| Ticket UUID: | 1233834 | |||
| Title: | Restrictions on window resizing in Motif wms | |||
| Type: | Bug | Version: | obsolete: 8.4.11 | |
| Submitter: | hobbs | Created on: | 2005-07-06 22:55:49 | |
| Subsystem: | 67. Unix Window Operations | Assigned To: | jenglish | |
| Priority: | 5 Medium | Severity: | ||
| Status: | Open | Last Modified: | 2026-09-23 19:35:18 | |
| Resolution: | None | Closed By: | ||
| Closed on: | ||||
| Description: |
It was reported that the attached code (admittedly arcane, but working) that does a pseudo-wizard still toplevel no longer resizes properly when moving to the 2nd screen (which is larger). It apparently works OK with non-Motif wms. | |||
| User Comments: |
serhiy.storchaka added on 2026-09-23 19:35:18:
Reproduced with 9.1b1 and the attached script under mwm 2.3.8: page B keeps the size of page A (221x216 while it requests 484x536). The same script resizes correctly under Window Maker, metacity, xfwm4, icewm, i3, dwm, ctwm, fvwm and with no window manager at all, as the report says. The trigger is wm resizable .t 0 0 followed by wm resizable .t 1 1 on a mapped toplevel: every later change of the requested size is then ignored by mwm. With only 0 0, or without the toggle, mwm resizes the window as asked, and wm geometry and wm minsize have nothing to do with it. Cause: while the window is not resizable Tk sends PMaxSize equal to the current size, and when it becomes resizable again Tk simply omits the flag. mwm keeps the maximum it saw last instead of dropping it. Setting an explicit maximum before the resize cures it, and the retained value is clearly in use: wm maxsize .t 300 200 -> window clamped to 300x200 (requested 484x424) wm maxsize .t 600 500 -> window resized to 484x424 wm withdraw plus wm deiconify before the resize works as well, because mwm re-reads the hints then. Neither helps after the resize request was already refused. Tk could work around it: when a previously sent maximum is withdrawn, emit PMaxSize with the default maximum (the screen size, which wm maxsize reports anyway) instead of omitting the flag. But that changes the hints seen by every window manager, and some of them use the presence of a maximum to decide whether a window may be maximized or tiled, so it trades a risk on current desktops for a 30 year old window manager which, by ICCCM, should not retain a constraint the client no longer advertises. A script can avoid it with one wm maxsize call. jenglish added on 2005-07-07 07:18:26: Logged In: YES user_id=68433 More info: (16:14:37)tclguyit's supposed to be a 2-screen wizard (16:14:57)tclguythe 2nd screen is larger than the first (16:15:24)tclguyand the complaint was that with 8.4.9 and mwm, the 2nd screen remained the same size as the first screen - not desired This is possibly related to the patch for #568861, backorted to the 8-4 branch 2005-01-14; also possibly related to #1006686 (backported 2004-08-10) or #915350/#922336 (backported 2004-04-21) jenglish added on 2005-07-07 06:09:31: Logged In: YES user_id=68433 Clarification needed: what did it formerly do? What is it doing now? It doesn't seem OK to me under metacity (Gnome 2.8), but I'm not sure if the problems I'm seeing are the same as the problems the original reporter is concerned about. hobbs added on 2005-07-07 05:55:52: File Added - 141121: wm_resize_issue.tcl | |||
Attachments:
- wm_resize_issue.tcl [download] added by hobbs on 2005-07-07 05:55:49. [details]
