movetodel function in acme not called

movetodel function in acme not called

[Tobias Heinicke]

[-- Attachment #1.1.1: Type: text/plain, Size: 1373 bytes --]

Hi,

first of all I would like to thank you for developing 9front and thus
keeping plan9 alive and kicking.

I was just wondering whether there is a reason for movetodel() being
committed, without being ever used. The function was added in p9p as
part of rsc's effort to allow deleting several acme windows in a column
without having to move the mouse pointer (committed on 2012-09-23 -
37f8ed2410ad5cbd46eda00a77f8bf4950bcf544).

In 9front the function definition was added on 2013-08-29 (Node ID:
3af7b07be732a15a30e323134f8056986ca4e0ed), but the code to call it is
missing (and thus deleting a sequence of windows in 9front's acme still
requires moving the mouse a lot).

So far I have recursively grep'd the mail archives for 'movetodel', but
have not found anything. If the behaviour implemented by rsc is
desirable I suggest adding it.

Attached is a diff of the code that is missing for rsc's feature to be
available in 9front's acme (created using hg diff, if you would rather
like to see a patch (e.g. committing the changes and hg export or
sth.?), feel free to tell me). Sorry for probably not complying to any
rules, as I am fairly new to using 9front (and been lurking this mailing
list even less [- subscribed with a different address]).

Best regards,

Tobias Heinicke

P.S.: Sorry for my English, I am not a native speaker.

[-- Attachment #1.1.2: utilizeMovetodelDiff --]
[-- Type: text/plain, Size: 2001 bytes --]

diff -r 0b97536e2bcd sys/src/cmd/acme/cols.c
--- a/sys/src/cmd/acme/cols.c	Mon Feb 05 19:16:29 2018 +0100
+++ b/sys/src/cmd/acme/cols.c	Tue Feb 13 08:45:51 2018 +0100
@@ -111,7 +111,7 @@
 colclose(Column *c, Window *w, int dofree)
 {
 	Rectangle r;
-	int i;
+	int i, didmouse, up;
 
 	/* w is locked */
 	if(!c->safe)
@@ -125,7 +125,7 @@
 	w->tag.col = nil;
 	w->body.col = nil;
 	w->col = nil;
-	restoremouse(w);
+	didmouse = restoremouse(w);
 	if(dofree){
 		windelete(w);
 		winclose(w);
@@ -137,17 +137,24 @@
 		draw(screen, r, display->white, nil, ZP);
 		return;
 	}
+	up = 0;
 	if(i == c->nw){		/* extend last window down */
 		w = c->w[i-1];
 		r.min.y = w->r.min.y;
 		r.max.y = c->r.max.y;
 	}else{			/* extend next window up */
+		up = 1;
 		w = c->w[i];
 		r.max.y = w->r.max.y;
 	}
 	draw(screen, r, textcols[BACK], nil, ZP);
-	if(c->safe)
+	if(c->safe) {
+		if(!didmouse && up)
+			w->showdel = TRUE;
 		winresize(w, r, FALSE);
+		if(!didmouse && up)
+			movetodel(w);
+	}
 }
 
 void
diff -r 0b97536e2bcd sys/src/cmd/acme/fns.h
--- a/sys/src/cmd/acme/fns.h	Mon Feb 05 19:16:29 2018 +0100
+++ b/sys/src/cmd/acme/fns.h	Tue Feb 13 08:45:51 2018 +0100
@@ -19,10 +19,11 @@
 char*	getname(Text*, Text*, Rune*, int, int);
 void	scrsleep(uint);
 void	savemouse(Window*);
-void	restoremouse(Window*);
+int		restoremouse(Window*);
 void	clearmouse(void);
 void	allwindows(void(*)(Window*, void*), void*);
 uint loadfile(int, uint, int*, int(*)(void*, uint, Rune*, int), void*);
+void	movetodel(Window*);
 
 Window*	errorwin(Mntdir*, int);
 Window*	errorwinforwin(Window*);
diff -r 0b97536e2bcd sys/src/cmd/acme/util.c
--- a/sys/src/cmd/acme/util.c	Mon Feb 05 19:16:29 2018 +0100
+++ b/sys/src/cmd/acme/util.c	Tue Feb 13 08:45:51 2018 +0100
@@ -366,12 +366,18 @@
 	mousew = w;
 }
 
-void
+int
 restoremouse(Window *w)
 {
-	if(mousew!=nil && mousew==w)
+	int did;
+
+	did = 0;
+	if(mousew!=nil && mousew==w) {
 		moveto(mousectl, prevmouse);
+		did = 1;
+	}
 	mousew = nil;
+	return did;
 }
 
 void

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 516 bytes --]

Re: [9front] movetodel function in acme not called

[Ethan Grammatikidis]

On Tue, Feb 13, 2018, at 10:02 AM, Tobias Heinicke wrote:
> 
> If the behaviour implemented by rsc is
> desirable I suggest adding it.

speaking as an acme user, the behaviour is somewhat 
desirable. it's one of a myriad tweaks which mitigate the 
discomfort caused by acme's window system design.

as might be obvious from my tone, i'm only using acme 
until i can find or (preferably) write another editor i can 
tolerate. still, if anyone is in the mood for fixing acme, 
there's also something wrong with the X command. 
(probably Y too.)

someone, in some mailing list, (i'm very sorry for the 
vagueness,) writing about acme on inferno,
found a variable pertaining to the X command 
could be used uninitialized. i took notice because 
i've developed a certain amount of fear of using the 
X command in acme on 9front. strange things may 
happen during or after using it, including sudden exits 
and freezes without any apparent cause.

again, i'm sorry for the vagueness. i didn't save the 
inferno mail because i was going through a hard time, 
and i don't want to look at acme's source because i'm 
still struggling with my code-reading skills.

-- 
The lyf so short, the craft so long to lerne. -- Chaucer