Tk Source Code

View Ticket
Login
Ticket UUID: 220791
Title: Window gives false geometry information
Type: Bug Version: obsolete: 8.0.3
Submitter: welch Created on: 2000-11-01 00:56:03
Subsystem: 65. Generic Window Operations Assigned To: jenglish
Priority: 5 Medium Severity: Minor
Status: Closed Last Modified: 2026-11-06 15:23:11
Resolution: Fixed Closed By: nemethi
    Closed on: 2026-11-06 15:23:11
Description:
OriginalBugID: 1441 Bug
Version: 8.0.3
SubmitDate: '1999-03-04'
LastModified: '2000-01-06'
Severity: MED
Status: Assigned
Submitter: pat
ChangedBy: hobbs
OS: HP-UX
OSVersion: B.10.20/A
Machine: Other
FixedInVersion: 8.3b1


Name: Stephen Totten

ReproducibleScript:
 /stephent $ wish
% winfo geometry .
200x200+120+120
% wm geometry . 200x200+120+120
% winfo geometry .
200x200+120+120
% wm geometry . 200x200+130+130
% winfo geometry .
200x200+138+160
% winfo x .
138
% winfo y .
160
% winfo rootx .
138
% winfo rooty .
160
%

ObservedBehavior:
Window gives out bad geometry information, i.e. 200x200+138+160.

DesiredBehavior:
Window gives out correct geometry information, i.e. 200x200+130+130.

Verified the screwy behavior, although it is partly due to
decoration, although there is likely more involved. 
-- 01/06/2000 hobbs
User Comments: nemethi (claiming to be Csaba Nemethi) added on 2026-11-06 15:23:11:

Serhiy, many thanks for fixing this very long-standing bug! Merged into main, core-9-0-branch, and core-8-6-branch by commits [6790de40], [9b7360b5], and [ce2acb17].


serhiy.storchaka added on 2026-11-05 20:36:18:

The Mutter case is [6fa2369ec6]: Mutter reports the invisible borders of its frame window in _GTK_FRAME_EXTENTS. With the proposed fix in branch x11-gtk-frame-extents, wm geometry .t +20+80 is reported as +20+80 after wm withdraw and wm deiconify. I cannot reproduce the metacity case now.


serhiy.storchaka added on 2026-09-30 13:51:37:

A related case is [d1a711b500]: wm geometry reports the requested position if the window manager moved the window out of the area reserved for a panel. The proposed fix there also makes wm geometry report the position of the decorated window under KWin on Wayland (Xwayland), which decorates windows without reparenting them, as for reparenting window managers.

The general issue remains: after wm withdraw and wm deiconify, metacity and mutter report the position of the frame window, which includes invisible borders, so wm geometry .t +20+80 is then reported as +-6+57 under metacity and +6+68 under mutter.


patthoyts added on 2010-01-04 22:04:15:
This does not occur with 8.5.7 on Windows but can be seen with 8.5.8 on Linux (compiz window manager). The [winfo] and [wm] geometries agree though.

% package require Tk
8.5.8
% winfo geometry .
200x200+1+25
% wm geometry . 200x200+120+120
% wm geometry .
200x200+121+145
% winfo geometry .
200x200+121+145
% list [winfo x .] [winfo y .] [winfo rootx .] [winfo rooty .]
121 145 121 145

pcaffin added on 2008-03-13 16:45:14:
Logged In: YES 
user_id=2029956
Originator: NO

I've tested this with Tk 8.5.2.0.284581 (teapot) on Linux and can confirm this bug is still present.

jenglish added on 2003-09-02 00:28:55:
Logged In: YES 
user_id=68433

Reassigning to me.  I know what causes this (details too
long to fit in the margin of this bug report), and it will
probably take a TIP to fix.

tak416 added on 2002-12-22 03:28:43:
Logged In: YES 
user_id=564039

I submitted this finding to tcl trackers. the tracker id is
657282 assigned to das. I also submiitted the test program
so you should be testing it and see the behavior.

Thanks,

dgp added on 2002-07-06 05:03:34:
Logged In: YES 
user_id=80530


this little buggy got lost.  will you give him a home?

dkf added on 2000-12-01 03:11:41:
winfo geometry gives info about the window itself, and wm geometry gives info about the decorated window (which is actually a separate window wrapped about Tk.)  Well, normally at least.  The specification of what the wm version *actually* means is somewhat non-obvious since it is really dealing with a hint to the window manager as to where the program/user would like the window to go, but the manner of implementation of this hint is not always obvious. Or seemingly even always consistent.  :^(

IOW, I've no idea if this is a real bug, or just strangeness completely beyond our control...