Re: [dev] [patch] map new window *after* arrange()

2011-07-09 Thread garbeam 
Hi Peter,

On 9 July 2011 00:40, Peter Hartlich bq...@hartlich.com wrote:
 I reviewed your patch and concluded I can't apply it as is, because
 focus() would be called twice, once on the unmapped window, and then
 on again.

 It seems to me that focusing and arranging should be independent
 operations. So here comes another patch to remove focus(NULL) from
 arrange and add it, if missing, everywhere arrange is called after
 either a window might have vanished from selmon's view, or selmon
 itself might have vanished.

I agree with this and applied both patches. It is not my ideal
solution, but I postpone an ideal version for 6.0.

All, please test hg tip. If there are no issues this will become 5.9 tomorrow.

Cheers,
--garbeam

Re: [dev] [patch] map new window *after* arrange()

2011-07-09 Thread Bogdan Ionuț 
On Sat, Jul 9, 2011 at 10:01, garbeam garb...@gmail.com wrote:

 Hi Peter,

 On 9 July 2011 00:40, Peter Hartlich bq...@hartlich.com wrote:
  I reviewed your patch and concluded I can't apply it as is, because
  focus() would be called twice, once on the unmapped window, and then
  on again.
 
  It seems to me that focusing and arranging should be independent
  operations. So here comes another patch to remove focus(NULL) from
  arrange and add it, if missing, everywhere arrange is called after
  either a window might have vanished from selmon's view, or selmon
  itself might have vanished.

 I agree with this and applied both patches. It is not my ideal
 solution, but I postpone an ideal version for 6.0.

 All, please test hg tip. If there are no issues this will become 5.9
 tomorrow.

 Cheers,
 --garbeam

 just one annoying issue with floating apps. eg. running save as in gimp,
will open a window in the background.

Re: [dev] [patch] map new window *after* arrange()

2011-07-09 Thread Peter Hartlich 
 just one annoying issue with floating apps. eg. running save as in gimp,
 will open a window in the background.

Oh man, sorry. I should've tested it more with floating windows.

So arrange calls arrangemon, which calls restack, but the new floating
window hasn't become selmon-sel yet.

Restacking again at the end of manage fixes it, but then there's
another redraw. To prevent the first restack, maybe something like
the attached patch? Or split off (un)select functions from (un)focus,
which also came up in connection to virtual keyboards some time ago.

Regards,
Peter
--- dwm/dwm.c.orig  2011-07-10 02:34:59.982409153 +0200
+++ dwm/dwm.c   2011-07-10 03:49:19.114387186 +0200
@@ -1150,9 +1150,12 @@ manage(Window w, XWindowAttributes *wa)
attachstack(c);
XMoveResizeWindow(dpy, c-win, c-x + 2 * sw, c-y, c-w, c-h); /* 
some windows require this */
setclientstate(c, NormalState);
+   unfocus(selmon-sel, False);
+   selmon-sel = NULL;
arrange(c-mon);
XMapWindow(dpy, c-win);
focus(c);
+   restack(c-mon);
 }
 
 void

Re: [dev] [patch] map new window *after* arrange()

2011-07-09 Thread Peter Hartlich 
 To prevent the first restack, maybe something like the attached patch?

... which causes the title bar to flicker, as restack calls drawbar
before bailing out. Not sure why that's needed.

Peter