[9front] rio autosize on null sweep

[9front] rio autosize on null sweep

[umbraticus]

I tend to stack my rio windows.  If you do too you may
enjoy the patch below, which takes some inspiration
from sam and interprets a bad sweep (in practice, a
click instead of a sweep) as a request for dimensions
equal to the those of either the window clicked on or
else the whole screen.  You can still cancel a sweep
with b1 or b2.  Perhaps there's a cleverer
pseudo-tiling thing that could be hooked in here but
this was good enough for my use.

umbraticus

diff 9126ee3eea90d639f4e877c01400248581d10f65 uncommitted
--- a//sys/src/cmd/rio/rio.c
+++ b//sys/src/cmd/rio/rio.c
@@ -805,6 +805,7 @@
 sweep(void)
 {
 	Image *i, *oi;
+	Window *w;
 	Rectangle r;
 	Point p0, p;
 
@@ -837,8 +838,13 @@
 	}
 	if(mouse->buttons != 0)
 		goto Rescue;
-	if(i==nil || !goodrect(r))
-		goto Rescue;
+	if(i == nil || !goodrect(r)){
+		w = wpointto(p0);
+		r = w == nil ? screen->r : w->screenr;
+		i = allocwindow(wscreen, r, Refnone, DNofill);
+		if(i == nil)
+			goto Rescue;
+	}
 	oi = i;
 	i = allocwindow(wscreen, oi->r, Refbackup, DNofill);
 	freeimage(oi);

Re: [9front] rio autosize on null sweep

[qwx]

On Wed Jun  1 12:17:35 +0200 2022, umbraticus@prosimetrum.com wrote:
> I tend to stack my rio windows.  If you do too you may
> enjoy the patch below, which takes some inspiration
> from sam and interprets a bad sweep (in practice, a
> click instead of a sweep) as a request for dimensions
> equal to the those of either the window clicked on or
> else the whole screen.  You can still cancel a sweep
> with b1 or b2.  Perhaps there's a cleverer
> pseudo-tiling thing that could be hooked in here but
> this was good enough for my use.
> 
> umbraticus

Nice, thanks :)

Cheers,
qwx