1874787 - changed coordinate system when using --geometry

Status: UNCONFIRMED

(Thunderbird :: Untriaged, defect)

Description

[paul]

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/115.0

Steps to reproduce:

I have a start-up application to populate screen with terminals, Firefox, and Thunderbird. The terminals are created with mate-terminal and use the --geometry option to position. Firefox and Thunderbird use wmctrl to position the window. For the left-most material, both mate-terminal and Thunderbird position the left side to be "x=20" so all of them are 20 pixels to the right of the left side of my screen

Actual results:

Changed behavior in Thunderbird 115:

The mate-terminals are where I expect but thunderbird is further to the right. To get it to match up to where it should be, I have to tell wmctrl to position "x=-10"

I am suspect of Thunderbird as wmctrl and Firefox are where they should be. Plus, wmctrl has not changed on my computer since it was installed on 19oct2020

Expected results:

I expect the mate-terminals and Thunderbird to get their left-hand x position from the same coordinate system and, therefore, if the use the same number they should be aligned on the left-side at that number

Comment 1

[Wayne Mery (:wsmwk)]

This happens also with Firefox?

Comment 2

[paul]

Wayne:

I did not notice any change in Firefox. On my Centos 7 system where Thunderbird went from 105 to 115 when I updated on 12jan24, I can see that Firefox updated from 102 to 115 at the same time.

Let me clear a window in my schedule so I can clone a test program from my hero to show exactly what my code is doing in the start-up app. This will also help me confirm that the problem isn't somewhere on my end.

Paul

Comment 3

[paul]

Wayne:

I reduced my code to what I thought would be a simple test case and discovered that, yes, Firefox is affected by this. So now I have to play with the original code and the test case I created for you to see how that happened.

I have some theories but they may turn out to be excuses

Please give a couple more days to sort this out on my end as to finding out what the real change is

Paul

Comment 4

[paul]

Wayne and Magnus:

I worked on this some more tonight and am seeing something which convinces me that I am seeing things incorrectly.

The first item I create in the start-up app is a mate window. The next thing is a Firefox window moved after creation by wmctrl.

I am seeing the Firefox window showing up in a different position depending on the coords I give the mate-terminal creation. But I am not seeing wmctrl moving the window after creation. Which I do see in my hero script.

I am going to need more time to track down what is going on (as in more than another 24 hours) and I appreciate your bearing with. This is an odd one to triage and gui stuff is not my domain (or shall we call it "comfort zone")

I suspect that I will end up submitting a test case written in python with options to show the various weirdness I am seeing. Or weirdness I am not understanding, I don't know

Paul

Comment 5

[paul]

Wayne and Magnus:

I figured out that the last problem was a pilot error on my part. Sorry for the noise

I am attaching a test program to show the problem (will show up on the next post)

The usage is:
Usage for >./test_tBird_1874787.py<
-d/--debug: turn on debug
-h/--help: this message
-n/--new: what works in tBird v115
-o/--old: what works in tBird v105
NOTE: either -h/--help, -n/--new, or -o/--old must be selected

I will note that my original "v115 requires me to change 20 to -10" is incorrect as the right number is "-6".

The actual wmctrl code are line 303-306.

Please let me know if you need me to provide more information and/or if my test program has a flaw that I need to correct before this apparent bug can be confirmed.

Thanks in advance for your help,
Paul

Comment 6

[paul]

Attachment: test_tBird_1874787.py

This python program was developed on Linux (in this case, Centos7) with executable python3 installed and just need to be run as ./test_tBird_1874787.py. I suspect it should work on other opSys but have no way to verify

Comment 7

[Wayne Mery (:wsmwk)]

Paul, does this still reproduce for you with a current version?

Comment 8

[paul]

Wayne:

Yes, this still exists. I ran a test to confirm even though I can see the hack I had to put in my code is still there (and it this issue had been fixed and the coordinate system matched what a terminal and firefox want, I would have had to of removed the hack)

Best,
Paul

Comment 9

[paul]

Wayne:

Forgot to add. I am running Rocky 9.4 with Thunderbird 128.4.0esr. I see the with the recently released Rocky 9.5 that 128.5 is available but I am unable to update to 9.5 owing to "The epel and rpmfusion repositories have conflicts with one another" (my report is https://bugs.rockylinux.org/view.php?id=8350).

Paul

Comment 10

[paul]

The packages in rpmfusion that prevented me from upgrading to rocky9 have been fixed and the last one was pushed today. So, I have been able to update to Rocky 9.5 and have tested this with Thunderbird 128.5.0.esr. It has not been fixed in this latest release