Bug in echotc ?

Bug in echotc ?

[Bart Schaefer]

I was fooling with something related to the discussion of nopromptcr
workarounds.  I tried this:

schaefer[505] echotc cm $LINES $COLUMNS ; sleep 5; echo here
                        here

That doesn't look like column 80 to me, but I checked "man termcap" which
says (and "man terminfo" confirms):

       cm   Cursor move to row %1 and column %2 (on screen)

Nevertheless, I tried "echotc cm $COLUMNS $LINES" instead, and got what I
expected the first to have done.

Is echotc broken, or is the manual page confused?

(Using "echoti cup $LINES $COLUMNS" works correctly.)

Re: Bug in echotc ?

[Clint Adams]

> Is echotc broken, or is the manual page confused?

echotc is broken.  I think it's still broken after this patch.

Index: Src/Modules/termcap.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/termcap.c,v
retrieving revision 1.18
diff -u -r1.18 termcap.c
--- Src/Modules/termcap.c	7 Dec 2004 16:55:11 -0000	1.18
+++ Src/Modules/termcap.c	2 Mar 2005 15:04:42 -0000
@@ -155,7 +155,7 @@
 	tputs(t, 1, putraw);
     else {
 	num = (argv[1]) ? atoi(argv[1]) : atoi(*argv);
-	tputs(tgoto(t, atoi(*argv), num), num, putraw);
+	tputs(tgoto(t, num, atoi(*argv)), num, putraw);
     }
     return 0;
 }

Re: Bug in echotc ?

[Clint Adams]

> echotc is broken.  I think it's still broken after this patch.

This is probably a bit more appropriate.

Index: Src/Modules/termcap.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/termcap.c,v
retrieving revision 1.19
diff -u -r1.19 termcap.c
--- Src/Modules/termcap.c	2 Mar 2005 15:12:30 -0000	1.19
+++ Src/Modules/termcap.c	2 Mar 2005 16:10:10 -0000
@@ -154,8 +154,9 @@
     if (!argct)
 	tputs(t, 1, putraw);
     else {
+	/* This assumes arguments of <lines> <columns> for cap 'cm' */
 	num = (argv[1]) ? atoi(argv[1]) : atoi(*argv);
-	tputs(tgoto(t, num, atoi(*argv)), num, putraw);
+	tputs(tgoto(t, num, atoi(*argv)), 1, putraw);
     }
     return 0;
 }

Re: Bug in echotc ?

[Bart Schaefer]

On Mar 2, 11:10am, Clint Adams wrote:
} Subject: Re: Bug in echotc ?
}
} > echotc is broken.  I think it's still broken after this patch.
} 
} This is probably a bit more appropriate.
} 
}      else {
} +	/* This assumes arguments of <lines> <columns> for cap 'cm' */
}  	num = (argv[1]) ? atoi(argv[1]) : atoi(*argv);
} -	tputs(tgoto(t, num, atoi(*argv)), num, putraw);
} +	tputs(tgoto(t, num, atoi(*argv)), 1, putraw);
}      }

Hmm.
       The  tputs  routine  applies  padding  information  to the
       string str and outputs it.  The str  must  be  a  terminfo
       string  variable  or the return value from tparm, tgetstr,
       or tgoto.  affcnt is the number of lines affected, or 1 if
       not  applicable.   putc is a putchar-like routine to which
       the characters are passed, one at a time.

So it seems like what we really want is more like

	num = atoi(*argv);
	t = tgoto(t, (argv[1] ? atoi(argv[1]) : num), num);
	tputs(t, num, putraw);

Because atoi(*argv) is "the number of lines affected".  Except it isn't,
quite, because really the number of lines affected is the difference
between the current line and the target line.  So maybe 1 is the right
value to use for the second argument of tputs() after all.  Except if
this isn't a "cm" but rather a relative line-motion such as "DO" or "UP"
then ... maybe what we want is

	num = atoi(*argv);
	if (argct == 1)
	    tputs(tgoto(t, num, num), num, putraw);
        else
            tputs(tgoto(t, atoi(argv[1]), num), 1, putraw);

Except it's not clear *that* does the right thing with "LE" or "RI".
Gag.

Re: Bug in echotc ?

[Clint Adams]

> Because atoi(*argv) is "the number of lines affected".  Except it isn't,
> quite, because really the number of lines affected is the difference
> between the current line and the target line.  So maybe 1 is the right
> value to use for the second argument of tputs() after all.  Except if

My thought was based on the assumption that affcnt is used only for
delay padding.

> this isn't a "cm" but rather a relative line-motion such as "DO" or "UP"
> then ... maybe what we want is
> 
> 	num = atoi(*argv);
> 	if (argct == 1)
> 	    tputs(tgoto(t, num, num), num, putraw);
>         else
>             tputs(tgoto(t, atoi(argv[1]), num), 1, putraw);
> 
> Except it's not clear *that* does the right thing with "LE" or "RI".
> Gag.

I think that there's probably a reason everyone says to use terminfo instead.