New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 635030 link

Starred by 10 users

Issue metadata

Status: Verified
Owner:
Last visit > 30 days ago
Closed: Oct 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: Bug



Sign in to add a comment

gyp/gn transition causing missing fonts on Adobe drmtest2 and subtitles on ESPN, TED

Project Member Reported by hsiangc@chromium.org, Aug 5 2016

Issue description

google Chrome	54.0.2819.0 (Official Build) dev   
Platform	8676.0.0 (Official Build) dev-channel 
JavaScript	V8 5.4.352
Flash	22.0.0.209 -r1

What steps will reproduce the problem?
(1)Go to http://drmtest2.adobe.com/AccessPlayer/player.html play the DRM video
(2)or go to www.ted.com, www.espn.com play video and turn on subtitle
(3)Observe the behavior, garbage string display

What is the expected output?
No garbage string

What do you see instead?
garbage string displaay

Feedback: will add later

*******************Note********
1. issue reproduce on all devices
2. issue no reproduce on M-53


 
add feedback from M-54
Report ID: 14592139880

Comment 2 by ihf@chromium.org, Aug 6 2016

Status: Assigned (was: Untriaged)
Yes, I noticed that. Will need to bisect this one day.
any update on this one?
Labels: -ReleaseBlock-Beta ReleaseBlock-Stable
btw. I do not see issues with espn.com
ESPN still show garbage string while turning on subtitle, please check screenshot
Screenshot 2016-09-12 at 11.01.42 AM.png
363 KB View Download

Comment 6 by ihf@chromium.org, Sep 12 2016

Labels: -Pri-1 -ReleaseBlock-Stable Pri-2
I don't have a repro on drmtest2 with Minnie M55 build 8775. Also no repro with Samus and M55 build 8782. I don't think it is worth spending a lot of time on this.
Is M-54 and M-55 using different version flash?

Comment 8 by ihf@chromium.org, Sep 12 2016

Right now both are using 23.0.0.162.
Yes, issue only see on M-54
Cc: jecl...@adobe.com smori...@adobe.com
Hey Satoshi/ Jeromie,

Have either of you seen reports of fonts getting replaced by [X] chars, we've recently seen a couple on Chrome OS w/ the latest builds of Chrome (53+) and Flash (23+)?

Comment 11 by esas...@adobe.com, Sep 20 2016

I was able to reproduce the issue with Chromebook ARM peach pi

PASS: Chromebook ARM M53.0.2785.103 running FP 23.0.0.180
FAIL: Chromebook ARM M54.0.2840.24 running FP 23.0.0.180

I saw garbage output on drmtest2, ted and espn.
Cc: gbirtchnell@chromium.org

Comment 13 by ihf@chromium.org, Sep 21 2016

On samus this started with build R54-8562.0.0. The 61 build is fine.
Nothing exciting in
https://crosland.corp.google.com/log/8561.0.0..8562.0.0
but we do have a Chrome uprev
https://chromium.googlesource.com/chromium/src/+log/54.0.2787.0..54.0.2790.0?pretty=fuller&n=10000


Comment 14 by ihf@chromium.org, Sep 22 2016

Cc: steve...@chromium.org
Summary: gyp/gn transition causing missing fonts on Adobe drmtest2 and subtitles on ESPN, TED (was: garbage string display on Adobe DRM file playback or subtitle on ESPN, TED )
Notice the last link shows a truncated list of commits. When use image 8561 and then build the commit below using simple chrome and gn I get missing/boxy fonts on drmtest2. If I build it using gyp I get the good fonts in Flash.

commit 17f6a103beaea57c285f64f12be5f3ede34376c7
Author: chrome-cron <chrome-cron@google.com>
Date:   Sat Jul 2 03:12:31 2016 +0000

    Updating trunk VERSION from 2786.0 to 2787.0
    
    Cr-Commit-Position: refs/heads/master@{#403598}

Comment 15 by ihf@chromium.org, Sep 22 2016

For sanity, this means the failure is actually caused by
https://crosland.corp.google.com/log/8542.0.0..8562.0.0
and indeed there is "Use GN by default for chromeos-chrome". So far so good.
It's possible that this is a bug in deploy_chrome ? There were some changes to that associated with the GYP -> GN migration.

I think that the most efficient way to debug this is to identify the problem and fix it, rather than attempt a bisect or otherwise diff the GN vs GYP output, there are too many possibilities.

If the problem occurs to be missing fonts, can we check to see if they are missing from the image, and if so whether they are missing in the generated output? That would tell us whether this is a GN issue or a deploy_chrome issue.

We have a similar issue around high dpi icons, see  issue 644408 .

Comment 17 by ihf@chromium.org, Oct 6 2016

The issue is fixed on ToT.

When building ToT chrome with M54 chroot we observe no failure.
When building M54 chrome with ToT chroot we observe failure.

This means the fix is in the Chromium sources.

Comment 18 by ihf@chromium.org, Oct 7 2016

Cc: drott@chromium.org josa...@chromium.org derat@chromium.org
Labels: Merge-Request-54
Status: Started (was: Assigned)
drott, your change fixes the Flash font issues on M55 but is also needed on M54:
https://codereview.chromium.org/2278143002
https://codereview.chromium.org/2394203004

I'd like to request a merge to M54. Do you have objections? Notice this is the same as  issue 648523  which is a release blocker.

Comment 19 by ihf@chromium.org, Oct 7 2016

Cc: ihf@chromium.org vsu...@chromium.org blumberg@chromium.org viktorv@chromium.org georgesak@chromium.org lafo...@chromium.org dskaram@chromium.org sdavern@chromium.org hsiangc@chromium.org bustamante@chromium.org
 Issue 648523  has been merged into this issue.
Cc: js...@chromium.org
ihf@, I don't have objections to merging this to 54, but I am not sure it's the full fix for Chrome OS. 

What is the default monospace font in the browser preferences on Chrome OS? 

I don't really understand how my fix affects the Flash issue.

jshin@ are there Type 1 fonts in the Chrome OS font configuration?

Comment 21 by ihf@chromium.org, Oct 7 2016

I know nothing about fonts. But the Fonts settings say:
Standard font: Tinos
Serif font: Tinos
Sans-serif font: Arimo
Fixed width font: Cousine

Comment 22 by ihf@chromium.org, Oct 7 2016

(And I think these did not change from M52 to M54/M55.)
This issue doesn't see on nyan-big and blaze. Can someone check for that?

Comment 24 by ihf@chromium.org, Oct 8 2016

Labels: -Pri-2 ReleaseBlock-Stable Pri-1
Re-adding labels.

Comment 25 by ihf@chromium.org, Oct 8 2016

Cc: h...@chromium.org
I verified the patch applies locally on top of M54

commit 142f5edf1e8e1731d55cbd8631c01cb20a575549
Author: Mitsuru Oshima <oshima@chromium.org>
Date:   Fri Oct 7 13:49:15 2016 -0700

and on samus fixes fonts for all three repro cases that I have.

Haixia, can you make sure this gets integrated on Monday as I am OOO?

Comment 26 by h...@chromium.org, Oct 8 2016

Re#23 yes this is expected. Nyan on M54 is not yet migrated to freon.

Comment 27 by dimu@chromium.org, Oct 8 2016

Labels: -Merge-Request-54 Merge-Approved-54 Hotlist-Merge-Approved
Your change meets the bar and is auto-approved for M54 (branch: 2840)

Comment 28 by dskaram@google.com, Oct 10 2016

Cc: -dskaram@chromium.org

Comment 29 by h...@chromium.org, Oct 10 2016

Status: Fixed (was: Started)
commit	b9d9cb9d8693d5fc3b4bc9e4067b3020f3a90b1c	[log] [tgz]
author	hshi <hshi@chromium.org>	Mon Oct 10 19:44:25 2016
committer	Commit bot <commit-bot@chromium.org>	Mon Oct 10 19:48:33 2016
tree	e43b249d00e5319d16e4737354e8e5d72c4657e9
parent	55a0dba38e5df7bea291c3aa0ce96b01892d74ec [diff]
Filter font list, use Fontconfig and build on Ozone

Since we dropped support for rendering Type 1 fonts in Chrome, we should
filter the list of fonts that can be configured as defaults to only
display fonts that have the right font format, i.e. TrueType or CFF.

A a side effect, this allows us to build the font list enumeration for
ozone as well, as it is only dependent on Fontconfig after this CL, not
dependent on Pango anymore.

For backwards compability with the old implementation, we're adding
three Fontconfig alias families Sans, Serif and Monospace to the list,
since our default settings on Linux included the "Monospace" family for
the fixed width font.

BUG= 630508 , 457307 
TBR=avi@chromium.org
TBR=derat@chromium.org
TBR=spang@chromium.org

NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2278143002
Cr-Commit-Position: refs/heads/master@{#414677}
(cherry picked from commit e70ca0a2e161e47a7f36f94c19e5a9c6448a15b9)

Review-Url: https://codereview.chromium.org/2406983002
Cr-Commit-Position: refs/branch-heads/2840@{#705}
Cr-Branched-From: 1ae106dbab4bddd85132d5b75c670794311f4c57-refs/heads/master@{#414607}

Owner: h...@chromium.org
hshi@, was this merged to M54 already?

Comment 31 by h...@chromium.org, Oct 11 2016

josafat@: Yes see comment above yours (#29). It was merged at refs/branch-heads/2840@{#705}

Sorry I had to manually copy & paste the commit message here because it was cherry-picking a CL whose description did not mention this bug number.

Comment 32 by h...@chromium.org, Oct 11 2016

Labels: -Merge-Approved-54 merge-merged-2840
thanks 
 Issue 655156  has been merged into this issue.
Does this need to be merged into M-55 as well since it went into ToT after the branch cut? 

Comment 36 by ihf@chromium.org, Oct 13 2016

No, it was cherry-picked from M55 into M54.
Status: Verified (was: Fixed)
Verified on 8743.65.0/54.0.2840.59

Comment 38 by js...@chromium.org, Oct 13 2016

Perhaps, this has been answered already. Just in case.... 
> jshin@ are there Type 1 fonts in the Chrome OS font configuration?

No. All of them are either TTF or OTF (with CFF table) for web contents + native UI rendering. 
 


Sign in to add a comment