And hard-bind the keys Ctrl+Up and Ctrl+Down to these functions.

Unlike the paragraph-jumping code, these new functions disregard
any indentation, treating only blank lines as separators.  They
also do an automatic smart home when that option is set.

This fulfills https://savannah.gnu.org/bugs/?48291.

The bug with a leading wide character has been fixed since ncurses-5.9.

See https://savannah.gnu.org/bugs/?31743 for reference.

Also, compose the coloring attributes during initialization
instead of every time when painting something.
Signed-off-by: Rishabh Dave <rishabhddave@gmail.com>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

Instead of referring to a window that might have been deleted by a
resize, just turn the keypad back on for the two windows from which
input might have been requested.

Also: don't insert KEY_WINCH as the verbatim keystroke.

This fixes https://savannah.gnu.org/bugs/?48532.

This fixes https://savannah.gnu.org/bugs/?48538.

This fixes https://savannah.gnu.org/bugs/?48455.

When the user changes her mind after having pressed Esc, she cannot
unpress Esc, so do that for her by ignoring the escape code when
the subsequent keycode is outside of the normal printable range.

This restores the behavior from before nano-2.3.5 -- except that
Ctrl+Alt+key continues to report an unbound key.

This fixes https://savannah.gnu.org/bugs/?48459.

This prevents many keycode sequences from being misconstrued when a
command key plus other keystrokes accumulate in the input buffer.

This fixes https://savannah.gnu.org/bugs/?48395.

Color-pair numbers and attributes can be OR'd together -- do so,
to save an attron() call whenever the hilite is on.

Also condense some comments, and group keys with the same result together.

Add them so that <Esc> followed by <Home> will simply do "Home" instead of
reporting an unbound key.  The same for <End>, <PageUp> and <PageDown>.

This restores the behavior from before nano-2.3.5.

Spaces and tabs and control codes never are multi-column glyphs, so
only look up the width for "normal", visible characters.

This avoids having to null-terminate every single-byte character.

Just allocate ample space up front and thus discard the delaying 'if' for
each and every character.  In most cases this will allocate far too much,
but that hardly matters: it is freed again as soon as the line is printed.

This fixes https://savannah.gnu.org/bugs/?48125.

(The fix is slightly wasteful; speeding things up will follow later.)

Any control character is represented by a ^ plus an ASCII character.

This fixes the first part of https://savannah.gnu.org/bugs/?48331

.
Reported-by: Mike Frysinger <vapier@gentoo.org>

This fixes https://savannah.gnu.org/bugs/?47962 reported by Cody Taylor.

Keystrokes are single integers (aided by the flags meta_key and
func_key) but in the input stream they can be encoded as escape
sequences (a series of bytes).  Characters are values in byte
range, but in UTF-8 one character can consist of multiple bytes.

Also rename two variables, because the secondary input buffer
that they refer to contains only characters (mostly just one),
never any shortcuts; and there are too many "kbinput" already.

Typing a Unicode code is always finished after at most six characters:
either retval == uni or retval == kbinput, but it can't be ERR.

Switch to https and drop the www. prefix for all occurrences of the
website URL.

Error messages about lock files should not get overwritten by purely
informational messages, only by alerting ones.

This fixes https://savannah.gnu.org/bugs/?47963.