1662980 - Low quality print output is triggered randomly

Status: RESOLVED WORKSFORME

(Core :: Printing: Output, defect, P2)

Description

[Vlad Lucaci, Desktop QA (:vlucaci)]

Attachment: IMG_20200903_120038.jpg

Affected versions

• 81.0b5

Affected platforms

• macOS 10.15.6
• Ubuntu 18.04

Steps to reproduce

1. Launch FF.
2. Download and open this document in FF
3. Open new Print Preview.
4. Print.

Expected result

• Printing is done without any visual artifacts in either Print preview or actual output.

Actual result

• Print output is very low quality and pixelated. Print preview content is also displayed blurry.

Suggested Severity

• I would consider this issue an S2 as it affects 2 OS's and it does affect the overall UX with low quality prints at random ocassions, but I am aware to provide a clear STR or regression because the issue occurs randomly.

Regression range

• This issue is a regression since this issue was not present in B4. However I am unable to reproduce this issue on mozreg builds at all.

Additional notes

• This issue occurs for Ubuntu, but with a lower repro rate.
• We were unable to determine if there is any manual trigger that displays the issue. This issue seems to occur randomly and disappears after a while.
• Searching for a regression was impossible since the issue was not present at all while using mozregression-ui tool builds.
• I have been not been able to reproduce this issue on nightly yet.
• This issue did not reproduce on the old UI.

Comment 1

[Alice0775 White]

Regression window from comment#0:
https://hg.mozilla.org/releases/mozilla-beta/pushloghtml?fromchange=dcc7a303edfb16d069cba8e3f10c919a0e0f3b29&tochange=60e3cf505151542c568586862381f323862fdf5d

Comment 2

[Hiroyuki Ikezoe (:hiro)]

Thank you Alice!

If the issue doesn't happen on Windows, it's probably a CUPS backend issue. CCing Emily and Erik. What I don't quite understand is why the issue doesn't happen on m-c.

Comment 3

[Hiroyuki Ikezoe (:hiro)]

I am suspecting bug 1661785 is the regressor.

Comment 4

[Hiroyuki Ikezoe (:hiro)]

Hopefully bug 1662518 fixes this issue in the next 81.0b6.

Comment 5

[Erik Nordin [:nordzilla]]

vlucaci, the STR says to download a particular document, but I don't see one attached. I'm happy to look into this, but I want to make sure I'm able to reproduce it first with the same STR.

Comment 6

[Emily McDonough [:alaskanemily]]

I would be a little surprised if bug 1661785 was what caused it, if it's happening randomly. Could this be related to the DPI changes you were working on, :nordzilla?

Comment 7

[Hiroyuki Ikezoe (:hiro)]

Presumably this; https://www-archive.mozilla.org/quality/browser/front-end/testcases/printing/indentation

Comment 8

[Vlad Lucaci, Desktop QA (:vlucaci)]

Hello Erik,

Emily is right, I am using that particular test-file

Comment 9

[Hiroyuki Ikezoe (:hiro)]

I am mostly convinced bug 1660857 which has been uplifted in 81.0b6 fixes this.

Comment 10

[Hiroyuki Ikezoe (:hiro)]

Okay, to be honest I am unsure whether the regressor is in the range in comment 1 because I haven't tested with real printers. So I've been testing with "Save to PDF".

In the case of "Save to PDF", on 81.0b5 user still can select Color mode because bug 1660857 hadn't been landed in 81.0b5, that means if it's "Black and white" mode, the grayscale filter is applied, thus the output result will be blurry.

In the case of real printers, I am still suspecting bug 1661785 is the regressor because that's the only one change affecting color mode.

Anyways, if the problem on printers happens only if the color mode is "B&W", then bug 1660857 should fix the issue.

Hey Vlad, are you able to reproduce the issue with "Color" color mode on 81.0b5? I hope it's not.

Comment 11

[Jonathan Watt [:jwatt]]

Vlad, does this still reproduce? We've had some resolution related changes such as bug 1662820 land that may have helped.

Comment 12

[Vlad Lucaci, Desktop QA (:vlucaci)]

I was unable to reproduce this issue anymore on Windows 10x64, macOS 10.15 and Ubuntu 20 using 82.0b5(20200929175845) and 83.0a1(20200930214529)

Comment 13

[Julien Cristau [:jcristau] (back April 22)]

Calling this fixed per comment 12.