1257071 - When unplugging a secondary low-dpi monitor on which Firefox is open, Windows moves Firefox to the high-density display and all chrome text is misplaced and content is mis-sized

Status: RESOLVED DUPLICATE of bug 1154125

(Core :: Widget: Win32, defect, P2)

Description

[bvanoudtshoorn]

Attachment: Capture.PNG

User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:47.0) Gecko/20100101 Firefox/47.0
Build ID: 20160315004019

Steps to reproduce:

1. Open Firefox on a low-density display in a mixed-density environment.
2. Unplug the low-density display; Firefox will be moved to the high-density display automatically.


Actual results:

* Chrome UI elements have their text inappropriately sized and positioned (tabs, addon badges, URL hover text, menus [including non-hamburger]).
* Maximising and restoring Firefox on the display does not resolve the issue


Expected results:

Firefox should have been appropriately sized in all contexts.

Comment 1

[bvanoudtshoorn]

I'm now struggling to recreate this issue; but the screenshot proves it's happened at least once... I've undertaken the following tests to try and recreate this issue:

* 2x low-density, 1x high-density. Firefox on low-density display.
  * Logged in and active.  Unplug Fx monitor => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Locked but awake.  Unplug Fx monitor, unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Locked and sleeping. Unplug Fx monitor, wake and unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Sleeping but unlocked. Unplug Fx monitor, wake and unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
* 1x low-density, 1x high-density. Firefox on low-density display.
  * Logged in and active.  Unplug Fx monitor => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Locked but awake.  Unplug Fx monitor, unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Locked and sleeping. Unplug Fx monitor, wake and unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.
  * Sleeping but unlocked. Unplug Fx monitor, wake and unlock => Fx moves to primary (high-density) display. Exhibits issue for a second or so, then resolves.

If it helps, I'm running a Surface Pro 3 in a dock with one monitor plugged into the dock and the other plugged directly into the device.

Comment 2

[Loic]

Does it happen with FF45?

Comment 3

[bvanoudtshoorn]

I'm not sure if it happens in 45, Loic, as I don't currently have it installed. Sorry! My *assumption* would be that it doesn't, and that it's related to the (fantastic) work to run nicely in mixed-density environments on Windows.

I've definitely never encountered this before on my setup (the configuration of which hasn't changed for at least 6 months), and I've been running Firefox Dev Edition all along, so I would have been running 45 for a while.

Comment 4

[Jonathan Kew [:jfkthame]]

This wouldn't happen in 45 (or 46), as we don't support mixed resolutions and dynamic resolution-changes there. It's related to that feature, which is currently on 47 (though might get held back to 48 if there are too many remaining issues).

As for this specific bug, it sounds quite similar to bug 1254019, where a patch has just landed on inbound. It should go into Nightly (48) within the next day or two, and I'll then request uplift to Aurora (47) as well. Let's see whether that resolves things here.

(Note that a temporarily-bad display that resolves itself after a moment, as described in comment 1, is to be expected when the display configuration suddenly changes. I'm more concerned about any cases where the incorrect scaling persists and doesn't "fix itself" automatically.)

Comment 5

[Jonathan Kew [:jfkthame]]

bvanoudtshoorn: Does this still occur with current Nightly or Developer Edition? I'm guessing it may have been resolved by bug 1154125, which should be fixed for Firefox 49 and later.

Comment 6

[bvanoudtshoorn]

jfkthame: I haven't encountered this for a good while now on the Developer Edition. I think your guess about bug 1154125 resolving it is correct. :D

Comment 7

[Jonathan Kew [:jfkthame]]

Great, thanks for letting us know; in that case, I'll close this as a duplicate.

If you do see similar problems again, please file a fresh report with all the details. Thanks!