Compute the base dpi from X

Compute the base dpi from X

X uses 96 as default, and alters the screen size
but it can be overwritten. Compute the dpi from X
to match what gtk sues as base.

Round the dsf to 1 decimals as finer grained value can cause rendering problem than benefit.

Gpu test expectation on nvidia needs to be updated, so mark these tests as failing now.

BUG=485183
Committed: https://crrev.com/5cb0c01e4233ff3eb1e7f6f677b4cf9dae9ffaf2
Cr-Commit-Position: refs/heads/master@{#330212}

[oshima, 2015-05-13 21:45:31]

Patchset #1 (id:1) has been deleted

[oshima, 2015-05-13 21:48:05]

oshima@chromium.org changed reviewers:
+ estade@chromium.org, mdempsky@chromium.org

[oshima, 2015-05-13 21:52:00]

Patchset #1 (id:20001) has been deleted

[oshima, 2015-05-13 22:00:45]

estade@ for owners review

mdempsky@, would you mind testing this patch and see if it fixes the issue on
your environment?

[Evan Stade, 2015-05-13 22:09:28]

https://codereview.chromium.org/1136913005/diff/60001/chrome/browser/ui/libgt...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/60001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:386: return
(static_cast<double>(DisplayWidth(xdisplay, xscreen)) * 25.4) /
is this casting truly necessary? 25.4 is already a double, so I think everything
else will be converted to a double

[oshima, 2015-05-13 22:21:24]

https://codereview.chromium.org/1136913005/diff/60001/chrome/browser/ui/libgt...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/60001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:386: return
(static_cast<double>(DisplayWidth(xdisplay, xscreen)) * 25.4) /
On 2015/05/13 22:09:27, Evan Stade wrote:
> is this casting truly necessary? 25.4 is already a double, so I think
everything
> else will be converted to a double

I was lazy, sorry. (was cut&paste) Fixed.

[Evan Stade, 2015-05-13 22:24:42]

I'll owner stamp after mdempsky review

[mdempsky, 2015-05-13 22:26:47]

I'll try it out.

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:386: return (DisplayWidth(xdisplay,
xscreen) * 25.4) /
gdk uses DisplayHeight and DisplayHeightMM, not DisplayWidth and DisplayWidthMM.
 It probably doesn't matter in practice, but seems like we should be consistent
with them.

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:1437: float scale = GetFontDPI() /
GetBaseDPI();
This formulation seems odd to me.  If gtk-xft-dpi isn't set (which I would
expect to be the common case), you're calculating GetBaseDPI() / GetBaseDPI(),
which will always evaluate to 1.0, but will require extra calls to DisplayWidth,
etc.

[oshima, 2015-05-13 22:40:33]

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:386: return (DisplayWidth(xdisplay,
xscreen) * 25.4) /
On 2015/05/13 22:26:46, mdempsky wrote:
> gdk uses DisplayHeight and DisplayHeightMM, not DisplayWidth and
DisplayWidthMM.
>  It probably doesn't matter in practice, but seems like we should be
consistent
> with them.

good point. done.

https://codereview.chromium.org/1136913005/diff/80001/chrome/browser/ui/libgt...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:1437: float scale = GetFontDPI() /
GetBaseDPI();
On 2015/05/13 22:26:46, mdempsky wrote:
> This formulation seems odd to me.  If gtk-xft-dpi isn't set (which I would
> expect to be the common case), you're calculating GetBaseDPI() / GetBaseDPI(),
> which will always evaluate to 1.0, but will require extra calls to
DisplayWidth,
> etc.

gtk-xft-dpi is used to set font scale so this should be available when a user
wants larger fonts, at least on ubuntu / gnome environment. I don't know about
kde/qt, but that's the outside of the scope of this CL, as this class is meant
for gtk integration.

DisplayWidth/MM are macros which just access the members of Display struct so it
should be very fast.

[mdempsky, 2015-05-13 23:14:36]

Looks right to me on my HP Z30i when I disable the "Xft.dpi: 96" workaround. 
This error message gets spammed a bajillion times per second to the console
while I move my mouse around though:

[11619:11619:0513/161402:ERROR:gtk2_ui.cc(1436)] GetBaseDPI:100.969

[oshima, 2015-05-13 23:34:48]

On 2015/05/13 23:14:36, mdempsky wrote:
> Looks right to me on my HP Z30i when I disable the "Xft.dpi: 96" workaround. 
> This error message gets spammed a bajillion times per second to the console
> while I move my mouse around though:
> 
> [11619:11619:0513/161402:ERROR:gtk2_ui.cc(1436)] GetBaseDPI:100.969

oops, sorry. removed. I think caller should cache it if appropriate. I'll look
into in a separate CL.

[oshima, 2015-05-14 22:23:01]

mdempsky: ping?

[mdempsky, 2015-05-14 22:26:06]

lgtm

[Evan Stade, 2015-05-14 22:43:04]

lgtm

https://codereview.chromium.org/1136913005/diff/120001/chrome/browser/ui/libg...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/120001/chrome/browser/ui/libg...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:1437: // Round to 1 decimals, e.g. to
1.4.
nit: 1 decimal

[oshima, 2015-05-14 22:45:13]

https://codereview.chromium.org/1136913005/diff/120001/chrome/browser/ui/libg...
File chrome/browser/ui/libgtk2ui/gtk2_ui.cc (right):

https://codereview.chromium.org/1136913005/diff/120001/chrome/browser/ui/libg...
chrome/browser/ui/libgtk2ui/gtk2_ui.cc:1437: // Round to 1 decimals, e.g. to
1.4.
On 2015/05/14 22:43:04, Evan Stade wrote:
> nit: 1 decimal

Done.

[oshima, 2015-05-14 22:45:23]

The CQ bit was checked by oshima@chromium.org

[oshima, 2015-05-14 22:45:24]

The patchset sent to the CQ was uploaded after l-g-t-m from estade@chromium.org,
mdempsky@chromium.org
Link to the patchset: https://codereview.chromium.org/1136913005/#ps140001
(title: "fix comment")

[commit-bot: I haz the power, 2015-05-14 22:46:29]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1136913005/140001

[commit-bot: I haz the power, 2015-05-15 00:50:17]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-15 00:50:18]

Try jobs failed on following builders:
  linux_chromium_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[oshima, 2015-05-15 01:13:12]

The CQ bit was checked by oshima@chromium.org

[commit-bot: I haz the power, 2015-05-15 01:13:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1136913005/140001

[commit-bot: I haz the power, 2015-05-15 03:14:17]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-15 03:14:18]

Try jobs failed on following builders:
  linux_chromium_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[oshima, 2015-05-15 18:09:48]

oshima@chromium.org changed reviewers:
+ kbr@chromium.org

[oshima, 2015-05-15 18:10:05]

+kbr@ for content/test/gpu/gpu_tests/pixel_expectations.py

[Ken Russell (switch to Gerrit), 2015-05-15 18:18:56]

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
File content/test/gpu/gpu_tests/pixel_expectations.py (right):

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
content/test/gpu/gpu_tests/pixel_expectations.py:19: [ ('nvidia', 0x104a)],
bug=485183)
Actually, since you need to update the revision number of all of these tests,
you'll have to mark them failing for all platforms. In this CL, please also
increment the version numbers for all of these tets in
src/content/test/gpu/page_sets/pixel_tests.py .

We need to keep the duration of how long these are marked failing as short as
possible. Once this is in the CQ, please prepare and send for review another CL
which un-marks these as failing.

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
content/test/gpu/gpu_tests/pixel_expectations.py:20: pass
I think you may need to remove this "pass". Did you try running the tests with
this patch?
https://www.chromium.org/developers/testing/gpu-testing#TOC-Running-the-GPU-T...

If it runs OK then no need to remove it.

[oshima, 2015-05-15 19:37:18]

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
File content/test/gpu/gpu_tests/pixel_expectations.py (right):

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
content/test/gpu/gpu_tests/pixel_expectations.py:19: [ ('nvidia', 0x104a)],
bug=485183)
On 2015/05/15 18:18:56, Ken Russell wrote:
> Actually, since you need to update the revision number of all of these tests,
> you'll have to mark them failing for all platforms. In this CL, please also
> increment the version numbers for all of these tets in
> src/content/test/gpu/page_sets/pixel_tests.py .
> 
> We need to keep the duration of how long these are marked failing as short as
> possible. Once this is in the CQ, please prepare and send for review another
CL
> which un-marks these as failing.

Done.

https://codereview.chromium.org/1136913005/diff/160001/content/test/gpu/gpu_t...
content/test/gpu/gpu_tests/pixel_expectations.py:20: pass
On 2015/05/15 18:18:56, Ken Russell wrote:
> I think you may need to remove this "pass". Did you try running the tests with
> this patch?
>
https://www.chromium.org/developers/testing/gpu-testing#TOC-Running-the-GPU-T...
> 
> If it runs OK then no need to remove it.

Done.

[oshima, 2015-05-15 19:52:59]

The CQ bit was checked by oshima@chromium.org to run a CQ dry run

[oshima, 2015-05-15 19:52:59]

The patchset sent to the CQ was uploaded after l-g-t-m from estade@chromium.org,
mdempsky@chromium.org
Link to the patchset: https://codereview.chromium.org/1136913005/#ps200001
(title: " ")

[oshima, 2015-05-15 19:54:16]

Patchset #10 (id:220001) has been deleted

[Ken Russell (switch to Gerrit), 2015-05-15 20:37:15]

Thanks. LGTM

[oshima, 2015-05-15 20:41:39]

The CQ bit was checked by oshima@chromium.org

[oshima, 2015-05-15 20:41:39]

The patchset sent to the CQ was uploaded after l-g-t-m from estade@chromium.org,
mdempsky@chromium.org
Link to the patchset: https://codereview.chromium.org/1136913005/#ps200001
(title: " ")

[commit-bot: I haz the power, 2015-05-15 20:42:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1136913005/200001

[commit-bot: I haz the power, 2015-05-15 22:16:19]

Committed patchset #9 (id:200001)

[commit-bot: I haz the power, 2015-05-18 11:28:15]

Patchset 9 (id:??) landed as
https://crrev.com/5cb0c01e4233ff3eb1e7f6f677b4cf9dae9ffaf2
Cr-Commit-Position: refs/heads/master@{#330212}