Corrections for 20021228

[Philippe Blain]

>From Philippe Blain, Bordeaux, France.
My old computer: P133 - 8,4 Go - 32 Mo Red Hat Linux 7.0

To maintainers of 'ncurses'.(and to Mr Dickey)
Subject: Corrections for ncurses-5.3-20021228+

Here are some problems I found :

1---------------------------------------------------------------------------
--
File : ncurses/base/tty_update.c

Correct padding in the following functions :

_nc_scrolln()   lines 1683, 1694, 1729, 1740
Padding for 'change_scroll_region' is not proportionnal according to
terminfo
man pages (P and not P*).
It's simply a   putp(tparm(change_scroll_region, XXX, YYY)); on these lines.


ClearScreen()   line 1336
'clr_eos' is noted as proportionnal (P*)
Should be       tputs(clr_eos, screen_lines, _nc_outch);


EmitRange()     line 478
'repeat_char' is noted as proportionnal (P*)
Should be       tputs(tparm(repeat_chars, CharOf(ntext0), rep_count),
                    rep_count, _nc_outch);

2---------------------------------------------------------------------------
File : ncurses/base/tty_update.c

ClrToEOL()     line 907
Safer checking 'clr_eol' and inversing the test:

if (needclear) {
    UpdateAttrs (AttrOf (blank));
    TPUTS_TRACE ("clr_eol");
==> if (clr_eol && SP->_el_cost <= (screen_columns - SP->_curscol)) {
==>     putp (clr_eol);
    }
    else {
        int count = (screen_columns - SP->_curscol);
        while (count-- > 0)
            PutChar (CHREF (blank));
    }
}

3---------------------------------------------------------------------------
File : ncurses/base/tty_update.c

The second part of ClrBottom (looking for blank chars in curscr) is
useless and prevents hard-clearing all blank lines when there are less
blank lines at bottom of curscr than in newscr.
Only look in newscr, then do a ClrToEOS() which blanks corresponding lines
in curscr.

static int ClrBottom (int total)
{
    int row;
    int col;
    int top = total;
    int last = min (screen_columns, newscr->_maxx + 1);
    NCURSES_CH_T blank = newscr->_line[total - 1].text[last - 1];
    bool ok;

    if (clr_eos && can_clear_with (CHREF (blank))) {

        for (row = total - 1; row >= 0; row--) {
            for (col = 0, ok = TRUE; ok && col < last; col++) {
                ok = (CharEq (newscr->_line[row].text[col], blank));
            }
            if (!ok) break;
==>         top = row;
        }

/* As we are in the case where 'clr_eos' (ed=\E[J) exists, there is no
disavantage using it here for just one line, instead of using
clr_eol/clr_bol in TransformLine() where we are going to check a second time
if all chars of that line are blank */

==>     if (top < total) {
            GoTo (top, 0);
            ClrToEOS (blank);
            if (SP->oldhash && SP->newhash) {
                for (row = top; row < screen_lines; row++)
                    SP->oldhash[row] = SP->newhash[row];
            }
        }
    }
==> return (top);
}

----------------------------------------------------------------------------
- Philippe