Project: chromium Issues People Development process History Sign in
New issue
Advanced search Search tips
Issue 25541 DirectWrite support
Starred by 531 users Project Member Reported by pkasting@chromium.org, Oct 22, 2009 Back to list

Comments by non-members will not trigger notification emails to users who starred this issue.
DirectWrite is (I think) a new font renderer in Windows 7.  It provides 
things like floating-point advances and has nice output.  Mozilla has a post 
about it at http://blog.mozilla.com/nattokirai/2009/10/22/better-postscript-
cff-font-rendering-with-directwrite/ .

However, it may be COM-only and thus unsuitable for use in our current 
sandboxed renderer.  If we move rendering to a hardware-accelerated 
architecture we may be able to use DirectWrite in that.
 
Comment 1 by karen@chromium.org, Oct 22, 2009
Labels: Mstone-X
Comment 2 by pkasting@chromium.org, Nov 13, 2009
A blog post I read said that MS has backported DirectWrite to Vista.
Comment 3 by brettw@chromium.org, Nov 13, 2009
MSDN says "Windows 7 or Windows Vista with Service Pack 2 (SP2) and Platform Update 
for Windows Vista" on http://msdn.microsoft.com/en-us/library/dd368038(VS.85).aspx
Comment 4 by pkasting@chromium.org, Nov 23, 2009
Labels: -OS-All OS-Windows
Comment 5 by dpranke@chromium.org, Nov 24, 2009
http://www.trustedreviews.com/software/news/2009/11/30/Firefox-3-7-Alpha-Launches-With-Direct2D-GPU-Rendering/p1

IE9 and even Firefox 3.7 (Minefield) already have development builds taking advantage of GPU rendering which increases performance 
dramatically; just tested a development build of Firefox and it does render websites much, much faster. Websites like twitter, facebook and 
google do show dramatic performance improvements, as much as 200%+, as indicated in the website mentioned above.

IE9 and Firefox 3.7 alpha both use the DirectX family of Windows APIs to enable many advances for web developers. The starting point is moving 
all graphics and text rendering from the CPU to the graphics card using Direct2D and DirectWrite. Graphics hardware acceleration means that 
rich, graphically intensive sites can render faster while using less CPU. This will allow web developers to take advantage of the hardware 
ecosystem’s advances in graphics while they continue to author sites with the same interoperable standards patterns they’re used to.

Mstone-X..By then Firefox and IE will have this..and I, amongst many other users, look for the fastest rendering between browsers.
Comment 7 Deleted
Please don't begin making impassioned pleas on this bug.  Adding noise just makes it 
less likely for us to look into this.

We've already filed the bug.  We already have plans to hardware-accelerate our 
pipeline.  We're always interested in improving speed, and this isn't flying under 
our radar.

It will take some time to adapt our sandboxed rendering system to be able to use 
tools like this.
Safari already has this in Safari 5 (which has recently been released), Firefox in nightly builds, and IE9 will have this when it is released. Please start work on this!
Comment 10 by kerz@chromium.org, Jul 8, 2010
Status: Available
Moving all bugs marked as untriaged and mstone X to be available rather than untriaged.  If you think this is in error, please feel free to set back to untriaged.
Comment 11 by dvy12...@gmail.com, Sep 17, 2010
Sub-pixel of DirectWrite positioning is good fact in small font size,otherwise anti-aliasing in y-direction of DirectWrite is good stuff in large font size.
Comment 12 by dami...@gmail.com, Oct 27, 2010
DierctWrite lets you have sub-pixel on the X and grey-scale AA on the Y at the same time :)
Comment 13 by bpgar...@gmail.com, Oct 27, 2010
Even if DirectWrite is unable to be implemented due to sandboxing issues, is it possible to configure Chrome's font renderer (Skia?) to match the DirectWrite font smoothing?
Comment 14 by hamaji@chromium.org, Nov 17, 2010
Comment 15 by yuzo@chromium.org, Nov 17, 2010
Comment 16 by bstell@chromium.org, Nov 30, 2010
> is it possible to configure Chrome's font renderer (Skia?) to match 
> the DirectWrite font smoothing?

If you are talking about FreeType based rendering I don't remember any option to enable vertical anti-aliasing. However, it could be added but I don't know how difficult at this time.

Is anyone doing any work on this? I know that when I website develop, text quality difference between FF and Chrome blows me away -- sadly in FF's favor. 

While performance is important, people spend seconds loading, minutes reading -- the way the text looks should not be overlooked.
I have to agree with the above user. As a Typekit user and occasional user of Google Web Fonts I rarely consider using 90% of the fonts because they look so bad in Chrome on windows. With IE9 and FF4 out it is only Opera and Chrome holding the web back (Safari is fine). People designing on Macs/Linux or just using FF and IE to preview will be unaware of a jagged mess they have made for use chrome users. 
Comment 19 by myutw...@gmail.com, Mar 27, 2011
I'm going to try avoid making an "impassioned plea" here, but agreeing with comment 17, I would love to hear when plans or this feature are expected to be realised. My own website currently renders better in IE9 and FF4 and I expect I'll be designing sites in the future to render with DirectWrite features. It would be great as a developer to know roughly when google was comming on board.
Comment 20 by Deleted ...@, Jun 22, 2011
Ok, so IE is BEATING Chrome?!?  Come on get with it.  This is embarrassing.
Labels: Restrict-AddIssueComment-Commit
Comment 22 by evan@chromium.org, Jan 28, 2012
Cc: reed@google.com bunge...@chromium.org
Issue 105944 has been merged into this issue.
Comment 23 by js...@chromium.org, Jun 29, 2012
Copied from bug 105944 (comment by reed) : 

DirectWrite can give us subpixel position, and prettier antialiasing. Need to do several things to get there:
1. provide DW backend for skia
2. update chrome/webkit to call skia for all of its metrics, rather than GDI directly
3. update chrome/webkit to go through some abstraction for complex layout, rather than calling Uniscribe directly (to be invented)

Mike,  #2 was done. How about #1?  I guess you've worked on that. (do we have bugs to block this issue?) 
I'm not sure about #3, either. 


Comment 24 by js...@chromium.org, Jun 29, 2012
Labels: Internals-Skia Webkit-Fonts
Comment 25 by reed@google.com, Jun 29, 2012
#1 is not done
#2 is basically done
#3 is not done
Cc: reed@chromium.org bashi@chromium.org
Issue 103242 has been merged into this issue.
Cc: msw@chromium.org kojii...@gmail.com asvitk...@chromium.org
Issue 124406 has been merged into this issue.
From the dupe:

"""The current status is that Skia (almost) has a fully complete DirectWrite font host; it isn't currently in the source tree but is out for review. The only thing currently holding it up are some details in association with some other changes. So DirectWrite in Skia should be fairly soon.

Using DirectWrite in Chrome will then require some work on the WebKit side to properly use it, making Chrome's FontPlatformDataChromiumWin more Skia aware and less tied to GDI. The largest part of this work will be ensuring GDI fallback and using DirectWrite for complex text layout. These changes may take some time.
Jun 24, 2012"""

Sounds like #1 might be done by now.
Project Member Comment 29 by bugdroid1@chromium.org, Mar 10, 2013
Labels: -Area-WebKit -Internals-Skia -Webkit-Fonts Cr-Content Cr-Content-Fonts Cr-Internals-Skia
Comment 30 by laforge@google.com, Mar 13, 2013
Labels: -Restrict-AddIssueComment-Commit Restrict-AddIssueComment-EditIssue
Comment 31 by e...@chromium.org, Apr 3, 2013
Cc: e...@chromium.org
Project Member Comment 32 by bugdroid1@chromium.org, Apr 6, 2013
Labels: -Cr-Content Cr-Blink
Project Member Comment 33 by bugdroid1@chromium.org, Apr 6, 2013
Labels: -Cr-Content-Fonts Cr-Blink-Fonts
Labels: -Pri-3 -Mstone-X Pri-2
Summary: DirectWrite support (was: Investigate DirectWrite support)
This seems to be the most canonical DW support bug, and we're actively working on DW support, so retitling slightly. 
Blockedon: chromium:249099
Blocking: chromium:116276
Issue 111777 has been merged into this issue.
Issue 124278 has been merged into this issue.
Cc: rponnada@chromium.org ligim...@chromium.org
Issue 136486 has been merged into this issue.
Issue 116540 has been merged into this issue.
Issue 116276 has been merged into this issue.
Blocking: -chromium:116276
Issue 137848 has been merged into this issue.
Issue 125297 has been merged into this issue.
Issue 136362 has been merged into this issue.
Comment 46 by rbyers@chromium.org, Jul 10, 2013
Blocking: chromium:137692
Comment 47 by rbyers@chromium.org, Jul 10, 2013
Owner: e...@chromium.org
Status: Started
Blocking: chromium:18383
Comment 49 by e...@chromium.org, Sep 12, 2013
Blocking: chromium:290411
Comment 50 by rsch...@chromium.org, Sep 16, 2013
Labels: Hotlist-GoogleApps
Comment 51 by e...@chromium.org, Oct 1, 2013
Blocking: chromium:242735
Issue 104137 has been merged into this issue.
Comment 53 by yuzo@chromium.org, Oct 3, 2013
Cc: -yuzo@chromium.org
Blocking: chromium:231379
Project Member Comment 55 by bugdroid1@chromium.org, Oct 8, 2013
The following revision refers to this bug:
    http://src.chromium.org/viewvc/blink?view=rev&rev=159071

------------------------------------------------------------------------
r159071 | eae@chromium.org | 2013-10-08T01:32:02.724306Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/core/platform/graphics/skia/FontCacheSkiaWin.cpp?r1=159071&r2=159070&pathrev=159071
   M http://src.chromium.org/viewvc/blink/trunk/Source/core/page/RuntimeEnabledFeatures.in?r1=159071&r2=159070&pathrev=159071

Add runtime flag for using DirectWrite on windows

Add runtime enabled feature for using the DirectWrite skia backend on
windows.

BUG= 25541 
R=bungeman@chromium.org, eseidel@chromium.org

Review URL: https://codereview.chromium.org/26335002
------------------------------------------------------------------------
Comment 56 by e...@chromium.org, Nov 5, 2013
Blockedon: chromium:313423
Project Member Comment 57 by bugdroid1@chromium.org, Nov 13, 2013
The following revision refers to this bug:
    http://src.chromium.org/viewvc/blink?view=rev&rev=161954

------------------------------------------------------------------------
r161954 | eae@chromium.org | 2013-11-13T21:33:42.610310Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/web/WebRuntimeFeatures.cpp?r1=161954&r2=161953&pathrev=161954
   M http://src.chromium.org/viewvc/blink/trunk/public/web/WebRuntimeFeatures.h?r1=161954&r2=161953&pathrev=161954

Expose DirectWrite RuntimeEnabledFeature to WebRuntimeFeatures

Expose the DirectWrite RuntimeEnabledFeature to WebRuntimeFeatures to
allow it to be set from chrome. This in turn will allow us to add a
--direct-write command line flag for the experimental DirectWrite font
rendering support.

BUG= 25541 
R=abarth@chromium.org

Review URL: https://codereview.chromium.org/71153006
------------------------------------------------------------------------
Comment 58 by vli@chromium.org, Nov 19, 2013
Labels: Hotlist-DevRel
Project Member Comment 60 by bugdroid1@chromium.org, Dec 12, 2013
The following revision refers to this bug:
    http://src.chromium.org/viewvc/blink?view=rev&rev=163758

------------------------------------------------------------------------
r163758 | eae@chromium.org | 2013-12-12T07:32:43.340585Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/platform/fonts/FontCache.h?r1=163758&r2=163757&pathrev=163758
   M http://src.chromium.org/viewvc/blink/trunk/Source/platform/fonts/skia/FontCustomPlatformDataSkia.cpp?r1=163758&r2=163757&pathrev=163758

Use SkFontMgr for webfont construction on windows

Change FontCache to expose the SkFontMgr implementation and use that to
construct webfonts on windows. This way the appropriate font manager
(DirectWrite or GDI) is used for webfonts just as it is for system
fonts.

BUG= 25541 
R=schenney@chromium.org,pdr@chromium.org

Review URL: https://codereview.chromium.org/113193005
------------------------------------------------------------------------
Labels: -Restrict-AddIssueComment-EditIssue
The latest Chrome Canary build for Windows now supports the --enable-direct-write command line flag which switches font rendering from GDI to DirectWrite.  It is applied to both system fonts and webfonts. We'd love your feedback.

However…

For the moment, it requires the --no-sandbox flag to function. Using this flag disables Chrome's sandbox ( http://www.chromium.org/developers/design-documents/sandbox ) which is a security risk. I would NOT recommend leaving this flag on your Chrome during everyday browsing.

Additionally, subpixel font scaling has been enabled and can be toggled by the --enable-experimental-web-platform-features flag on Windows, Linux and ChromeOS.  Enabling this flag will also toggle on 40 other additional experimental features: http://goo.gl/BtLHtY This may also decrease the stability of some pages.
On Windows, the --enable-direct-write flag is required to also see the subpixel font scaling). 

We're currently working on supporting DirectWrite within the sandbox. When it's finished the feature will move to the --enable-experimental-web-platform-features flag. ( This can also be turned on easily in about:flags ). We'll tell you on this issue thread when that happens.


So in summary:

These experimental flags will *relax your browser security*. ONLY use for testing, NOT everyday browsing.
You can test against Chrome Canary ( https://www.google.com/intl/en/chrome/browser/canary.html ) or a fresh Chromium build ( https://download-chromium.appspot.com ) 
Follow these instructions to apply the flags: http://www.chromium.org/developers/how-tos/run-chromium-with-flags

    # To enable DirectWrite:
    --enable-direct-write --no-sandbox

    # To enable DirectWrite and sub-pixel font scaling:
    --enable-direct-write --no-sandbox --enable-experimental-web-platform-features 

We'd appreciate your feedback.

Comment 62 by ebrami...@gmail.com, Dec 22, 2013
Seems it has serious font rendering issue (attached), but obvious and reproducible one: https://code.google.com/p/chromium/issues/detail?id=330458 (Version 33.0.1750.5 latest on Dev Channel)
EnglishWikipedia.png
179 KB View Download
#62, please be more explicit when you're reporting bugs. I don't see anything obviously wrong in your screenshot. Chrome and Opera are drawing the fonts differently, and that's to be expected, since we've completely replaced the font engine inside Chrome. Chrome vs IE might be a relevant comparison?
Comment 64 by dragory...@gmail.com, Dec 23, 2013
#63: As far as I know, newer IE versions use DirectWrite without subpixel AA, so maybe Chrome vs Firefox would be the most relevant one here?
2bcTDFm.png
290 KB View Download
#63 The font is clearly different. Chrome seems to have problems with choosing correct font and it's not about drawing. Above is Arial Narrow from what it looks like. I've attached the comparison between Chrome, Fx and IE10. I should note though, that IE has really bad contrast settings somehow.

#64 Looks like you didn't enable DirectWrite for Arial font in Fx, thus comparison incorrect. To enable you have to manually remove it from about:config variable gfx.font_rendering.cleartype_params.force_gdi_classic_for_families, otherwise Fx uses GDI for "Classic" fonts like Arial, Tahoma, etc, because there was too much whining from people about "blurry" fonts.
w-ie.PNG
248 KB View Download
w-chrome.PNG
258 KB View Download
w-fx.PNG
255 KB View Download
Comment 66 by sush...@gmail.com, Dec 23, 2013
I have the same issues with fonts on my install. Screenshots attached. 


1.png
126 KB View Download
3.png
264 KB View Download
2.png
280 KB View Download
Comment 67 by ebrami...@gmail.com, Dec 23, 2013
First of all, thank you very much for supporting DirectWrite on Chrome, this is very nice and needed feature as number of stars given show it.

Before commenting here and filing the more specific bug, Issue 330458, I tested Firefox and IE 11 and their result was same with disabled DirectWrite.

I prepared a simple test to show the different on Latin script also:

data:text/html;charset=utf-8,<div style="font-size: 1000%; font-family: arial;">Simple</div><div style="font-size: 1000%; font-family: tahoma;">Simple</div>

I guess there is a specific issue on rendering Arial on Latin script. Anyway hope to be useful.
ChromeDW IE11 Firefox.png
66.7 KB View Download
Comment 68 by m...@rcel.cz, Dec 23, 2013
Enabling DirectWrite fixes my issue 322510.
Comment 69 by e...@chromium.org, Dec 25, 2013
Blockedon: chromium:330458
Comment 70 by e...@chromium.org, Dec 25, 2013
Blockedon: chromium:330639
I believe the problem with the skinny fonts comes from different font selection paradigm between GDI and DirectWrite. In GDI, fonts with "narrow" variants were presented as entirely separate fonts. DirectWrite only presents the font family, and you control the stretch of the font as a parameter.

http://msdn.microsoft.com/en-us/library/windows/desktop/dd368078(v=vs.85).aspx

It looks like the folks at Paint.NET ran into this issue when moving from GDI to DirectWrite:

http://blog.getpaint.net/2009/12/21/paintnet-v352-is-on-the-way-%E2%80%93-better-font-support-for-win7vista/

As you can see from my attachments, if I remove actual font TTF files for Arial Narrow from my fonts folder, the problem resolves. Referencing ebraminio's image above, Arial also renders with the narrow font in Chrome, but since there is no Tahoma Narrow font, it renders consistently. You can also see dev tools thinks it's rendering Arial, not Arial Narrow. 

I would hazard a guess that the issue comes from interpreting an undefined stretch property as the narrowest font available. This isn't so much of a problem in practice because very few fonts have condensed versions. It just so happens Arial does on default Windows installations, and it also happens to be one of the most widely-used fonts on the web.

This also seems to only be a problem with local fonts. (Or maybe even just Arial?) Web fonts over at google fonts all seem to be working correctly.

I can also confirm a font selection problem with Chrome, similar to the problem mentioned in the Paint.NET blog. If you go to Settings -> Advanced -> Customize Fonts... you will find some of the fonts don't preview when you select them, nor do they seem to work in practice. Instead, it renders some kind of fallback font. Specifically, only regular fonts seems to work. Bold, Italic, etc. all render the fallback font.

A temporary workaround for testers would be to move the Arial Narrow font (or other offenders) out of the fonts folder. 
arial-narrow.png
97.7 KB View Download
arial-style.png
21.4 KB View Download
arial-normal.png
139 KB View Download
Comment 72 by ebrami...@gmail.com, Dec 26, 2013
What a great info on #71, I confirm Issue 330458 will be fixed when Arial Narrow is deleted. I guess because Arial Narrow has not Arabic glyphs and that is why Arial will fallback to Tahoma for Arabic script.
Comment 73 by e...@chromium.org, Dec 30, 2013
Blockedon: chromium:331115
Comment 74 Deleted
Blockedon: chromium:333029
Comment 76 by ebrami...@gmail.com, Jan 10, 2014
"Issue 333430: [DirectWrite] WebFonts won't get italic style" seems related to here as commenting on Issue 137692 also is restricted.
Project Member Comment 77 by bugdroid1@chromium.org, Jan 23, 2014
The following revision refers to this bug:
    http://src.chromium.org/viewvc/blink?view=rev&rev=165591

------------------------------------------------------------------------
r165591 | eae@chromium.org | 2014-01-23T01:52:22.375226Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/platform/fonts/skia/FontCustomPlatformDataSkia.cpp?r1=165591&r2=165590&pathrev=165591

Use SkFontMgr for synthetic font styles on windows

Change FontCustomPlatformData::fontPlatformData to use the SkFontMgr
API on Windows for synthetic bold and italic. Without it we'd fall
back on GDI even if DirectWrite is enabled.

BUG= 25541 
R=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/141653017
------------------------------------------------------------------------
Comment 78 Deleted
Comment 79 by dashi...@gmail.com, Jan 24, 2014
Question: Will this also unify artificial small-caps appearance with the way it looks in other browsers? Traditionally Chrome's small-caps has scaled down the 'lower case' letters more than other browsers, producing some potential spacing issues (and not looking as good, in my opinion).
Comment 80 by e...@chromium.org, Jan 28, 2014
Issue 335581 has been merged into this issue.
Comment 81 by rsch...@chromium.org, Jan 28, 2014
Blocking: chromium:132239
Comment 82 Deleted
Thanks a lot for working on this and font smoothing effect in DirectWrite enabled Chromium really looks fantastic! However it has negative effect on Chinese characters. See the attached screenshot. Left is Chromium with DirectWrite and right is Chrome with GDI. The DirectWrite one is too faint.
Screenshot 2014-01-30 19.46.36.png
594 KB View Download
Comment 84 by e...@chromium.org, Jan 31, 2014
That looks too faint indeed, thanks for the report!
Comment 85 by adria...@gmail.com, Feb 7, 2014
Could you review bug that light weight fonts are not displaying correctly in directwrite http://code.google.com/p/chromium/issues/detail?id=335323
Comment 86 by hcm@chromium.org, Feb 13, 2014
Cc: hcm@chromium.org
Just to update, the large remaining issues are:

1. Getting DirectWrite working in the sandbox. This is being worked on and we should have something working soon.

2. Selecting weight and width variations. DirectWrite uses a different set of names from GDI (OpenType fonts have a number of different names). We will need to be able to select fonts using either naming style.

3. Complex text shaping. We are now using HarfBuzz instead of Uniscribe with GDI, and would like to do so with DirectWrite as well. This has uncovered a number of issues with how we use HarfBuzz which are currently being remedied.

In addition, as stated in Comment #83, no matter how 'ugly' many web developers think GDI may be, GDI is still preferred by many CJK users. Until every CJK user has a >200 dpi display, this is likely to continue to be the case. As a result we should allow these users to continue using GDI when available.
Comment 88 by Deleted ...@, Feb 18, 2014
I expect probably everyone already know this (?) but Firefox uses DirectWrite by default, and after receiving backlash for blurry or light font rendering as in post #83, they implemented a set of rules for adjusting the rendering parameters for certain standard system fonts at small sizes so that it closely resembles GDI in style, but still using DW to render. By default it switches the rendering parameters based on size and font family, and IMO it all works pretty seamlessly now with any site. More information in this post (particularly the last paragraph) https://blog.mozilla.org/nattokirai/2011/08/11/directwrite-text-rendering-in-firefox-6/

Attached is the difference between Arial rendered with DW in Firefox, with the GDI style being applied, and Arial actually rendered by GDI in Firefox (by disabling direct2d acceleration). Notice the top still looks better. For the larger page title FF uses a different rendering mode because pixel-snapping does not need to be as strong, resulting in smoother and better shaped characters. Maybe chrome could implement some similar rules before shipping this
example.png
46.3 KB View Download
#88 It should be noted though, that switching to GDI style for "classic" font families is optional in Fx. For example, I always enable DW for everything and see no "blurry" fonts. It's a matter of taste, hardware display quality and so on, so it should be optional.
Comment 90 by papal...@gmail.com, Feb 21, 2014
> GDI is still preferred by many CJK users. […] As a result we should allow these users to continue using GDI when available.

Why not simply behave like IE and Firefox? I believe they have a hardcoded list of CJK fonts that have bitmaps included and they display those bitmaps at the appropriate sizes (it's odd that DirectWrite itself isn't doing it).
As https://code.google.com/p/chromium/issues/detail?id=137692#c123 is closed to comments and directing people here, I'm not sure if I'm doing something wrong, but it appears to suggest that Canary builds do not require --no-sandbox or the chrome://flags/#enable-experimental-web-platform-features flag, just the chrome://flags/#enable-direct-write flag. However, on both Version 35.0.1867.2 canary and Version 34.0.1847.11 beta-m the hideously ugly bad font rendering is gone, but ONLY if I enable --no-sandbox, which the other ticket warns against. Is there any other way of being sandbox-safe AND having clean fonts apart from switching browser for the time being?
@comment 91

No, there is currently no way to get DirectWrite in the sandbox in a released version of Chrome. This work is being tracked by https://code.google.com/p/chromium/issues/detail?id=333029 which is blocking this issue. I am currently working on DirectWrite in the sandbox. DirectWrite is not yet officially supported; the flags are provided so that early adopters can give feedback on any issues they might come across.
Comment 93 Deleted
Comment 94 by tkent@chromium.org, Mar 18, 2014
Issue 352943 has been merged into this issue.
Comment 95 by tkent@chromium.org, Mar 18, 2014
Issue 352944 has been merged into this issue.
Comment 96 by e...@chromium.org, Mar 26, 2014
Blockedon: chromium:356780
Comment 97 by e...@chromium.org, Mar 26, 2014
Blockedon: chromium:356823
Comment 98 by scot...@chromium.org, Mar 27, 2014
Blockedon: chromium:356950
Does this also occur if you start with --no-sandbox? This issue's discussion is already rather long and difficult to follow, if this is what you are seeing you should open another issue and be sure to state the exact Chromium version and flags you're using.
#99 That looks like a sandbox issue I'm debugging at the moment, I hope to have it fixed in the next Canary (or so). Thanks for testing.
I'm running canary 35.0.1915.0 and --no-sandbox doesn't even work so I didn't bother using it, but the fonts still look A LOT better.  The only issue it seems is that the page posted in #99 is not rendered fast enough because there are checkerboards instead.
Comment 103 by e...@chromium.org, Mar 29, 2014
Blockedon: chromium:357864
Comment 104 by Deleted ...@, Mar 29, 2014
In reply to #99. one of the code.google.com issue pages blew up just like that on the current dev build, but cleared up when it was restarted.  Possible memory leak?
Blocking: chromium:333011
Blockedon: chromium:361019
Comment 107 by adria...@gmail.com, Apr 10, 2014
I tried the latest canary build with directwrite enabled but some sites use cleartype http://nytimes.com while others use grayscale https://onedrive.live.com/about/en-us. Is this expected?

BTW: Im using windows magnifier (Windows Key+"+") and Windows Key + "Esc"
Comment 108 by girard@chromium.org, Apr 14, 2014
Blocking: chromium:149881
Blockedon: chromium:360416
Project Member Comment 110 by bugdroid1@chromium.org, Apr 17, 2014
The following revision refers to this bug:
  http://src.chromium.org/viewvc/blink?view=rev&rev=171817

------------------------------------------------------------------
r171817 | eae@chromium.org | 2014-04-16T23:59:21.833932Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/platform/fonts/win/FontCacheSkiaWin.cpp?r1=171817&r2=171816&pathrev=171817
   M http://src.chromium.org/viewvc/blink/trunk/Source/platform/fonts/win/FontPlatformDataWin.cpp?r1=171817&r2=171816&pathrev=171817
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/TestExpectations?r1=171817&r2=171816&pathrev=171817

Fix subpixel font scaling for DirectWrite

Implement support for subpixel font scaling for the Windows DirectWrite
backend and change how the subpixel flag is set to use the SkPaint::
kSubpixelText_Flag flag instead of the method as otherwise the setFlags
call clears the subpixel flag.

BUG= 25541 
R=leviw@chromium.org

Review URL: https://codereview.chromium.org/239283010
-----------------------------------------------------------------
Project Member Comment 111 by bugdroid1@chromium.org, Apr 17, 2014
The following revision refers to this bug:
  http://src.chromium.org/viewvc/blink?view=rev&rev=171886

------------------------------------------------------------------
r171886 | eae@chromium.org | 2014-04-17T20:17:40.452824Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/Source/web/WebRuntimeFeatures.cpp?r1=171886&r2=171885&pathrev=171886
   M http://src.chromium.org/viewvc/blink/trunk/public/web/WebRuntimeFeatures.h?r1=171886&r2=171885&pathrev=171886

Expose subpixel font scaling flag through WebRuntimeFeatures

Expose the SubpixelFontScaling RuntimeEnabledFeatures flag to chromium
through WebRuntimeFeatures as we want to enable it for Windows Direct
Write builds.

R=abarth@chromium.org
BUG= 25541 

Review URL: https://codereview.chromium.org/239433009
-----------------------------------------------------------------
Blockedon: chromium:365208
Comment 113 by e...@opera.com, Apr 28, 2014
Blocking: chromium:367774
Does the point 2 of #87 have a tracking bug?
http://www.dafont.com/larabie-font.font
This font has a different name under DirectWrite and is not picked up with existing setting.
Comment 115 by e...@chromium.org, May 5, 2014
Blockedon: chromium:333430
Project Member Comment 117 by bugdroid1@chromium.org, May 7, 2014
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c5a8ee06577ee8223577b419678b7d1f2ad04780

commit c5a8ee06577ee8223577b419678b7d1f2ad04780
Author: eae@chromium.org <eae@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>
Date: Wed May 07 18:47:47 2014 +0000

[DirectWrite] Force subpixel font sizing to on when DW is enabled

Force subpixel font sizing to be enabled when direct write is enabled
using the --enable-direct-write command line switch. This used to be the
case but broke when sandbox support was added.

BUG= 25541 
R=cpu@chromium.org

Review URL: https://codereview.chromium.org/248593002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@268879 0039d316-1c4b-4281-b951-d872f2087c98


Project Member Comment 118 by bugdroid1@chromium.org, May 7, 2014
------------------------------------------------------------------
r268879 | eae@chromium.org | 2014-05-07T18:47:47.245975Z

Changed paths:
   M http://src.chromium.org/viewvc/chrome/trunk/src/content/renderer/renderer_main_platform_delegate_win.cc?r1=268879&r2=268878&pathrev=268879

[DirectWrite] Force subpixel font sizing to on when DW is enabled

Force subpixel font sizing to be enabled when direct write is enabled
using the --enable-direct-write command line switch. This used to be the
case but broke when sandbox support was added.

BUG= 25541 
R=cpu@chromium.org

Review URL: https://codereview.chromium.org/248593002
-----------------------------------------------------------------
Glad to see DirectWrite to work with sandbox support! However CJK support is still kinda broken. The strokes are just too thin and the characters too faint and they strain eyes easily.
Untitled.png
357 KB View Download
Comment 120 by e...@chromium.org, May 12, 2014
Blocking: chromium:338644
Comment 121 by i...@behe.st, May 12, 2014
Monospace font-family broken since v37?
see screenshot - top right window: font is too small and lacks any AA. 
br.PNG
288 KB View Download
Comment 122 by e...@chromium.org, May 13, 2014
Re 121; that lack of AA is expected, for certain fonts at smaller font sizes we now use the embedded bitmaps. The reduced size is due to subpixel font scaling, the specified size is 11.4px, before we rounded this up to 12 but with subpixel scaling that no longer happens and thus the text is smaller. 
Comment 123 by i...@behe.st, May 13, 2014
#122, thanks for the reply. I tested it further and AA is disabled for sizes up to 21px (default monospace font = courier new) which does not qualify as small and looks kind of ugly now.
I believe using embedded bitmaps was a requested feature for CJK characters only, not whole font families.
#123, second that; for non-CJK characters, AA should be preferably at small sizes, as it actually improves legibility.
Comment 125 Deleted
Comment 126 by e...@chromium.org, May 13, 2014
Please use bug 356823 to discuss monospace font anti aliasing. It has more details and back story.
Allow me pinging about do I need to file #114 as a separate bug?
I found Chrome's font settings actually lacks the access to the alternate names rendering such fonts effectively unreachable under DirectWrite.
I found a regression when updating from 36.0.1976.2 dev-m to 36.0.1985.2 dev-m: When Direct Write is enabled, the Google Mail window is unusable. Some controls are moved from top right to bottom left and the messages aren't displayed at all. Screenshots are attached.

Hardware: Dell XPS 15, 3200x1800 pixel display
OS: Windows 8.1 Pro 64 Bit, Update 1
Chrome-Settings: 
- high-dpi-support enabled in registry
- Direct Write enabled in chrome://flags

It worked fine on the previous version 36.0.1976.2 dev-m. The issue occured after updating to 36.0.1985.2 dev-m. When I switch off the Direct Write support, the GMail window is displayed correctly again.


Chrome_36.0.1985.2_dev-m_DirectWrite_enabled_Gmail_issue.png
110 KB View Download
Chrome_36.0.1985.2_dev-m_DirectWrite_disabled_Gmail_ok.png
250 KB View Download
Comment 129 by e...@chromium.org, May 14, 2014
Re 127: Please do, makes it easier to keep track of.
Do what?
Comment 131 by e...@chromium.org, May 14, 2014
File a separate bug.
File as a separate bug, referring to comment #127.
Ah, sorry for the confusion, I misread the numbers. 

Should a file a seperate bug for #128 as well or should we keep it here?
Filed #114 as issue 373494.
Update to #128:
Yesterdays version 36.0.1985.5 dev-m is also affected.

36.0.1976.2 dev-m: ok
36.0.1985.2 dev-m: not ok
36.0.1985.5 dev-m: not ok

Can anybody confirm this or am I the only one who's affected? I can demonstrate this on 2 different (but similar) laptops and with 4 different GMail accounts (free and business). 
Cc: dominik....@intel.com
Is it possible to not use SimSun?
Chrome seems to fallback to SimSun when theres kanji or chinese characters mixed in.

Serach term : "chrome simsun"

The same effect is seen at 
http://bandai-hobby.net/site/index.html
In the scrolling div just below the header

I'm using the Chrome Stable 35.0.1916.114 m with the experimental DirectWrite flag on.
dw.png
136 KB View Download
Update to #128 and #135:

The problem still exists for 36.0.1985.18 dev-m and 37.0.2008.2 dev-m.
Project Member Comment 139 by bugdroid1@chromium.org, May 29, 2014
The following revision refers to this bug:
  http://src.chromium.org/viewvc/blink?view=rev&rev=175038

------------------------------------------------------------------
r175038 | chrome-bot@google.com | 2014-05-29T05:57:40.362170Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/html/css3-modsel-72-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/custom/stroked-pattern-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/struct-image-05-b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/pasteboard/paste-line-endings-002-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/overflow/overflow-rtl-inline-scrollbar-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xml/css3-modsel-145b-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/custom/coords-relative-units-transforms-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/flexbox/005-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug17548-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/table/relative-position-offsets-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/html/css3-modsel-167a-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/dom/SVGPathSegList-appendItem-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/compositing/overflow/nested-scrolling-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/deleting/delete-at-paragraph-boundaries-006-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug139524-1-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/block-replaced-width-001-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/html/css3-modsel-82-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/dynamic/004-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/repaint/line-flow-with-floats-2-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/absolute-replaced-width-029-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug3977-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/fake-italic-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t0803-c5502-mrgn-r-01-c-a-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xml/css3-modsel-170d-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/body-propagation/background-image/006-xhtml-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/justify-ideograph-simple-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/table/cell-coalescing-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/virtual/gpu/compositedscrolling/overflow/paint-neg-z-order-descendants-into-scrolling-contents-layer-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/layers/inline-dirty-z-order-lists-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/html/css3-modsel-170d-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug28928-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/body-propagation/overflow/005-xhtml-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/execCommand/paste-2-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/css/pseudo-element-line-break-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/selectors/166a-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug113235-3-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/lists/marker-before-empty-inline-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/coords-trans-03-t-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/dynamic/014-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/button-align-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/ietestcenter/css3/text/textshadow-003-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/repaint/filter-repaint-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/media/video-layer-crash-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/input-align-image-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t040105-import-00-b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t170602-bdr-conflct-w-10-d-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/dom/Window/btoa-pnglet-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/encoding/utf-16-big-endian-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla_expected_failures/dom/appendCol1-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/other/slashlogo-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/paths-data-09-t-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/block/positioning/abs-inside-inline-rel-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css1/text_properties/text_indent-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug647-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/table/dynamic-caption-add-remove-before-child-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/custom/path-bad-data-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-I18N/text-anchor-dirRTL-anchorEnd-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/pservers-grad-01-b-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/overflow/overflow-with-local-background-attachment-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/animate-elem-03-t-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/block/float/intruding-float-remove-from-sibling-block-on-fixed-position2-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/dom/insertCellsExpand1-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/custom/object-sizing-no-width-height-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug45486-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/selectors/167a-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla_expected_failures/bugs/bug91057-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t0511-c21-pseud-anch-00-e-i-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/selectors/lang-vs-xml-lang-xhtml-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/x_th_width-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/http/tests/misc/object-embedding-svg-delayed-size-negotiation-2-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/ruby/ruby-beforeafter-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/style/unbold-in-bold-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/ietestcenter/css3/bordersbackgrounds/background_repeat_space_content_box-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/virtual/softwarecompositing/overflow/textarea-scroll-touch-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/basic-inputs-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/table-height-algorithm-023-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t170602-bdr-conflct-w-22-d-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/selection/4889598-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-d2-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/images/png-suite/test-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/trailing-white-space-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug16252-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t0602-c13-inheritance-00-e-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-170b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/x_th_nowrap-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/html/details-remove-child-1-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t1205-c566-list-stl-00-e-ag-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/absolute-non-replaced-width-010-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/animate-elem-15-t-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/lists/008-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/inserting/before-after-input-element-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/animate-elem-32-t-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/as-background-image/svg-as-background-3-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-43b-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/absolute-replaced-height-026-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/selectors/170d-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-21c-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css1/pseudo/pseudo_elements_in_selectors-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug8411-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/inserting/insert-paragraph-04-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/core/table_widths-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-56-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/x_table_border-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/input-value-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t040302-c61-rel-len-00-b-ag-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css1/box_properties/margin-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/tr_bgcolor_purple_rgb-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t170602-bdr-conflct-w-51-d-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla_expected_failures/marvin/table_overflow_caption_hidden-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-122-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-103b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css1/box_properties/border_left_width_inline-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/absolute-non-replaced-width-020-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t0803-c5502-mrgn-r-00-c-ag-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/deleting/delete-ws-fixup-004-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/animate-elem-44-t-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/absolute-replaced-height-036-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/search-cancel-button-style-sharing-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/deleting/5144139-2-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/block/margin-collapse/063-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/basic/generic-family-reset-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/animate-elem-61-t-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/body-propagation/background-image/006-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/selection/iframe-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-66-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/text/select-x-list-1-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t060402-c31-important-00-b-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/compositing/visibility/visibility-simple-video-layer-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/borders/mixed-border-styles-radius-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/reset-emptyRun-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/international/rtl-caret-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xml/css3-modsel-44-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-132-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-104b-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla_expected_failures/other/empty_cells-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/xsl/xslt-enc-cyr-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/inserting/insert-tab-001-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/20110323/float-non-replaced-width-004-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t170602-bdr-conflct-w-80-d-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/selection/4932260-3-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xml/css3-modsel-130-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug10565-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/W3C-SVG-1.1/filters-light-04-f-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/text/text-appearance-basic-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/bugs/bug46268-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/editing/deleting/delete-block-merge-contents-006-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/x_tr_id-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/block/float/overhanging-float-remove-from-fixed-position-block2-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/forms/indeterminate-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css3/selectors3/xhtml/css3-modsel-177b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/fast/text/font-initial-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t0804-c5507-padn-r-03-f-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/svg/custom/object-sizing-expected.txt?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/css2.1/t040105-atrule-00-b-expected.png?r1=175038&r2=175037&pathrev=175038
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/platform/win/tables/mozilla/marvin/x_td_align_c
Project Member Comment 140 by bugdroid1@chromium.org, May 29, 2014
The following revision refers to this bug:
  http://src.chromium.org/viewvc/blink?view=rev&rev=175043

------------------------------------------------------------------
r175043 | eae@chromium.org | 2014-05-29T09:56:43.085385Z

Changed paths:
   M http://src.chromium.org/viewvc/blink/trunk/LayoutTests/TestExpectations?r1=175043&r2=175042&pathrev=175043

Skip redundant DirectWrite virtual test suite

BUG= 25541 
TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/304903002
-----------------------------------------------------------------
Update to #128, #135 and #138:

The problem still exists for 37.0.2024.2 dev-m.
Comment 142 by wu.s...@gmail.com, Jun 6, 2014
The latest dev update finally fixed the non-antialiased fonts at small sizes issue, but kerning looks really bad for some fonts at small sizes. Calibri in particular (see attached screenshot from chrome font settings page).
Calibri_kerning.png
7.0 KB View Download
Works well most of the time for me, but occasionally when I switch between tabs, text disappears partly, and gets even worse when I hover over that that text.
chrome_bug.png
5.9 KB View Download
Comment 144 by meh@google.com, Jun 9, 2014
Will this be shipping enabled by default in M37?  If yes, Emil, could you please update http://www.chromestatus.com/features/4725550652325888?
I'm on 37.0.2040.0 Win x64 build and System font kerning is horrible. Web fonts appear to be mostly okay. I have font size set to medium in chrome settings, but system DPI setting is 125%. Windows 7.
webfonts-2.png
87.8 KB View Download
webfonts-1.png
78.8 KB View Download
Sorry for double post but I decided I should upload two screenshots comparing Canary and the Release builds. Identical (default) settings - and same system of course.
fontissue-main.png
227 KB View Download
fontissue-canary.png
246 KB View Download
Comment 147 by e...@chromium.org, Jun 9, 2014
Re #144, meh: Yes, will ship in M37 by default. Will send out an update and update chromestatus.com once the flag is flipped (currently in CQ).
Comment 148 by e...@chromium.org, Jun 9, 2014
Re #146, mark@earthshine.it: The size of the chrome UI also changed, that is unrelated to DirectWrite.
Looks like the change to enable high dpi by default (on high dpi systems) is causing the change in appearance, see bug 149881 for details.
Since enabling DW, I've noticed in dev that going back to pages is extremely slow to repaint, or, just fails to and I have to kill the tab. See bug 382146
Project Member Comment 150 by bugdroid1@chromium.org, Jun 9, 2014
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/5b263b6a59875886517835ab1a236a30255bd9a5

commit 5b263b6a59875886517835ab1a236a30255bd9a5
Author: scottmg@chromium.org <scottmg@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98>
Date: Mon Jun 09 21:47:34 2014

Enable DirectWrite by default

Flips the sense of the flag for now. If all seems OK as we move through
channels, I will remove it later.

R=cpu@chromium.org, jam@chromium.org
BUG= 25541 

Review URL: https://codereview.chromium.org/323563005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@275860 0039d316-1c4b-4281-b951-d872f2087c98


Comment 152 by e...@chromium.org, Jun 11, 2014
Status: Fixed
As of Chrome 37.0.2042.0 (current canary) DirectWrite has been enabled by default.

Please keep using this bug to report bugs related to DW text rendering. Thank you.
Comment 153 by ebrami...@gmail.com, Jun 12, 2014
Good job.

However I still see one issue, I can not see CJK http://zh.wikipedia.org http://ja.wikipedia.org http://ko.wikipedia.org The screenshot is on Version 37.0.2044.2 canary on Windows 7. I am not user each of this languages and may I don't have the related font however using Disable DirectWrite flag the issue is not available. I have clean profile chrome --user-data-dir=/dev/null and my about:gpu is available on Issue 382084#c5. May just a Skia updated is needed? I just wanted to let you know.
CJK.png
213 KB View Download
Pinging for issue 373494.
I'm also experiencing the issue referenced in #99, #104 and #143 where random characters are not rendered. Sample screenshot attached.

I don't have anything repeatable and don't know enough about DirectWrite to provide any technical information, but here are my observations:

- this usually happens after an extended browsing session

- once the issue starts in one tab, it gets progressively worse in any new tabs and can start happening in previously loaded tabs, especially in parts of the viewport that are not initially visible

- reloading the page does not fix the issue; Chrome as a whole must be restarted

- it isn't specific to any single site that I've noticed (I've seen it in Facebook comment iframes and many other random sites)

- disabling DirectWrite text rendering completely resolves the issue

---

Chrome: 35.0.1916.153 m

OS: Windows 7 SP1
Processor: Intel Core2 Quad (Q8300 @ 2.5GHz)
RAM: 8GB
DirectX: 11.0

Graphics: NVIDIA GeForce GTX 550 Ti
- Driver Version: 311.06 (as reported by Nvidia Control Panel)
- Driver Version: 9.18.13.1106 (as reported by Windows Device Manager)
- Dedicated video memory: 1024 MB
- Shared system memory: 3071 MB

5-28-2014 4-36-33 PM.png
54.6 KB View Download
Comment 156 by e...@chromium.org, Jun 16, 2014
Blocking: chromium:176351
#155 This! I see this bug randomly once a day at least, with some or all characters gone, both stable and canary-x64 channels affected with DW enabled on 2 different PCs.
See https://code.google.com/p/chromium/issues/detail?id=377932 and https://code.google.com/p/chromium/issues/detail?id=377081 

Also, text underlining is broken in recent canary versions, see 2nd attachment, top window. Stable channel is fine, both have dw enabled.
bug.png
48.5 KB View Download
text-underline.PNG
133 KB View Download
Comment 158 by ebrami...@gmail.com, Jun 18, 2014
From CJK issue I had (#153) still Korean and Japanese issue is available.
A.png
255 KB View Download
Comment 159 by ben.cl...@gmail.com, Jul 17, 2014
I've been using the directwrite flag for many months, and I've seen the bug mentioned in #99, #104 and #143, #155, and #157 at least weekly, including today. It always happens after Chrome has been running non-stop for at least a day with multiple browsers/tabs running.

Perhaps this should be tracked in a new issue? Enabling DirectWrite by default without fixing this first seems hasty.
This doesn't seem to work well under Windows 7. The new rendering is considerably worse (not better) for almost all fonts. Contrast is reduced and what used to be simple 1px vertical lines for characters like l, are spread across 2 pixels. This really impares readability. If this can't be made to work well on Windows 7, it should be disabled by default until it can be improved to render as well as the old method. 
Chrome-font.png
10.9 KB View Download
Comment 161 by Deleted ...@, Sep 4, 2014
Nick: what they need to do is take a page from Firefox's DirectWrite implementation. DirectWrite needs different settings for smaller fonts. You need to use GDI-style rendering for standard screen fonts under 16pts or so.

Both projects are open source, so they should learn from each other.
Comment 162 by Deleted ...@, Sep 4, 2014
The switch to DirectWrite degrades the display of non-anti aliasing text, ie bitmapped fonts, when ClearType is disabled : the spaces between the letters is uneven/irregular.  Kerning is also degraded when anti-aliasing is swictched off in IE9 (with dwrite.dll), whereas IE8 was perfect.  Kerning is also perfect in Firefox 31.
Fortunately the option exists in Chrome to return to the old GDI.  Please never remove this option.

chrome_directwrite.png
10.6 KB View Download
chrome_gdi.png
10.5 KB View Download
Cc: -dpranke@chromium.org
Underlining of most web safe fonts is broken — http://jsfiddle.net/SerGen/3opq3k1v/

Stable, 37.0.2062.103
issue--chrome-underline.png
22.1 KB View Download
Comment 165 by laforge@google.com, Jun 13, 2015
Cc: drott@chromium.org
Comment 166 by laforge@google.com, Jun 13, 2015
Attempting to remove dominik.rottsches@intel.com from cc.
using 150% dpi scaling in windows 10 (1080p) resolution.. as others fonts are blurry in chrome (Version 50.0.2638.0 dev-m (64-bit)) which is not in firefox. Whats the fix?

Comment 168 by e...@chromium.org, Feb 16, 2016
Cc: -dominik....@intel.com
Re 167: Adjust your cleartype settings or disable antialiasing system wide if you prefer it aliased.
Sign in to add a comment