danakj wrote:
> > > https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
> > > File cc/output/gl_renderer.cc (left):
> > > 
> > >
> >
>
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
> > > cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
> > > If it's clipped then the coordinates returned by MapPoint are not valid:
> > >
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....
> > 
> > Actually, they are.
> > 
> > Well, unless w happens to be zero, but I've never seen that happen.
> 
> Well, I should be more clear on that.  The value they return produces results
> that you would expect in that situation.

To get correct values it should be doing something more like this with the
point, no?

https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....

[hendrikw, 2014-09-26 21:34:04]

On 2014/09/26 21:21:58, danakj wrote:
> On 2014/09/26 21:15:02, Hendrik wrote:
> > On 2014/09/26 21:11:31, Hendrik wrote:
> > > On 2014/09/26 21:09:19, danakj wrote:
> > > >
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
> > > > File cc/output/gl_renderer.cc (left):
> > > > 
> > > >
> > >
> >
>
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
> > > > cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
> > > > If it's clipped then the coordinates returned by MapPoint are not valid:
> > > >
> > >
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....
> > > 
> > > Actually, they are.
> > > 
> > > Well, unless w happens to be zero, but I've never seen that happen.
> > 
> > Well, I should be more clear on that.  The value they return produces
results
> > that you would expect in that situation.
> 
> To get correct values it should be doing something more like this with the
> point, no?
> 
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....

I'm assuming pixels drawn outside of the viewport will not be drawn?  Right? 
The near plane is zero (i.e. right at the 'camera', if you can call it that), so
I would never expect the quad to be clipped in such a way that it would be
visible (unless its normal is perpendicular to the view direction in which case
it could only draw a line, or nothing maybe).  That code seems like overkill for
this.  Maybe I'm wrong.

[hendrikw, 2014-09-26 21:47:23]

On 2014/09/26 21:34:04, Hendrik wrote:
> On 2014/09/26 21:21:58, danakj wrote:
> > On 2014/09/26 21:15:02, Hendrik wrote:
> > > On 2014/09/26 21:11:31, Hendrik wrote:
> > > > On 2014/09/26 21:09:19, danakj wrote:
> > > > >
> https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
> > > > > File cc/output/gl_renderer.cc (left):
> > > > > 
> > > > >
> > > >
> > >
> >
>
https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
> > > > > cc/output/gl_renderer.cc:1426: DCHECK(!clipped);
> > > > > If it's clipped then the coordinates returned by MapPoint are not
valid:
> > > > >
> > > >
> > >
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....
> > > > 
> > > > Actually, they are.
> > > > 
> > > > Well, unless w happens to be zero, but I've never seen that happen.
> > > 
> > > Well, I should be more clear on that.  The value they return produces
> results
> > > that you would expect in that situation.
> > 
> > To get correct values it should be doing something more like this with the
> > point, no?
> > 
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/cc/base/math_util....
> 
> I'm assuming pixels drawn outside of the viewport will not be drawn?  Right? 
> The near plane is zero (i.e. right at the 'camera', if you can call it that),
so
> I would never expect the quad to be clipped in such a way that it would be
> visible (unless its normal is perpendicular to the view direction in which
case
> it could only draw a line, or nothing maybe).  That code seems like overkill
for
> this.  Maybe I'm wrong.

For fun, I tried to set up the situation where we sweep across the camera (and
behind).

http://jsfiddle.net/1n8qppoo/2/

Looks right to me.

[danakj, 2014-09-26 23:04:34]

ok LGTM

https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc
File cc/output/gl_renderer.cc (left):

https://codereview.chromium.org/610603003/diff/1/cc/output/gl_renderer.cc#old...
cc/output/gl_renderer.cc:1425: bottom_right =
MathUtil::MapPoint(device_transform, bottom_right, &clipped);
can you add comment explaining why we are ignoring the clipped value here and
point it at the bug

[hendrikw, 2014-09-26 23:15:07]

Updated comment

[danakj, 2014-09-26 23:30:15]

Thanks! LGTM

[hendrikw, 2014-09-26 23:30:50]

The CQ bit was checked by hendrikw@chromium.org

[commit-bot: I haz the power, 2014-09-26 23:31:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/610603003/20001

[commit-bot: I haz the power, 2014-09-27 00:29:27]

Committed patchset #2 (id:20001) as 7541b6443a24fd77ffb6ef75babe24762e4b16d9

[commit-bot: I haz the power, 2014-09-27 00:30:04]

Patchset 2 (id:??) landed as
https://crrev.com/06aebb2f4b89f7bb0a511fa12881e919a860287a
Cr-Commit-Position: refs/heads/master@{#297082}