Metro/HiDPI: Mark Chrome as HiDPI aware on Windows

Metro/HiDPI: Mark Chrome as HiDPI aware on Windows

With this CL Chrome is marked as HiDPI aware if the enable_hidpi build flag is set.

BUG=114311
TEST=Changed Windows to 2x DPI scale. Ran Chrome. Verified that 2x icons were used.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=133826

[sail, 2012-04-13 21:01:36]

jam: OWNERS review
cpu: HiDPI code

[jam, 2012-04-15 22:23:14]

lgtm

[jam, 2012-04-15 22:23:53]

does this need to be called for all process types, or just the browser process?

[sail, 2012-04-16 17:30:45]

On 2012/04/15 22:23:53, John Abd-El-Malek wrote:
> does this need to be called for all process types, or just the browser
process?

I thought all processes would need this but now I'm not sure.
CCing fsamuel who should know.

[sail, 2012-04-16 17:31:21]

Hi Fady. On Windows do we need to mark all processes as HiDPI aware or just the
browser process?

[cpu_(ooo_6.6-7.5), 2012-04-16 19:09:15]

lgtm

[Fady Samuel, 2012-04-16 19:31:25]

On 2012/04/16 17:31:21, sail wrote:
> Hi Fady. On Windows do we need to mark all processes as HiDPI aware or just
the
> browser process?

Hi, right now to enable high DPI in WebKit, you would use the
--default-device-scale-factor=x where x > 1. This needs to be passed to all the
renderers (and is). However, this flag currently exists for debugging purposes
and things may be done differently a little later.

[sail, 2012-04-16 19:35:45]

On 2012/04/16 19:31:25, Fady Samuel wrote:
> On 2012/04/16 17:31:21, sail wrote:
> > Hi Fady. On Windows do we need to mark all processes as HiDPI aware or just
> the
> > browser process?
> 
> Hi, right now to enable high DPI in WebKit, you would use the
> --default-device-scale-factor=x where x > 1. This needs to be passed to all
the
> renderers (and is). However, this flag currently exists for debugging purposes
> and things may be done differently a little later.

Hi Fady. I'm asking a slightly different question.
On Windows, does the renderer process need to be HiDPI aware? If a process is
not marked as HiDPI aware then when it calls system APIs everything is given in
logical coordinates. So for example, mouse coordinates might be different
between the browser and the renderer.

[Fady Samuel, 2012-04-16 19:41:00]

On 2012/04/16 19:35:45, sail wrote:
> On 2012/04/16 19:31:25, Fady Samuel wrote:
> > On 2012/04/16 17:31:21, sail wrote:
> > > Hi Fady. On Windows do we need to mark all processes as HiDPI aware or
just
> > the
> > > browser process?
> > 
> > Hi, right now to enable high DPI in WebKit, you would use the
> > --default-device-scale-factor=x where x > 1. This needs to be passed to all
> the
> > renderers (and is). However, this flag currently exists for debugging
purposes
> > and things may be done differently a little later.
> 
> Hi Fady. I'm asking a slightly different question.
> On Windows, does the renderer process need to be HiDPI aware? If a process is
> not marked as HiDPI aware then when it calls system APIs everything is given
in
> logical coordinates. So for example, mouse coordinates might be different
> between the browser and the renderer.

I haven't really played with High DPI on Windows, unfortunately so I'm not sure
of the answer to your question. I want to say yes, since WebKit is High-DPI
aware, the renderers should be marked as high-DPI aware, but I'm not sure,
sorry.

[sail, 2012-04-16 19:43:02]

On 2012/04/16 19:41:00, Fady Samuel wrote:
> On 2012/04/16 19:35:45, sail wrote:
> > On 2012/04/16 19:31:25, Fady Samuel wrote:
> > > On 2012/04/16 17:31:21, sail wrote:
> > > > Hi Fady. On Windows do we need to mark all processes as HiDPI aware or
> just
> > > the
> > > > browser process?
> > > 
> > > Hi, right now to enable high DPI in WebKit, you would use the
> > > --default-device-scale-factor=x where x > 1. This needs to be passed to
all
> > the
> > > renderers (and is). However, this flag currently exists for debugging
> purposes
> > > and things may be done differently a little later.
> > 
> > Hi Fady. I'm asking a slightly different question.
> > On Windows, does the renderer process need to be HiDPI aware? If a process
is
> > not marked as HiDPI aware then when it calls system APIs everything is given
> in
> > logical coordinates. So for example, mouse coordinates might be different
> > between the browser and the renderer.
> 
> I haven't really played with High DPI on Windows, unfortunately so I'm not
sure
> of the answer to your question. I want to say yes, since WebKit is High-DPI
> aware, the renderers should be marked as high-DPI aware, but I'm not sure,
> sorry.

Sounds good. I think it's probably worth landing as is just so we can start
testing things. We can always move this out to the exe_main code if needed.

[jam, 2012-04-17 03:15:31]

On 2012/04/16 19:43:02, sail wrote:
> On 2012/04/16 19:41:00, Fady Samuel wrote:
> > On 2012/04/16 19:35:45, sail wrote:
> > > On 2012/04/16 19:31:25, Fady Samuel wrote:
> > > > On 2012/04/16 17:31:21, sail wrote:
> > > > > Hi Fady. On Windows do we need to mark all processes as HiDPI aware or
> > just
> > > > the
> > > > > browser process?
> > > > 
> > > > Hi, right now to enable high DPI in WebKit, you would use the
> > > > --default-device-scale-factor=x where x > 1. This needs to be passed to
> all
> > > the
> > > > renderers (and is). However, this flag currently exists for debugging
> > purposes
> > > > and things may be done differently a little later.
> > > 
> > > Hi Fady. I'm asking a slightly different question.
> > > On Windows, does the renderer process need to be HiDPI aware? If a process
> is
> > > not marked as HiDPI aware then when it calls system APIs everything is
given
> > in
> > > logical coordinates. So for example, mouse coordinates might be different
> > > between the browser and the renderer.
> > 
> > I haven't really played with High DPI on Windows, unfortunately so I'm not
> sure
> > of the answer to your question. I want to say yes, since WebKit is High-DPI
> > aware, the renderers should be marked as high-DPI aware, but I'm not sure,
> > sorry.
> 
> Sounds good. I think it's probably worth landing as is just so we can start
> testing things. We can always move this out to the exe_main code if needed.

can you please confirm before landing? it seems easy to do so (i.e. you can even
just get the process-type switch and only do it if it's empty)

[sail, 2012-04-17 03:21:00]

> can you please confirm before landing? it seems easy to do so (i.e. you can
even
> just get the process-type switch and only do it if it's empty)

Hi John, I'm not sure how to confirm before landing.

[jam, 2012-04-17 04:34:41]

On 2012/04/17 03:21:00, sail wrote:
> > can you please confirm before landing? it seems easy to do so (i.e. you can
> even
> > just get the process-type switch and only do it if it's empty)
> 
> Hi John, I'm not sure how to confirm before landing.

see a few lines above where you added for an example of checking when you're in
the browser process:

if (!process_type.empty())
  ui::EnableHighDPISupport();

[commit-bot: I haz the power, 2012-04-24 23:06:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sail@chromium.org/10086009/2001

[commit-bot: I haz the power, 2012-04-25 00:56:15]

Change committed as 133826