1413133 - Windows window snap state not restored for Thunderbird

Status: NEW

(Thunderbird :: OS Integration, defect)

Description

[Stanzilla]

Attachment: 2017-10-31_10-32-37.png

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.18 Safari/537.36

Steps to reproduce:

I have a multi (3) monitor setup, not sure how relevant this is.
Snap Thunderbird to a window edge, I used the left side, so WindowsKey+Left Arrow to do that. I have Thundebird on the right one of the 3 monitors.

Shutdown your computer
Start your computer
Start Thunderbird


Actual results:

Thunderbird was not exactly snapped to the edge anymore, instead it was a few pixels away from the edge.


Expected results:

Thunderbird should have restored in the snapped state.

Comment 1

[Jorg K (CEST = GMT+2)]

Does Firefox store the snapped state? If not, please file a bug for Firefox or move this bug to the Firefox queue.

Thunderbird uses window management from common Mozilla core modules. I could well imagine that the window size and position is reinstated but not the snap state.

Comment 2

[Wayne Mery (:wsmwk)]

(In reply to Jorg K [Almost not working on Thunderbird (some bustage-fix only) due to non-renewal of contract] from comment #1)
> Does Firefox store the snapped state? If not, please file a bug for Firefox
> or move this bug to the Firefox queue.
> 
> Thunderbird uses window management from common Mozilla core modules. I could
> well imagine that the window size and position is reinstated but not the
> snap state.

stan?

Comment 3

[Stanzilla]

Sorry, missed the reply. Firefox does in fact restore the snap state correctly. It seems to have some special code to do it. If I start it up again after closing, it restores in the "wrong" position but then snaps itself back into position after a split second.

Comment 4

[Wayne Mery (:wsmwk)]

Sorry, I don't have expertise in this area

Comment 5

[Wayne Mery (:wsmwk)]

Rob can you reproduce?
If so, please change status to new

Comment 6

[Stanzilla]

Who would be able to confirm this? I can still repro it on 61.0a1 (2018-03-21) (64-bit).

Comment 7

[Wayne Mery (:wsmwk)]

I can't but maybe rctgamer3 can

Comment 8

[rctgamer3]

Tested this with both Thunderbird and Firefox. Both reopen a couple of pixels from the screen border (likely default Windows behaviour, probably the window size + shadow width). However, Firefox fixes this, but Thunderbird doesn't.

Comment 9

[Stanzilla]

Yeah that is exactly what I am seeing. Thanks for confirming.

Comment 10

[Jorg K (CEST = GMT+2)]

Richard, any idea where we should start looking?

Comment 11

[Richard Marti (:Paenglab)]

Unfortunately not. Maybe a regression window where FX was fixed or TB regressed could help.

Comment 12

[Jorg K (CEST = GMT+2)]

Jim, are you aware of any code in Firefox that deals with the window snap state that we could port to TB? Looks like FF is handling it well and TB doesn't.

Comment 13

[Wayne Mery (:wsmwk)]

Can you reproduce with nightly from https://archive.mozilla.org/pub/thunderbird/nightly/latest-comm-central/ ?
Find a regression range?

Comment 14

[Stanzilla]

I tested with a recent nightly (62.0a1 (2018-05-19) (64-bit)) and it is still broken there.

Comment 15

[Jim Mathies [:jimm]]

(In reply to Jorg K (GMT+2) (bustage-fix only, NI for urgent reviews) from comment #12)
> Jim, are you aware of any code in Firefox that deals with the window snap
> state that we could port to TB? Looks like FF is handling it well and TB
> doesn't.

it's all down in nsWindow, I'd suggest turning on the native message debugging and looking at that code for clues.

Comment 16

[rctgamer3]

Still broken with newest Daily (2018-06-03)

Comment 17

[Stanzilla]

Still broken with 63.0a1 (2018-07-25) (64-bit)

Comment 18

[jjurkus]

I think I have seen this as well, slightly different, but it could be the same issue.

Thunderbird 60.0 64-bit on windows. Start with a new profile by first deleting the Thunderbird folders in Roaming and Local. Maximise Thunderbird, close thunderbird, reopen, it is not maximised anymore.

With Thunderbird 52.9.1 all was well. We use some code to copy xulstore.json to a new profile. This works fine, without this code this issue is also there. After closing thunderbird we look into the xulstore.json in the profile:

> "messengerWindow":{"screenX":"0","screenY":"0","width":"2560","height":"1052","sizemode":"maximized"},

But it doesn't respect this setting! You can even make an empty xulstore.json with only this line:

>{"chrome://messenger/content/messenger.xul":{"messengerWindow":{"sizemode":"maximized"}}}

And it will not maximise the window.

Comment 19

[jjurkus]

Of course I found the cause of my issue:
privacy.resistFingerprinting was set to true.

Perhaps the reporter has this same issue?

Comment 20

[Stanzilla]

(In reply to jjurkus from comment #19)
> Of course I found the cause of my issue:
> privacy.resistFingerprinting was set to true.
> 
> Perhaps the reporter has this same issue?

privacy.resistFingerprinting is set to false here

Comment 22

[Wayne Mery (:wsmwk)]

Stanzilla, please don't change the version# - it should stay at it's original value.

Richard/jorg, do you know a core dev who might give us insight?

Comment 23

[Jorg K (CEST = GMT+2)]

Jim already answered in comment #15. You meant Mozilla core, I assume.

Comment 24

[Richard Marti (:Paenglab)]

Unfortunately not.

Comment 25

[Stanzilla]

Oh sorry, want me to change it back?

Comment 26

[Wayne Mery (:wsmwk)]

(In reply to Stanzilla from comment #25)

Oh sorry, want me to change it back?

I already changed it back ;)

Comment 27

[Wayne Mery (:wsmwk)]

You will want to use a test profile to find the regression range using https://mozilla.github.io/mozregression/

Comment 28

[Arthur K. (he/him)]

Confirming this still seems to be present in TB 91.0b4. See bug 1576738.

Comment 29

[Wayne Mery (:wsmwk)]

Still reproduces?

Comment 30

[omgitsraven]

I'm another person who's been experiencing this bug; It still reproduces, even on Windows 11.

Comment 31

[Arthur K. (he/him)]

(In reply to Wayne Mery (:wsmwk) from comment #29)

Still reproduces?

I only have single monitor setups at work and home now so I can't test with original STR.

Comment 32

[Arthur K. (he/him)]

(In reply to omgitsraven from comment #30)

I'm another person who's been experiencing this bug; It still reproduces, even on Windows 11.

And this is on Win 11 25H2 26200.7623 and TB v147?

Comment 33

[omgitsraven]

I'm on esr so I'm actually only on TB 140.7.0esr, but yes to Win 11 25H2 26200.7623.