1318845 - [e10s] Print output are garbage

Status: VERIFIED FIXED

(Core :: Printing: Output, defect, P3)

Description

[baffclan]

1. Open Yahoo 路線情報(transfer information)
   http://transit.yahoo.co.jp/
2. type "東京" in 出発 and "新宿" in 到着
3. Click [検索]
4. Click [ルート1]
5. Click [印刷する]
6. Appear New tab
7. Click Print Preview on File menu
9. Appear Print Preview and Click [Print...]
10. Appear Print dialog, and Click [OK]

Print output are garbage.

Comment 1

[baffclan]

Attachment: Screenshot

Comment 2

[baffclan]

Application Basics:
Name: Firefox
Version: 53.0a1
Build ID: 20161118030222
User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:53.0) Gecko/20100101 Firefox/53.0

Comment 3

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Could you describe the difference between the results you expected and the results you got in slightly more detail than "garbage"?

Comment 4

[yassan]

Attachment: test case

Comment 5

[yassan]

Good
https://archive.mozilla.org/pub/firefox/nightly/2016/11/2016-11-05-03-02-11-mozilla-central/firefox-52.0a1.en-US.win32.zip

Bad
https://archive.mozilla.org/pub/firefox/nightly/2016/11/2016-11-06-03-02-03-mozilla-central/firefox-52.0a1.en-US.win32.zip

Comment 6

[Astley Chen (inactive)]

I can't reproduce on Window 10 Nightly 53 with STR in comment 0 and attached test case either.
I'm not sure if the issue is about specific font or not.

Comment 7

[Xidorn Quan [:xidorn] UTC+11]

I cannot reproduce this issue either, on Windows 10 with Japanese locale and Firefox 52.0a1 20161106030203 32bit version (which is refered as bad version in comment 5). I don't even reproduce this issue with printing it out...

Does this bug only appear when sending to a real printer, but not on print preview? If you print it to a PDF file and print that file, does the same issue still appear?

Comment 8

[yassan]

This issue is not reproduce by print preview. Print is necessary for reproduce. I can reproduce with a real printer or virtual printer.

Comment 9

[yassan]

Attachment: Test case printed by "Microsoft Print to PDF"

Comment 10

[Xidorn Quan [:xidorn] UTC+11]

If this is reproducible with virtual printer, could you help using mozregression (https://mozilla.github.io/mozregression/install.html) to further narrow down the regression range?

Comment 11

[yassan]

(In reply to Xidorn Quan [:xidorn] (UTC+10) from comment #10)
I tried that tool.

2016-11-25T21:25:54: INFO : Narrowed inbound regression window from [c04f84af, 253395be] (3 revisions) to [c04f84af, b4ade2b0] (2 revisions) (~1 steps left)
2016-11-25T21:25:54: DEBUG : Starting merge handling...
2016-11-25T21:25:54: DEBUG : Using url: https://hg.mozilla.org/integration/mozilla-inbound/json-pushes?changeset=b4ade2b0841c5825968f34752d86ee09507a0a9f&full=1
2016-11-25T21:25:55: DEBUG : Found commit message:
Enable the GPU process for Nightly Windows users. (bug 1314133, r=milan)

2016-11-25T21:25:55: INFO : The bisection is done.
2016-11-25T21:25:55: INFO : Stopped

Comment 12

[Xidorn Quan [:xidorn] UTC+11]

That bug seems to just toggle a pref... So if you set "layers.gpu-process.dev.enabled" to false in about:config, does this issue disappear?

Comment 13

[yassan]

(In reply to Xidorn Quan [:xidorn] (UTC+10) from comment #12)
> That bug seems to just toggle a pref... So if you set
> "layers.gpu-process.dev.enabled" to false in about:config, does this issue
> disappear?
Yes.

And, If I set "layers.gpu-process.dev.enabled" to true, I reproduce it with https://archive.mozilla.org/pub/firefox/nightly/2016/11/2016-11-05-03-02-11-mozilla-central/firefox-52.0a1.en-US.win32.zip .

Comment 14

[baffclan]

Xidorn Quan, 
Thanks for information.

yassan-san, 
Thanks for a confirmation of this problem.
ありがとうございます。

Comment 15

[Milan Sreckovic [:milan] (needinfo for best results)]

[Tracking Requested - why for this release]: If printing is broken, we probably want to block on it.

Comment 16

[Marcia Knous [:marcia]]

Tracking 53+ for this printing regression - comment 15 says it all.

Comment 17

[George Wright (:gw280) (needinfo me!)]

Can I get your about:support information?

Comment 18

[Alice0775 White]

Attachment: about:suppot

I can also reproduce the problem on Windows10.

https://hg.mozilla.org/mozilla-central/rev/13736e2db6eb94b02dd28cc88f2943b8109aa374
Mozilla/5.0 (Windows NT 10.0; WOW64; rv:53.0) Gecko/20100101 Firefox/53.0 ID:20161130030206

Comment 19

[baffclan]

Attachment: about:support (Troubleshooting Information)

(In reply to George Wright (:gw280) (:gwright) from comment #17)
> Can I get your about:support information?

attached my about:support.

Comment 20

[Sotaro Ikeda [:sotaro]]

I checked with 3 Win10 PCs, I succeeded to reproduce on 2 PC, and I failed to reproduce on 1 PC.

Comment 21

[Alice0775 White]

Setting print.print_via_parent = false seems to fix the problem.

Comment 22

[Alice0775 White]

(In reply to Alice0775 White from comment #21)
> Setting print.print_via_parent = false seems to fix the problem.
Though, There are serious side effects Bug 1156742.

Comment 23

[atlanto]

Attachment: another test case

Comment 24

[atlanto]

Attachment: the result of another test case

Comment 25

[George Wright (:gw280) (needinfo me!)]

Interesting, so when the GPU process is in use, we hit this gfxWarning:

http://searchfox.org/mozilla-central/source/gfx/2d/NativeFontResourceGDI.cpp#53

When not using the GPU process, same build, we don't. Bas, do you have any idea why this would be? I'm investigating further for now.

Comment 26

[George Wright (:gw280) (needinfo me!)]

nvm, think I've found the issue. Patch incoming.

Comment 27

[George Wright (:gw280) (needinfo me!)]

Attachment: 0001-Bug-1318845-Ensure-we-support-DWrite-fonts-in-the-pa.patch

I think this is a fix for this. Unfortunately, I can't reproduce it so sotaro, can you test it for me?

Comment 28

[Sotaro Ikeda [:sotaro]]

(In reply to George Wright (:gw280) (:gwright) from comment #27)
> Created attachment 8818720 [details] [diff] [review]
> 0001-Bug-1318845-Ensure-we-support-DWrite-fonts-in-the-pa.patch
> 
> I think this is a fix for this. Unfortunately, I can't reproduce it so
> sotaro, can you test it for me?

I confirmed that the problem is addressed on my PC!

Comment 29

[Bas Schouten (:bas.schouten)]

Comment on attachment 8818720 [details] [diff] [review]
0001-Bug-1318845-Ensure-we-support-DWrite-fonts-in-the-pa.patch

Review of attachment 8818720 [details] [diff] [review]:
-----------------------------------------------------------------

Hrm, perhaps we should use this DWriteFactory in other portions of Moz2D as well then, if this is the one we're going to check for here. This will do for now though.

Comment 30

[Milan Sreckovic [:milan] (needinfo for best results)]

Not really "needinfo" more of a "here's some info" for Mason.

Comment 31

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/d82dc57ab9d6
Ensure we support DWrite fonts in the parent process when the GPU process is enabled r=Bas

Comment 32

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/d82dc57ab9d6

Comment 33

[baffclan]

Cannot Reproduce Newest Nightly.

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:53.0) Gecko/20100101 Firefox/53.0
Build ID: 20161217030205

Thanks for Fixing this!

Comment 34

[Andrei Vaida [:avaida]]

Updating status flags based on Comment 33, thanks baffclan!

Comment 35

[Ryan VanderMeulen [:RyanVM]]

[Tracking Requested - why for this release]: needed for bug 1373363

Comment 36

[Lee Salzman [:lsalzman]]

Comment on attachment 8818720 [details] [diff] [review]
0001-Bug-1318845-Ensure-we-support-DWrite-fonts-in-the-pa.patch

[Approval Request Comment]
If this is not a sec:{high,crit} bug, please state case for ESR consideration: Printing brokenness on Windows.
User impact if declined: Printing text does not work on Windows when Direct2D is disabled. See bug 1373363
Fix Landed on Version: 53
Risk to taking this patch (and alternatives if risky): Will now print with DirectWrite fonts when Direct2D is disabled (like we are doing from release 53 onward) instead of GDI fonts, which might cause some slight perceptual differences in printed text for those users, but which will just make it the same as when Direct2D is enabled. 
String or UUID changes made by this patch: None

Comment 37

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Tracked as a dot release driver.

Comment 38

[Ritu Kothari (:ritu) (Inactive, please n-i to RyanVM, jcristau, or pascal)]

Comment on attachment 8818720 [details] [diff] [review]
0001-Bug-1318845-Ensure-we-support-DWrite-fonts-in-the-pa.patch

Approved for uplift to ESR52 branch. This is the only fix so far that will be included in the ESR52.2.1 release.

Comment 39

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-esr52/rev/512efd480dac (FIREFOX_ESR_52_2_X_RELBRANCH - 52.2.1)

Leaving this set to affected until it lands on the default branch for 52.3 as well.

Comment 40

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-esr52/rev/a5bbba7a442d (default - 52.3.0)

Comment 41

[Camelia Badau [:cbadau], Desktop QA]

We reproduced the initial issue on Windows 7 x64 using Nightly 53.0a1 (2016-11-19).
Verified fixed on Windows 7 x64 and Windows 10 x64 using Firefox 52.2.1 ESR (buildID: 20170627155318).

Comment 42

[Ko Nagase]

Our customer reported the following crashes when using ExtJS 6.0.1 app
on Firefox ESR 52.2.1esr Windows x86 version,
but is it related with this update ?
(Because those include `mozilla::gfx::Matrix::Rotation(float) gfx/2d/Matrix.cpp:87` at frame 3.)
https://crash-stats.mozilla.com/report/index/adebe006-dba4-4445-8b31-5e5310170705
https://crash-stats.mozilla.com/report/index/64d27960-a516-486f-a07e-5b9bb0170705
https://crash-stats.mozilla.com/report/index/cb94f255-bc03-420c-999d-5f29a0170705
https://crash-stats.mozilla.com/report/index/137290f6-2f05-44a2-ae99-445400170705

These crashes also seem to be related with bug 1338646 from frame 0 signature.

Comment 43

[Catalin Sasca, Desktop QA [:csasca]]

Attachment: normal is bold in pdf output.png

Hey Jwatt, seems that this has regressed on Win 10 - prin to PDF output. I'll attach a print with the issue. Should we reopen this issue or log another one?

Comment 44

[Jonathan Watt [:jwatt]]

Please file a new bug for bugs that have been marked as FIXED rather than reopening them some time later, especially years later. Do link to this bug from the new bug though. Thanks!