From: Benjamin Stürz <benni+openbsd@stuerz.xyz>
Subject: Re: [new/wip] sysutils/apmtop
To: Omar Polo <op@omarpolo.com>
Cc: James Cook <falsifian@falsifian.org>, ports@openbsd.org
Date: Thu, 11 Jan 2024 22:01:21 +0100

Hi,

thanks for the contributions.
I have attached a new version of the port.

On 1/11/24 11:14, Omar Polo wrote:
> Hello,
> 
> Well, the port actually looks good, even if a manpage (or at least
> installing the README) would be useful :)
apmtop(1) now has a man page.
It's still quite bare bones, but so is apmtop itself.
> 
> I tried it on my desktop and I can say it seems to mostly work.  It
> report the cpu temperature as -1 C which I assume means 'not available'.
> There is actually one sensor on this machine that reports a temperature,
> but since it doesn't contain the "cpu" string in its name it gets
> skipped.
> 
> 	% sysctl -a | grep degC
> 	hw.sensors.ksmn0.temp0=25.25 degC (Tctl)
It looks like ksmn(4) is actually a CPU temperature sensor.
But why isn't it called 'hw.sensors.cpu0.temp0' then?
> 
> it also reports the power as 0, but I guess it's just a metric not
> available on my handware.
> 
> (BAT is obviously 0 for me too, but it's a desktop)
> 
>>> [...]
>>> Nice utility! Don't know if I'll use it since I like to squeeze such
>>> things into my status bar, but maybe I can learn from it. One annoyance:
>>> xterm flashes on every update as the TUI is redrawn, which is
>>> distracting when I'm trying to look at something else.
> 
> I agree, it's a nice little utility :)
Thanks, I'll continue improving it,
once I'm done studying for my exams :)
> 
> I got a bit annoyed by the screen flashes at each update and gave it a
> try at improving it, see diff below.
Thanks, I wasn't able to solve it on my own.
But I did solve the wrong-temperature-bug.
It was a simple conversion mistake.
> 
>> I too put these things into my status bar,
>> but I wanted to have a nice graph,
>> to make optimizing battery life a little easier.
> 
> Index: apmtop.c
> --- apmtop.c.orig
> +++ apmtop.c
> @@ -259,7 +259,7 @@ draw_graph (struct graph *g, const char *fmt, int min,
>   	w -= 2;
>   	h -= 3;
>   
> -	wclear (g->win);
> +	werase (g->win);
>   
>   	for (int i = 0; i < entries; ++i) {
>   		int v = g->log[i];
> @@ -287,7 +287,7 @@ draw_graph (struct graph *g, const char *fmt, int min,
>   static void
>   draw (struct display *dpy)
>   {
> -	clear ();
> +	erase ();
>   
>   	draw_graph (&dpy->g[G_CPUUSAGE],	"CPU: %d%%",	0, 100);
>   	draw_graph (&dpy->g[G_CPUTEMP],		"CPU: %d C",	0, 100);
> @@ -302,12 +302,10 @@ draw (struct display *dpy)
>   
>   	printw (" %c delay=%d", dpy->running ? 'R' : 'S', dpy->delay);
>   
> -	refresh ();
> +	wnoutrefresh (stdscr);
>   	for (int i = 0; i < NGRAPH; ++i)
>   		wnoutrefresh (dpy->g[i].win);
>   	doupdate ();
> -	move (0, 0);
> -
>   }
>   
>   static void
>