Re: [dwm] nmaster strikes back

On 2009-02-11 9:56 +0100, Szabolcs Nagy wrote:
Hi,

> nice, but tweak() has too much magic (if there were many runtime
> tunable int parameters in dwm then it would be ok).

The nice thing about int + bounds checking is that it should be general
enough for other layout algorithms as well, even where they now might
use bools, chars, floats etc.

I have to agree though rel/abs handling does not belong in tweak() proper
and should be done by the caller instead.

>> Configuration example:
>> [...]

> this way master width cannot be set to a fixed value

{ MODKEY, XK_n, tweakmwp, {.v = (int[]){Abs, 60}} }

> is nmaster that useful?

It's pretty much essential for me on a small screen when I want to look
at two pieces of code at the same time.

>> keyboards thread. It also corrects a discrepancy between documentation and
>> code: mfact was described as [0.05..0.95], but was actually [0.1..0.9].

> nice catch, config.def.h or setmfact() should be fixed

Actually, now that adjustborder() gives an indication if there is more
than one tiled client, why not make the right boundary 100%? This way,
you get pure horizontal splitting (as was recently requested IIRC) for
free:

        #include <limits.h>
        ...
        { MODKEY, XK_x, tweakmn, {.v = (int[]){Abs, INT_MAX}} },
        { MODKEY, XK_x, tweakmwp, {.v = (int[]){Abs, 100}} },

With all that in mind, what do you think about the attached patch
(on top of the last)? It also does the bounds handling in tweak()
more constructively.

Regards,
Peter