On Wed 2016.11.30 at 15:20 +0100, Vadim Vygonets wrote: > Quoth Okan Demirmen on Sat, Nov 26, 2016: > > Hopefully these new action names make more sense; they will allow adding > > some > > more functions that will more closely match what they do. > > The new action names are definitely better. I'm a bit divided > about breaking compatibility, but it's probably worth doing it > once.
Yeah, this is that binding on un-matched function names implies exec 'command', so no one would ever know, thus these bind keyword changes forces that issue. I suppose a future change could change the implicit 'command' binding to something more explicit, such as "bind-key M-m command /usr/local/bin/magic" While that requires a bit of surgery, I could make that a part of this overall change? > > diff -u -p -r1.224 conf.c > > @@ -510,6 +512,10 @@ conf_bind_key(struct conf *c, const char > > const char *key; > > unsigned int i; > > > > + if (strcmp(bind, "*") == 0) { > > + conf_unbind_key_all(c); > > + goto out; > > + } > > I'd rather move this down under "if (cmd == NULL)", to avoid > "bind-key * something". (And the same about bind-mouse.) Never thought anyone would do that, but I guess something like: if (cmd == NULL) { free(kb); if (strcmp(bind, "*") == 0) conf_unbind_key_all(c); goto out; } Likewise for bind-mouse. > > diff -u -p -r1.63 cwmrc.5 > > -.It group[n] > > +.It group-toggle-[n] > > Toggle visibility of group n, where n is 1-9. > > -.It grouponly[n] > > +.It group-only-[n] > > Like > > -.Ar group[n] > > +.Ar group-toggle-[n] > > but also hides the other groups. > > This is not strictly correct: group-only-[n] never hides group n. > Perhaps something like this would be better: > > Show group n, where n is 1-9, and hide the other groups. True; the current wording is that way, and I didn't change it. However, your point is taken, maybe even to the point of using "group-showonly-[n]" might be better?