On Wed, Sep 10, 2008 at 10:44:05AM +0200, Martin von Gagern <[EMAIL PROTECTED]> wrote: > Debugging the issue I found out that while kwin understands > MWM_HINTS_DECORATIONS, If it indeed does, it should say so by creating the _MOTIF_WM_INFO atom which window managers implementing mwm hints have to to declare their compatibility to mwm.
> having kwin handle the window seems the cleaner approach to me. Just teach kwin the mwm_hints protocol then, which urxvt and many other apps support for just that case. Or use a working window manager, such as fvwm, which has mwm hints support implement for a very long time. > for _MOTIF_WM_INFO. People who want borderless windows on a window > manager that doesn't honour MWM_HINTS_DECORATIONS would be required to > enable override-redirect mode manually in order to get the old behaviour. I don't think people who want a borderless window should jump through hoops just because of one broken window manager. Note that the borderless mode is designed to result in a borderless window (e.g. for status displays), it's not meant to do the job of the wm and change decorations - if you want that, teach your wm to change how the rxvt window should look like, this is not the job of urxvt. If your wm fails to do that, use a better one. > MWM_HINTS_DECORATIONS and the interdependence between these two. I guess > there is a high probability you will be going to argue that rxvt-unicode > behaves correctly and instead kwin is broken. kwin doesn't support mwm hints, so the alternative to override-redirect is total failure - not very useful. > http://bugzilla.xfce.org/show_bug.cgi?id=1877 . STill, patching a single > failing application instead of an otherwise working window manager seems > the pragmatic solution to me. Of course rxvt-unicode is not the only failing application as you want it to look like: Try google codesearch and you will easily find lots of applications relying on correct implementation of the mwm hints protocol (trivial example: eterm, which additionally has no explicit override-redirect switch). Note that *the single failing application* is, like so many times in the past, kwin here, so if you take your pragmatic approach seriously, the obvious fix is to teach kwin mwm hints. Note that kwin is not an "otherwise working window manager" - kwin is a low quality window manager that fails a lot of properly working applications. Making lots of applications jump through hoops and implement hacks (instead of relying on de-facto protocols much older than kwin that kwin fails to implement) just to please kwin is absurd. If kwin wants to support mwm hints it has to implement the protocol for it, as virtually every other window manager does. Fixing all applications because of kwin's lack of correctness is not going to scale. kwin needs to implement the protocols just like everybody else. > And auto-enabling override-redirect mode feels bad in any case. Instead you want urxvt to silently fail - not an option. > Would you apply this patch to the main repository? Certainly not - it would break things. -- The choice of a Deliantra, the free code+content MORPG -----==- _GNU_ http://www.deliantra.net ----==-- _ generation ---==---(_)__ __ ____ __ Marc Lehmann --==---/ / _ \/ // /\ \/ / [EMAIL PROTECTED] -=====/_/_//_/\_,_/ /_/\_\ _______________________________________________ rxvt-unicode mailing list [email protected] http://lists.schmorp.de/cgi-bin/mailman/listinfo/rxvt-unicode
