Just to throw a bit more weight on this; I use windows and linux on my Retina mac, and this seems to be a UI issue regardless of window manager (even is an issue in windows if you enable "Windows XP scaling".

I think it's most likely because the themes for linux/windows don't have the high-dpi GUI elements included. I'd think the work for OSX would mostly be transferable...?

No HiDpi Support is one of the bigger blocker for using Linux on a Retina Macbook.

Using site scaling in the options is no real solution since websites / apps behaves different sometimes.

I don't really expect this to be solved prior to high-dpi screens becoming more common, short of doing it myself.

..I'd really quite like to do it myself. Sadly, I'm completely unfamiliar with the Chromium codebase. Could anyone give me a few hints of where to start?

Just got a Dell XPS 13 Ubuntu laptop.  I can barely read the titles of tabs and other icons.  While being able to set the global default scaling is great, it still does not fix the problem of small UI elements.

I've now figured out that the bookmark bar obeys the GTK font size (and icon size?) (if using the GTK theme). However, the tabs still do not. So this bug may be a duplicate of #146291.

The code for this has to exist, to some degree. It works on the Chromebook Pixel, right?

On the Chromebook pixel running stock Ubuntu with the dpi settings turned up, it doesn't seem to work properly. Screenshot attached.

	pixel-screenshot.png 2.6 MB View Download

That's exactly what I get on ubuntu and arch on a macbook pro retina.

I think the difference between ChromeOS and stock linux is that ChromeOS has deeper integration with X than normal Chromium. I wouldn't expect it to 'just work' because ChromeOS can do it.

Still, it _is_ doable, and I'd love to see this fixed... using anything but OS X on my Retina's an exercise in squinting. (And as such, I use other browsers. Opera's surprisingly less bad about this.)

I'm facing the same issue on the Chromebook Pixel running Ubuntu 13.04. It's deplorable that Google uses Linux & Chromium as a base for their ChromeOS and has proper high DPI support but the OpenSource versions are left without such basic support. 

ChromeOS doesn't even use X11. It uses its own window manager called Aura. I think short-term X11 wasn't up to the task of supporting multi-touch display and also variable DPI on multiple screens.
Aura handles both these tasks admirably. But it's very unfortunate that X11 still does not support proper multi-touch nor proper DPI across multiple displays (plug in an external monitor to your ChromeOS and see how well it handles everything, it's very impressive!)

Thanks for the explanation. I did not know ChromeOS was using a different X server. Google needs to provide some explanation around it. Is it so difficult to get the multi-touch capabilities into X11? On a related topic, they have made changes to the Linux kernel 3.4 which haven't yet been pushed upstream (we're upto 3.10-rc2 now). This makes it very difficult to work with Pixel on vanilla Linux. And require a bunch of workarounds to make it functional. Thankfully those workarounds at least make Pixel useable under Linux.

A couple of comments.

First, ChromeOS does use X11, but has a custom window manager, Aura. Second, X can definitely support what we need to get proper dpi scaling (look at the recent work going on in KDE to solve problems with high-dpi screens).

Third, let's not turn this into a discussion about the PIxel; it's entirely beside the point; Chromium on Linux does not scale it's DPI properly, we we want it fixed. 

Can we please get some sort of a stance from the developers on this?

Issue 108335 has been merged into this issue.

I ran a simple test on Linux Chromium 28.0.1500.71 (209842), by changing the DPI with "xrandr --dpi 300".

These are the problems I saw:
- size of 100% zoom for webpages is not affected by DPI (it should scale directly)
- tab bar has a fixed height (see Issue 17866)
- icons in the "main bar" (to the left and right of the omnibar) are not scaled
- icons in the main menu (zoom, settings, exit, etc.) are not scaled
- omnibar dropdown rows have a fixed height, causing text to be cut off
- scrollbars have fixed width
- UI lines are still fixed width (eg. 1px underline of webpage text, 1px line separator in context menus, 1px omnibox border) (maybe this is OK though)
- zoom popup extends off the right edge of the screen

Here's what seems to be working:
- height of omnibar and omnibar text size
- all context menu text (including bookmarks menu)

As far as I can tell from looking at the code and reading around, HiDPI will only be supported on Linux through the Aura framework. Which means we'll have to wait until Aura is stable on Linux proper (issue 125106).

I compiled chromium with Aura turned on and was able to get it to properly scale content on a hidpi screen (after forcing the device scale) but there were a slew of other issues that rendered it unusable.

In 2013 ASUS, Samsung, Fujitsu and Acer released ultrabook with HiDPI support, so it is not only MacBook problem. Now only this Chrome issue block me to moving to Asus Zenbook UX301L :(.

There are two bits that needs to be fixed/implemented in order to get high DPI working on linux aura.

1) Hook up xrandr -> gfx::Display::device_scale_factor()
2) Enable 2x resource on linux aura.

I'm not sure which (or if) scale factor we want to support on linux-aura though (we support only fixed set of scale factors).

Can you at least start with *something*, e.g. a command-line flag only
supporting exactly 2x?

I think a flag to tune the scale will remain useful even after you figure
out good default from xrandr - some people have bad eyesight, some don't
mind squinting to fit 20 untruncated tab titles on their screen...  (HiDPI
screens have a wider range of usable sizes, since tiny font != entirely
unredable.)

Unfortunately the ways to customize system-wide UI density on linux are a
mess, badly exposed and badly working.  There is X DPI (requires editing
Xorg.conf), there are diffent (frequently hidden - e.g. gnome-tweak-tool)
text scaling options per toolkit/desktop, one can just tweak fonts...
I doubt you will be able to automatically discern the desired size in all
cases.
The best bet is maybe respecting Chrome's font settings - whatever the
default site font ends up (whether inherited from system fonts or
customized in settings) is a reasonable indication of user intent for UI
font.

GNOME 3.10 has full HiDPI support. Maybe their config will be useful too?

There is already a flag that is mentioned in #18. (--force-device-scale-factor=2)
Unfortunately, it will not utilize 2x resources which is required for chrome to work correctly in 2x DSF. There is a fallback mechanism to use 1x images, although there are many special cases where this fallback may not work. If you want to get this work, you need 2) at least.

Is there a flag in Chrome version 33 to enable HiDPI support in Fedora 20 (Gnome 3.10) on MBP 13 Retina?

I played with it a bit (with 200 assets). There are at least two issues with high DPI.

1) Some 200dpi assets are missing on linux, which is causing visual issues.
2) popup location/size are treating DIP as native. I guess native <-> DIP conversion is not implemented properly on linux aura. I was hoping we can simply rely on Win high DPI work but this seems to be platform specific?

I also noticed that chrome is using different assets for back/forward/reload buttons. I wonder if this because it's using linux native theme?

In any case, I'll land the patch to get enable_hidpi=1 working. I don't know how much work we need to get everything working though.

Oshima, thank! :)

Yeap, on Linux Chromium use default icons. But on HiDPI we should have HiDPI native icons, so question is only, what icon size we request from OS.

@25: Under the GTK theme, assets are being generated in chrome/browser/ui/libgtk2ui/gt2_ui.cc. Take a look at things like Gtk2UI::GeneratedTintedIcon() or Gtk2UI::GenerateGTKIcon(). These things work by setting up a Gtk widget tree and then scraping the X11 bits and returning them to chrome.

I suspect that if you go under settings and select Use Classic Theme, things will look a lot better under HiDPI.

#27, yep, that fixed the resource issue. thanks!

------------------------------------------------------------------------
r244302 | oshima@chromium.org | 2014-01-11T01:59:50.414738Z

Changed paths:
   M http://src.chromium.org/viewvc/chrome/trunk/src/ui/base/resource/resource_bundle.cc?r1=244302&r2=244301&pathrev=244302

Include 2x assets when enable_hidpi=1 is specified on linux_aura

BUG=143619
R=erg@chromium.org

Review URL: https://codereview.chromium.org/132183004
------------------------------------------------------------------------

Hi guys,
After enabling hidpi it looks almost good, but I have two issues.

1. Open new is tab hidden by strange red square
2. All menus are not rendered properly, it seems that size for them is incorrectly calculated.

Version 34.0.1788.0 aura (245082)

	Screenshot from 2014-01-21 21:24:52.png 274 KB View Download

Same here with 34.0.1809.0 dev aura

RE: #30 - It appears that we're missing a resource from the new tab button (red spot)

My view of the default GTK theme is vastly different, I have Gnome 3.10 with retina support fully working. I ran chrome with "google-chrome-unstable --enable_hidpi=1"

What gives? I have 34.1825.4dev

	hidpi.png 53.4 KB View Download

34.1825.4 disabled aura since we aren't going to make it into M34. It will be back with the first M35 dev channel.

We'll work on linux-aura's high DPI support once win-aura high DPI support is done.

Re #34: I used Google Version 34.0.1825.4 dev, Gnome 3.10.2, Archlinux. The tab is still very small.

Tabs are large if compiled with "-Denable_hidpi=1 -Duse_aura=1", so if you want hidpi mode the only way to get it in archlinux is to use https://aur.archlinux.org/packages/chromium-dev with specified options. And as I said above It still has at least two issues: all popup menus are broken, and new tab button is hidden by a red square.

after change the ebuild file "    #enable Hidpi
    myconf+=" -Denable_hidpi=1" "
I emerge chromium  35.0.1862.2

Tabs are large  but all popup menus are broken

the issue is  always in 35.0.1897.2
nobody  to solve this problem?

Same here. I think a lot of users are waiting for HighDPI support on linux desktop. Or at least some kind of workaround.

I just bought a Yoga 2 Pro which has a 3200x1800 resolution in a 13" screen and this issue is really killing me. For now I've set the default zoom to 250% and the text is relatively good. The only issue after settings the zoom to 250% is the images of the UI that didn't scale at all. So I have an image of 16px next to a text with a font-size of 16*2.5 (the favicons, extensions etc).

Firefox has a perpx setting.  I'm confused as to why Chrome/Chromium doesn't have the same option.   Sooner that later would be good as the hidpi screens are becoming more prevalent.

I installed 35.0.1916.27 beta aura on Fedora 20, but neither the UI nor page content is appropriately scaled. I noticed that my xdpyinfo originally reported my dpi as 96x96, so I set my dpi manually via xrandr. I have attached my xpdyinfo output.

xdpyinfo.txt 6.7 KB View Download

Hidpi only works (in a way), if you compile it with -Denable_hidpi=1. By default it is turned off, but it will break all menus if enabled.

I did not compile chrome, I installed the linux beta rpm. Is hidpi support going to make it into 35?

@46: No. Getting hidpi working requires getting aura shipped, and we're focusing on that right now.

Could you point us to the Aura issue so we can track it's progress?

Linux Aura is tracked as Issue 125106. However, it is now marked as Type-Launch, which is supposed to require some sort of permission to see.

Just FYI, here is the list of issues which is relevant to and/or required for Aura transition on Linux.
https://code.google.com/p/chromium/issues/list?can=2&q=OS%3DLinux+Proj%3DDesktopAura+&sort=-modified

Issue 146291 has been merged into this issue.

Seems like Ubuntu has fixed it in their packaged Chromium in 14.04. It looks much better than Chrome.

To say it is "fixed" in Ubuntu is generous.

My drop-down menus appear half-way across the screen from where they should be and are sized so as to not be able to read more than the first three characters of each entry.

But, yes, the chrome is high res and the fonts are right.

UI is scaled properly in Chromium (but not in Chrome stable/unstable), but the dropdowns are completely broken.

Reporting the same problem with menus and drop-down things, they're in wrong positions and sizes.

Also it doesn't properly detect the DPI of the main (external) screen, or of the other screen (built-in but configured in a lower resolution); see attached screenshot.

Is there a command line switch to override the DPI detection?

	Screenshot from 2014-06-03 12:33:31.png 455 KB View Download

Oshima, is the current plan here to automatically use high-DPI assets (and scale UI fonts, and maybe also Blink fonts, accordingly) when the system API (as reported by GtkSetting's gtk-xft-dpi property) is above some threshold?

Yes. Windows is using 1.25 and I'm planning to do the same.

I took a look at the state of what's at head with -Denable_hidpi=1.

1. Fonts get double-scaled with --force-device-scale-factor > 1. GetFontDPI  https://code.google.com/p/chromium/codesearch#chromium/src/chrome/browser/ui/libgtk2ui/gtk2_ui.cc&q=getfontdpi&sq=package:chromium&type=cs&l=681 is looking at gtk-xft-dpi directly. I guess it should instead be (only) ui::gfx::Display that scales based on that. I don't know if it that's configurable per-display? but I guess doing one global scale at first is still an improvement over what we have now. (Windows doesn't deal with multiple DPIs yet either, though mostly because it's a Win81+ API. I don't know if higher level code will work properly if we try to support that.)

2. The 2x resource bundle always gets loaded when ENABLE_HIDPI. I think that should be only when > 1x. I'm not sure if the bundle is getting built properly either.

3. Windows are being created at the wrong location (popups, the main browser window when re-opened). It looks like points<->pixels isn't handled so hopefully this is just a matter of tracking down the places where that needs to be considered. Not sure which the system is expecting when > 1x.

I'll try to write patches for those. (I barely know a window manager from an ubuntu, so feel free to correct misconceptions.)

https://codereview.chromium.org/659883002/ sort of works (fixes window locations/sizes, tabstrip size, and font sizes) if anyone wants to patch it in and see what's still broken.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/13b1371303258fe761e907d5874cdfdf8e3ed082

commit 13b1371303258fe761e907d5874cdfdf8e3ed082
Author: scottmg <scottmg@chromium.org>
Date: Thu Oct 23 05:07:37 2014

remove IsInHighDPIMode

Some simplifications before using this code on Linux.

R=sky@chromium.org
BUG=143619

Review URL: https://codereview.chromium.org/669213004

Cr-Commit-Position: refs/heads/master@{#300840}

[modify] https://chromium.googlesource.com/chromium/src.git/+/13b1371303258fe761e907d5874cdfdf8e3ed082/ui/gfx/screen_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/13b1371303258fe761e907d5874cdfdf8e3ed082/ui/gfx/win/dpi.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/13b1371303258fe761e907d5874cdfdf8e3ed082/ui/gfx/win/dpi.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/13b1371303258fe761e907d5874cdfdf8e3ed082/ui/views/widget/desktop_aura/desktop_window_tree_host_win.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048

commit 23a0f218f7126e3abc8ce39ca9986411a465d048
Author: scottmg <scottmg@chromium.org>
Date: Thu Oct 23 16:56:21 2014

Remove IsDIPEnabled from Screen

Appears to be unused.

TBR=ben@chromium.org
R=sky@chromium.org
BUG=143619

Review URL: https://codereview.chromium.org/672823002

Cr-Commit-Position: refs/heads/master@{#300906}

[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ash/display/screen_ash.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ash/display/screen_ash.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/athena/env/athena_env_impl.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/chrome/browser/android/shortcut_helper_unittest.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/chrome/browser/extensions/api/system_display/system_display_apitest.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/chrome/browser/ui/window_sizer/window_sizer_common_unittest.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/content/browser/media/capture/web_contents_video_capture_device.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/content/browser/media/capture/web_contents_video_capture_device_unittest.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/extensions/browser/api/capture_web_contents_function.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/mojo/aura/screen_mojo.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/mojo/aura/screen_mojo.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/aura/test/test_screen.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/aura/test/test_screen.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/base/layout.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen_android.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen_ios.mm
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen_mac.mm
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/gfx/screen_win.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/views/widget/desktop_aura/desktop_screen_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/views/widget/desktop_aura/desktop_screen_win.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/views/widget/desktop_aura/desktop_screen_x11.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/23a0f218f7126e3abc8ce39ca9986411a465d048/ui/views/widget/desktop_aura/desktop_screen_x11.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/aaa54ca8dc57118f0f81cf51198adb879ca6bfcc

commit aaa54ca8dc57118f0f81cf51198adb879ca6bfcc
Author: scottmg <scottmg@chromium.org>
Date: Thu Oct 23 18:57:35 2014

remove ui/base/win/dpi_setup.cc

Unnecessary wrapper.

TBR=jam@chromium.org
R=sky@chromium.org
BUG=143619

Review URL: https://codereview.chromium.org/672583003

Cr-Commit-Position: refs/heads/master@{#300923}

[modify] https://chromium.googlesource.com/chromium/src.git/+/aaa54ca8dc57118f0f81cf51198adb879ca6bfcc/content/app/content_main_runner.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/aaa54ca8dc57118f0f81cf51198adb879ca6bfcc/ui/base/BUILD.gn
[modify] https://chromium.googlesource.com/chromium/src.git/+/aaa54ca8dc57118f0f81cf51198adb879ca6bfcc/ui/base/resource/resource_bundle.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/aaa54ca8dc57118f0f81cf51198adb879ca6bfcc/ui/base/ui_base.gyp
[delete] https://chromium.googlesource.com/chromium/src.git/+/b78241aa63cc312c4c9fbee134347e75c31e2333/ui/base/win/dpi_setup.cc
[delete] https://chromium.googlesource.com/chromium/src.git/+/b78241aa63cc312c4c9fbee134347e75c31e2333/ui/base/win/dpi_setup.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd

commit 67a57736b1cd516ff05116a5c27c4c5237f151fd
Author: scottmg <scottmg@chromium.org>
Date: Thu Oct 23 19:47:50 2014

remove ui::gfx::GetDeviceScaleFactor

GetDPIScale() handles overrides, having two accessors is extra confusion.

R=sky@chromium.org
BUG=143619

Review URL: https://codereview.chromium.org/674763002

Cr-Commit-Position: refs/heads/master@{#300931}

[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/chrome/browser/icon_loader_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/chrome/browser/ui/views/apps/app_window_desktop_window_tree_host_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/chrome/browser/ui/views/frame/minimize_button_metrics_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/base/layout.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/gfx/canvas_paint_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/gfx/screen_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/gfx/win/dpi.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/gfx/win/dpi.h
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/views/controls/menu/menu_separator_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/67a57736b1cd516ff05116a5c27c4c5237f151fd/ui/views/win/hwnd_message_handler.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34

commit 2525dae3107f78a01bf0f7f623db630c93375c34
Author: scottmg <scottmg@chromium.org>
Date: Thu Oct 23 23:43:03 2014

Remove IsHighDPIEnabled, move EnableHighDPISupport to only place it's used

R=sky@chromium.org
BUG=143619

Review URL: https://codereview.chromium.org/672673005

Cr-Commit-Position: refs/heads/master@{#300986}

[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ash/display/display_info.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/chrome/app/chrome_exe_main_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/base/layout.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/base/resource/resource_bundle.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/base/resource/resource_bundle_unittest.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/base/resource/resource_bundle_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/base/test/run_all_unittests.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/gfx/screen_win.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/gfx/win/dpi.cc
[modify] https://chromium.googlesource.com/chromium/src.git/+/2525dae3107f78a01bf0f7f623db630c93375c34/ui/gfx/win/dpi.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/aa051ee8d1a7ff725225d783d1a96663d5aedb08

commit aa051ee8d1a7ff725225d783d1a96663d5aedb08
Author: stapelberg <stapelberg@google.com>
Date: Thu Mar 19 22:40:36 2015

Fix Pango font rendering with HiDPi displays on Linux.

Don’t scale fonts up twice when running with device_scale_factor > 1.0.

Chrome’s UI is scaled up to device_scale_factor (e.g. 2x) by calling
SkCanvas->scale(). See Chrome’s gfx::Canvas class which calls
canvas.scale(image_scale, image_scale):
https://code.google.com/p/chromium/codesearch#chromium/src/ui/gfx/canvas.cc

Here is a skia fiddle which renders text with .setTextSize(21):
https://fiddle.skia.org/c/098cb32edac7e394aba5b9ff4c08b06c

Here is the same fiddle, but calling canvas.Scale(2.0, 2.0) before rendering to simulate what Chrome does:
https://fiddle.skia.org/c/a7b456338776239278e398741e1098d9

As you can see, the text is twice as large.

Therefore, this commit changes Chrome to take into account the DPI that
GTK provides, but set it in perspective to the device scale factor. This
preserves the same behavior when device scale factor == 1 but yields
correct rendering when device scale factor > 1.0 (hi-dpi displays).

Before this change:
http://t.zekjur.net/chrome-hidpi-enabled.png

After this change:
http://t.zekjur.net/chrome-hidpi-enabled-scalefix.png

This change also enables subpixel rendering on HiDPi displays, which is what Chrome OS also does.
Without subpixel rendering:
http://t.zekjur.net/without-subpixel.png

With subpixel rendering:
http://t.zekjur.net/with-subpixel.png

BUG=143619

Review URL: https://codereview.chromium.org/929733002

Cr-Commit-Position: refs/heads/master@{#321457}

[modify] http://crrev.com/aa051ee8d1a7ff725225d783d1a96663d5aedb08/chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc
[modify] http://crrev.com/aa051ee8d1a7ff725225d783d1a96663d5aedb08/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/aa051ee8d1a7ff725225d783d1a96663d5aedb08/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/aa051ee8d1a7ff725225d783d1a96663d5aedb08/ui/gfx/font_render_params_linux.cc
[modify] http://crrev.com/aa051ee8d1a7ff725225d783d1a96663d5aedb08/ui/views/linux_ui/linux_ui.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/53630bda6cd2145038f20108d13fb1cf4b979472

commit 53630bda6cd2145038f20108d13fb1cf4b979472
Author: stapelberg <stapelberg@google.com>
Date: Mon Mar 23 16:18:38 2015

Fix tab drag and drop with hi-dpi.

This change converts between DIP and pixels in a few places that were
not converted before, but are necessary for drag & drop of tabs to work
correctly.

R=sadrul@chromium.org
BUG=458804,143619

Review URL: https://codereview.chromium.org/1003863005

Cr-Commit-Position: refs/heads/master@{#321787}

[modify] http://crrev.com/53630bda6cd2145038f20108d13fb1cf4b979472/chrome/browser/ui/views/tabs/window_finder_x11.cc
[modify] http://crrev.com/53630bda6cd2145038f20108d13fb1cf4b979472/ui/views/widget/desktop_aura/desktop_screen_x11.cc
[modify] http://crrev.com/53630bda6cd2145038f20108d13fb1cf4b979472/ui/views/widget/desktop_aura/x11_topmost_window_finder.cc
[modify] http://crrev.com/53630bda6cd2145038f20108d13fb1cf4b979472/ui/views/widget/desktop_aura/x11_topmost_window_finder.h

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/a27daaff43d65055a62aee8aca3415901ec3eb33

commit a27daaff43d65055a62aee8aca3415901ec3eb33
Author: stapelberg <stapelberg@google.com>
Date: Mon Mar 23 21:47:51 2015

Enable hi-dpi support on Linux.

With the recent bugfixes (see BUG=), hi-dpi support on Linux works well enough
to enable it by default.

BUG=143619

Review URL: https://codereview.chromium.org/1028703004

Cr-Commit-Position: refs/heads/master@{#321865}

[modify] http://crrev.com/a27daaff43d65055a62aee8aca3415901ec3eb33/build/common.gypi
[modify] http://crrev.com/a27daaff43d65055a62aee8aca3415901ec3eb33/build/config/ui.gni

Font rendering, as of this screenshot, is not fixed:

> Comment #71
> With subpixel rendering:
> http://t.zekjur.net/with-subpixel.png

Notice how the kerning is off in "Welcome to Chrome", especially in the "me" of "Welcome".  It looks more like "Welcom e  to".  (ChromeOS also has this problem.)

I tried these patches on Linux (actually, Google Chrome 43.0.2327.5 dev, so perhaps a bit out of date) and can reproduce the kerning problem.

(Issue 415929 is the related ChromeOS issue.)

	Screenshot from 2015-03-23 20:26:22.png 6.5 KB View Download

jrockway: The UI and Blink text-rendering paths are (almost) completely separate. The change that you're referring to updated the UI. I think that https://codereview.chromium.org/1023343003/ updates Blink.

stapelberg@ asked me to mention that the --enable-webkit-text-subpixel-positioning command-line flag can be used as a workaround for the Blink hinting issue until Oshima's fix to enable subpixel positioning automatically in Blink for DSF>1.0 is in.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/717c2773a86fb7ec73c3b77e9048fd541c3cf876

commit 717c2773a86fb7ec73c3b77e9048fd541c3cf876
Author: stapelberg <stapelberg@google.com>
Date: Sun Mar 29 15:12:02 2015

hi-dpi: convert _NET_WORKAREA from pixels.

Just like the gfx::Display member |bounds_| is not stored in pixels, we need to
convert |work_area_| from pixels.

Both are used in MenuController::UpdateInitialLocation (|work_area_| is
preferred over |bounds_|), which gets events with already converted
coordinates.

BUG=469373,143619

Review URL: https://codereview.chromium.org/1036263003

Cr-Commit-Position: refs/heads/master@{#322724}

[modify] http://crrev.com/717c2773a86fb7ec73c3b77e9048fd541c3cf876/ui/views/widget/desktop_aura/desktop_screen_x11.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/442a5a6b163e6f9f7d798b5501c5143367e5c8c7

commit 442a5a6b163e6f9f7d798b5501c5143367e5c8c7
Author: stapelberg <stapelberg@google.com>
Date: Tue Mar 31 19:22:31 2015

Force --enable-webkit-text-subpixel-positioning for DSF > 1.0.

We already do this for fonts in the UI, and we should also do this for fonts in
the actual web content (i.e. everywhere where chromium falls back on an
operating system font, e.g. when specifying just “monospace”).

before this change:
http://t.zekjur.net/2015-03-21-003244_2357x1809_scrot.png

after this change:
http://t.zekjur.net/2015-03-21-003308_2357x1775_scrot.png

BUG=396659,143619

Review URL: https://codereview.chromium.org/1020263004

Cr-Commit-Position: refs/heads/master@{#323085}

[modify] http://crrev.com/442a5a6b163e6f9f7d798b5501c5143367e5c8c7/chrome/browser/chromeos/login/chrome_restart_request.cc
[modify] http://crrev.com/442a5a6b163e6f9f7d798b5501c5143367e5c8c7/ui/gfx/font_render_params_linux.cc
[modify] http://crrev.com/442a5a6b163e6f9f7d798b5501c5143367e5c8c7/ui/gfx/switches.cc
[modify] http://crrev.com/442a5a6b163e6f9f7d798b5501c5143367e5c8c7/ui/gfx/switches.h

The google-chrome-unstable linux package 43.0.2351.3 contains most of the the commits (all except for the very latest one that makes the --enable-webkit-text-subpixel-positioning flag unnecessary), but doesn’t contain the chrome_200_percent.pak resource file.

I’m working on fixing that.

I gave it a try with that version and this invocation:

$ google-chrome-unstable  --force-device-scale-factor=1.50 --enable-webkit-text-subpixel-positioning

Looks good to me.  100% usable.  Font rendering looks better than CrOS 43 dev.  Yes, things like the "SSL OK" icon look a tiny bit blocky, as I'd expect from the missing 200% assets.

The bug tracker is probably the wrong place to say "nice work", but nice work :)

Actually, wouldn't you know it... Inbox has blurry fonts, but only at 1.5x, not at, say 5x.  Other random websites (G+, this site) look just fine.

(Screenshots attached.)

	Screenshot from 2015-04-01 00:16:54.png 40.2 KB View Download

	Screenshot from 2015-04-01 00:20:50.png 35.1 KB View Download

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/081ef20d521321c62353126088194624b424b1e1

commit 081ef20d521321c62353126088194624b424b1e1
Author: stapelberg <stapelberg@google.com>
Date: Wed Apr 01 06:25:32 2015

Include chrome_200_percent.pak in Linux builds.

Verified this by building the linux_packages_stable target, and now chrome_200_percent.pak is indeed included.

BUG=143619

Review URL: https://codereview.chromium.org/1048123004

Cr-Commit-Position: refs/heads/master@{#323193}

[modify] http://crrev.com/081ef20d521321c62353126088194624b424b1e1/chrome/installer/linux/common/installer.include
[modify] http://crrev.com/081ef20d521321c62353126088194624b424b1e1/chrome/tools/build/linux/FILES.cfg

jrockway, can you provide screenshots of a working website and a non-working website?

Is the problem related to a specific font-family?

Also, I think you should open a separate issue for this.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/575097ce8654423b32a15a6cd630398beb0734d3

commit 575097ce8654423b32a15a6cd630398beb0734d3
Author: stapelberg <stapelberg@google.com>
Date: Mon Apr 06 18:31:38 2015

Remove obsolete for_web_contents parameter in FontRenderParamsQuery.

BUG=143619

Review URL: https://codereview.chromium.org/1063603003

Cr-Commit-Position: refs/heads/master@{#323920}

[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/android_webview/native/aw_settings.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ash/display/display_manager_unittest.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/chrome/browser/renderer_preferences_util.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/content/browser/renderer_host/sandbox_ipc_linux.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/font_render_params.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/font_render_params.h
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/font_render_params_linux.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/font_render_params_linux_unittest.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/platform_font_linux.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/platform_font_mac.mm
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/platform_font_win.cc
[modify] http://crrev.com/575097ce8654423b32a15a6cd630398beb0734d3/ui/gfx/render_text_harfbuzz.cc

I just installed and tested today’s google-chrome-unstable release (version 43.0.2357.2) and all my fixes are included and seem to be working fine.

I’m closing this issue. For any further issues with hi-dpi on linux, please create new issues.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9

commit 06cda64742b45e1eb10bbbb66bca4b4da2efb3c9
Author: stapelberg <stapelberg@google.com>
Date: Sat Apr 11 14:08:50 2015

device scale factor detection: use gtk-xft-dpi consistently

Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
a more or less random display. This CL consistently uses gtk-xft-dpi which
is the better alternative: users can directly influence it (by changing the
Xft.dpi X resource, or via their gtkrc).

See also https://github.com/derat/font-config-info for displaying your
current system settings.

BUG=473089,143619

Review URL: https://codereview.chromium.org/1070433002

Cr-Commit-Position: refs/heads/master@{#324782}

[modify] http://crrev.com/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9/ui/views/widget/desktop_aura/desktop_screen_x11.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/5d03b6013e39b55de40a6789991f6240a79fd2f4

commit 5d03b6013e39b55de40a6789991f6240a79fd2f4
Author: timloh <timloh@chromium.org>
Date: Mon Apr 13 03:00:56 2015

Revert of device scale factor detection: use gtk-xft-dpi consistently (patchset #6 id:100001 of https://codereview.chromium.org/1070433002/)

Reason for revert:
This patch makes content_shell segfault during startup.

Program received signal SIGSEGV, Segmentation fault.
0x00000000049d190a in views::DesktopScreenX11::BuildDisplaysFromXRandRInfo (this=0x35ed099f9200) at ../../ui/views/widget/desktop_aura/desktop_screen_x11.cc:291
291           views::LinuxUI::instance()->GetDeviceScaleFactor();
(gdb) bt
#0  0x00000000049d190a in views::DesktopScreenX11::BuildDisplaysFromXRandRInfo (this=0x35ed099f9200) at ../../ui/views/widget/desktop_aura/desktop_screen_x11.cc:291
#1  0x00000000049d096f in views::DesktopScreenX11::DesktopScreenX11 (this=0x35ed099f9200) at ../../ui/views/widget/desktop_aura/desktop_screen_x11.cc:107
#2  0x00000000049d1fab in views::CreateDesktopScreen () at ../../ui/views/widget/desktop_aura/desktop_screen_x11.cc:370
#3  0x0000000000451f28 in content::Shell::PlatformInitialize (default_window_size=...) at ../../content/shell/browser/shell_views.cc:421
#4  0x0000000000440158 in content::Shell::Initialize () at ../../content/shell/browser/shell.cc:146

Original issue's description:
> device scale factor detection: use gtk-xft-dpi consistently
>
> Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
> a more or less random display. This CL consistently uses gtk-xft-dpi which
> is the better alternative: users can directly influence it (by changing the
> Xft.dpi X resource, or via their gtkrc).
>
> See also https://github.com/derat/font-config-info for displaying your
> current system settings.
>
> BUG=473089,143619
>
> Committed: https://crrev.com/06cda64742b45e1eb10bbbb66bca4b4da2efb3c9
> Cr-Commit-Position: refs/heads/master@{#324782}

TBR=sadrul@chromium.org,oshima@chromium.org,erg@google.com,erg@chromium.org,stapelberg@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=473089,143619

Review URL: https://codereview.chromium.org/1079883002

Cr-Commit-Position: refs/heads/master@{#324810}

[modify] http://crrev.com/5d03b6013e39b55de40a6789991f6240a79fd2f4/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/5d03b6013e39b55de40a6789991f6240a79fd2f4/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/5d03b6013e39b55de40a6789991f6240a79fd2f4/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/5d03b6013e39b55de40a6789991f6240a79fd2f4/ui/views/widget/desktop_aura/desktop_screen_x11.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34

commit d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34
Author: stapelberg <stapelberg@google.com>
Date: Mon Apr 13 22:44:50 2015

device scale factor detection: use gtk-xft-dpi consistently

Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
a more or less random display. This CL consistently uses gtk-xft-dpi which
is the better alternative: users can directly influence it (by changing the
Xft.dpi X resource, or via their gtkrc).

See also https://github.com/derat/font-config-info for displaying your
current system settings.

This is a re-do of https://codereview.chromium.org/1070433002/, but with
https://codereview.chromium.org/1084453002/ applied (fixes content_shell).

BUG=473089, 143619

Review URL: https://codereview.chromium.org/1085603002

Cr-Commit-Position: refs/heads/master@{#324928}

[modify] http://crrev.com/d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34/ui/views/widget/desktop_aura/desktop_screen_x11.cc

------------------------------------------------------------------
r294979 | stapelberg@google.com | 2015-04-24T14:26:08.346364Z

Changed paths:
   M http://src.chromium.org/viewvc/chrome/trunk/tools/build/scripts/slave/xvfb.py?r1=294979&r2=294978&pathrev=294979

xvfb: set dpi to 96 (xvfb defaults to 100).

Tests assume that the DPI is effectively 96. Up until recently, this was
the case because chromium only supported 100% (corresponding to 96 dpi)
or 200% (corresponding to 192 dpi) scaling in its UI.

With https://codereview.chromium.org/1078423002/, we want to change that
to better reflect the hardware that many people have today. Applying
that CL without also changing the DPI results in tests failing because
of off-by-a-few-pixel errors.

BUG=477716,143619

Review URL: https://codereview.chromium.org/1105903002
-----------------------------------------------------------------

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b3ca230ddd7d1238ee96ed26ea23e369f10dd655

commit b3ca230ddd7d1238ee96ed26ea23e369f10dd655
Author: stapelberg <stapelberg@google.com>
Date: Fri Apr 24 19:04:13 2015

Allow arbitrary scale factors.

As discussed in https://codereview.chromium.org/1070433002/, this CL
allows arbitrary scale factors, which works because chromium uses the
resource most closely matching the scale factor (e.g. 1x for 1.25x,
but 2x for 1.9x) and scales it.

BUG=143619

Review URL: https://codereview.chromium.org/1078423002

Cr-Commit-Position: refs/heads/master@{#326851}

[modify] http://crrev.com/b3ca230ddd7d1238ee96ed26ea23e369f10dd655/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/b3ca230ddd7d1238ee96ed26ea23e369f10dd655/testing/xvfb.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7cec497feff0d8fa0de1fc59271a29f3a0f3f043

commit 7cec497feff0d8fa0de1fc59271a29f3a0f3f043
Author: Magnus Jedvert <magjed@google.com>
Date: Tue Apr 28 11:42:03 2015

Allow arbitrary scale factors.

As discussed in https://codereview.chromium.org/1070433002/, this CL
allows arbitrary scale factors, which works because chromium uses the
resource most closely matching the scale factor (e.g. 1x for 1.25x,
but 2x for 1.9x) and scales it.

BUG=143619,477716

Review URL: https://codereview.chromium.org/1078423002

Cr-Commit-Position: refs/heads/master@{#326851}
(cherry picked from commit b3ca230ddd7d1238ee96ed26ea23e369f10dd655)

device scale factor detection: use gtk-xft-dpi consistently

Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
a more or less random display. This CL consistently uses gtk-xft-dpi which
is the better alternative: users can directly influence it (by changing the
Xft.dpi X resource, or via their gtkrc).

See also https://github.com/derat/font-config-info for displaying your
current system settings.

This is a re-do of https://codereview.chromium.org/1070433002/, but with
https://codereview.chromium.org/1084453002/ applied (fixes content_shell).

BUG=473089, 143619

Review URL: https://codereview.chromium.org/1085603002

Cr-Commit-Position: refs/heads/master@{#324928}
(cherry picked from commit d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34)

R=stapelberg@google.com
TBR=erg, estade, oshima

Review URL: https://codereview.chromium.org/1103383004

Cr-Commit-Position: refs/branch-heads/2357@{#246}
Cr-Branched-From: 59d4494849b405682265ed5d3f5164573b9a939b-refs/heads/master@{#323860}

[modify] http://crrev.com/7cec497feff0d8fa0de1fc59271a29f3a0f3f043/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/7cec497feff0d8fa0de1fc59271a29f3a0f3f043/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/7cec497feff0d8fa0de1fc59271a29f3a0f3f043/testing/xvfb.py
[modify] http://crrev.com/7cec497feff0d8fa0de1fc59271a29f3a0f3f043/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/7cec497feff0d8fa0de1fc59271a29f3a0f3f043/ui/views/widget/desktop_aura/desktop_screen_x11.cc

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/bling/chromium.git/+/7cec497feff0d8fa0de1fc59271a29f3a0f3f043

commit 7cec497feff0d8fa0de1fc59271a29f3a0f3f043
Author: Magnus Jedvert <magjed@google.com>
Date: Tue Apr 28 11:42:03 2015

+laforge@(M43 owner) for m43 owner.

Hi, we'd like to unmerge the CL in #93 (#323860/7cec497feff0d8fa0de1fc59271a29f3a0f3f043), as this included changes that weren't part of approved changes, and it's causing issue on 43.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1e83acd552618ab337d286187ed23d84434bd122

commit 1e83acd552618ab337d286187ed23d84434bd122
Author: Mitsuru Oshima <oshima@chromium.org>
Date: Tue May 26 20:46:41 2015

Revert "Allow arbitrary scale factors."

Reason for revert: crbug.ccom/484400

BUG=143619,484400
TBR=erg@chromium.org, estade@chromium.org, oshima@chromium.org, stapelberg@google.com

> Allow arbitrary scale factors.
>
>    As discussed in https://codereview.chromium.org/1070433002/, this CL
>    allows arbitrary scale factors, which works because chromium uses the
>    resource most closely matching the scale factor (e.g. 1x for 1.25x,
>    but 2x for 1.9x) and scales it.
>
>    BUG=143619,477716
>
>    Review URL: https://codereview.chromium.org/1078423002
>
>    Cr-Commit-Position: refs/heads/master@{#326851}
>    (cherry picked from commit b3ca230ddd7d1238ee96ed26ea23e369f10dd655)
>
>    device scale factor detection: use gtk-xft-dpi consistently
>
>    Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
>    a more or less random display. This CL consistently uses gtk-xft-dpi which
>    is the better alternative: users can directly influence it (by changing the
>    Xft.dpi X resource, or via their gtkrc).
>
>    See also https://github.com/derat/font-config-info for displaying your
>    current system settings.
>
>    This is a re-do of https://codereview.chromium.org/1070433002/, but with
>    https://codereview.chromium.org/1084453002/ applied (fixes content_shell).
>
>    BUG=473089, 143619
>
>    Review URL: https://codereview.chromium.org/1085603002
>
>    Cr-Commit-Position: refs/heads/master@{#324928}
>    (cherry picked from commit d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34)
>
>    R=stapelberg@google.com
>    TBR=erg, estade, oshima
>
>    Review URL: https://codereview.chromium.org/1103383004
>
>    Cr-Commit-Position: refs/branch-heads/2357@{#246}
>    Cr-Branched-From: 59d4494849b405682265ed5d3f5164573b9a939b-refs/heads/master@{#323860}

Review URL: https://codereview.chromium.org/1159703005

Cr-Commit-Position: refs/branch-heads/2357@{#440}
Cr-Branched-From: 59d4494849b405682265ed5d3f5164573b9a939b-refs/heads/master@{#323860}

[modify] http://crrev.com/1e83acd552618ab337d286187ed23d84434bd122/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/1e83acd552618ab337d286187ed23d84434bd122/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/1e83acd552618ab337d286187ed23d84434bd122/testing/xvfb.py
[modify] http://crrev.com/1e83acd552618ab337d286187ed23d84434bd122/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/1e83acd552618ab337d286187ed23d84434bd122/ui/views/widget/desktop_aura/desktop_screen_x11.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/da20b3fc0aed6674160b02eba4919f6ec2379ec7

commit da20b3fc0aed6674160b02eba4919f6ec2379ec7
Author: Michael Moss <mmoss@google.com>
Date: Wed Jun 10 20:38:05 2015

device scale factor detection: use gtk-xft-dpi consistently

Before this CL, we used either gtk-xft-dpi (UI fonts) or the physical DPI of
a more or less random display. This CL consistently uses gtk-xft-dpi which
is the better alternative: users can directly influence it (by changing the
Xft.dpi X resource, or via their gtkrc).

See also https://github.com/derat/font-config-info for displaying your
current system settings.

This is a re-do of https://codereview.chromium.org/1070433002/, but with
https://codereview.chromium.org/1084453002/ applied (fixes content_shell).

BUG=473089,143619,498426

Review URL: https://codereview.chromium.org/1085603002

Cr-Commit-Position: refs/heads/master@{#324928}
(cherry picked from commit d459a5c49a02507b9ac77ffbbf33dffeb3bc9c34)

TBR=laforge@chromium.org, oshima@chromium.org

Review URL: https://codereview.chromium.org/1171693008.

Cr-Commit-Position: refs/branch-heads/2357@{#469}
Cr-Branched-From: 59d4494849b405682265ed5d3f5164573b9a939b-refs/heads/master@{#323860}

[modify] http://crrev.com/da20b3fc0aed6674160b02eba4919f6ec2379ec7/chrome/browser/ui/libgtk2ui/gtk2_ui.cc
[modify] http://crrev.com/da20b3fc0aed6674160b02eba4919f6ec2379ec7/chrome/browser/ui/libgtk2ui/gtk2_ui.h
[modify] http://crrev.com/da20b3fc0aed6674160b02eba4919f6ec2379ec7/ui/views/linux_ui/linux_ui.h
[modify] http://crrev.com/da20b3fc0aed6674160b02eba4919f6ec2379ec7/ui/views/widget/desktop_aura/desktop_screen_x11.cc

Issue 491882 has been merged into this issue.

A change has landed for this issue, but it's been open for over 6 months. Please review and close it if applicable. If this issue should remain open, remove the "Hotlist-OpenBugWithCL" label. If no action is taken, it will be archived in 30 days.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Oshima, this is essentially done now, right?

Removing Hotlist-OpenBugWithCL label as this bug has blocked on issue.