On Mon, Mar 26, 2007 at 08:34:08PM +0200, Johannes Klauser wrote: > >The behavior of vanille dwm-3.8:incnmaster() is intended as is. > >Actually manipulating nmaster has no side-effects and doesn't > >depends on the layout in use or the amount of windows managed by > >dwm in view. > I did the patch because if nmaster is different from aktual view, you see > no effect by pressing "MODKEY|ShiftMask,XK_j/k". If you didn't pay > attention to this (or if you press it for some seconds for some reason) > you don't know the actual value of nmaster. Therefore it may be useful to > limit nmaster (by number of clients of actual tag, or (beeing more > konservative) by number of all clients), or having a view of the actual > nmaster value (e.g. at the topbar).
I know this problem. But it is intended, because nmaster is independent from any amount of windows - calling incnmaster(NULL) will restore the nmaster value to the initial one (NMASTER) defined in config.h - so if you assume #define NMASTER 666 your patch would break anyways... Actually, a more straight solution seems to me replacing incnmaster() with setnmaster() which sets the nmaster value directly. Usually I rarely use setups with nmaster > 3. > The second approach (nmaster at topbar) may be the best solution to make > dwm more comfortable without changing the current nmaster implementation. I tested this in some release (I think it was 2.8), but I didn't liked this approach. Regards, -- Anselm R. Garbe >< http://www.suckless.org/ >< GPG key: 0D73F361
