+</PRE><H2><a name="h2-PORTABILITY">PORTABILITY</a></H2><PRE>
+ All these functions are described in the XSI Curses standard, Issue 4.
+ The defaults specified for forms-drawing characters apply in the POSIX
+ locale.
+
+
+</PRE><H3><a name="h3-ACS-Symbols">ACS Symbols</a></H3><PRE>
+ X/Open Curses states that the <STRONG>ACS_</STRONG> definitions are <STRONG>char</STRONG> constants. For
+ the wide-character implementation (see <STRONG>curs_add_wch</STRONG>), there are
+ analogous <STRONG>WACS_</STRONG> definitions which are <STRONG>cchar_t</STRONG> constants. Some
+ implementations are problematic:
+
+ <STRONG>o</STRONG> Some implementations define the ACS symbols to a constant (such as
+ Solaris), while others define those to entries in an array.
+
+ This implementation uses an array <STRONG>acs_map</STRONG>, as done in SVr4 curses.
+ NetBSD also uses an array, actually named <STRONG>_acs_char</STRONG>, with a <STRONG>#define</STRONG>
+ for compatibility.
+
+ <STRONG>o</STRONG> HPUX curses equates some of the <STRONG>ACS_</STRONG> symbols to the analogous <STRONG>WACS_</STRONG>
+ symbols as if the <STRONG>ACS_</STRONG> symbols were wide characters. The
+ misdefined symbols are the arrows and other symbols which are not
+ used for line-drawing.
+
+ <STRONG>o</STRONG> X/Open Curses (issues 2 through 7) has a typographical error for
+ the ACS_LANTERN symbol, equating its "VT100+ Character" to <STRONG>I</STRONG>
+ (capital I), while the header files for SVr4 curses and the various
+ implementations use <STRONG>i</STRONG> (lowercase).
+
+ None of the terminal descriptions on Unix platforms use uppercase-
+ I, except for Solaris (i.e., <STRONG>screen</STRONG>'s terminal description,
+ apparently based on the X/Open documentation around 1995). On the
+ other hand, the terminal description <EM>gs6300</EM> (AT&T PC6300 with EMOTS
+ Terminal Emulator) uses lowercase-i.
+
+ Some ACS symbols (ACS_S3, ACS_S7, ACS_LEQUAL, ACS_GEQUAL, ACS_PI,
+ ACS_NEQUAL, ACS_STERLING) were not documented in any publicly released
+ System V. However, many publicly available terminfos include <STRONG>acsc</STRONG>
+ strings in which their key characters (pryz{|}) are embedded, and a
+ second-hand list of their character descriptions has come to light.
+ The ACS-prefixed names for them were invented for <STRONG><A HREF="ncurses.3x.html">ncurses(3x)</A></STRONG>.
+
+ The <EM>displayed</EM> values for the <STRONG>ACS_</STRONG> and <STRONG>WACS_</STRONG> constants depend on
+
+ <STRONG>o</STRONG> the library configuration, i.e., <STRONG>ncurses</STRONG> versus <STRONG>ncursesw</STRONG>, where the
+ latter is capable of displaying Unicode while the former is not,
+ and
+
+ <STRONG>o</STRONG> whether the <EM>locale</EM> uses UTF-8 encoding.
+
+ In certain cases, the terminal is unable to display line-drawing
+ characters except by using UTF-8 (see the discussion of
+ <STRONG>NCURSES_NO_UTF8_ACS</STRONG> in <STRONG><A HREF="ncurses.3x.html">ncurses(3x)</A></STRONG>).
+
+
+</PRE><H3><a name="h3-Character-Set">Character Set</a></H3><PRE>
+ X/Open Curses assumes that the parameter passed to <STRONG>waddch</STRONG> contains a
+ single character. As discussed in <STRONG><A HREF="curs_attr.3x.html">curs_attr(3x)</A></STRONG>, that character may
+ have been more than eight bits in an SVr3 or SVr4 implementation, but
+ in the X/Open Curses model, the details are not given. The important
+ distinction between SVr4 curses and X/Open Curses is that the non-
+ character information (attributes and color) was separated from the
+ character information which is packed in a <STRONG>chtype</STRONG> to pass to <STRONG>waddch</STRONG>.
+
+ In this implementation, <STRONG>chtype</STRONG> holds an eight-bit character. But
+ ncurses allows multibyte characters to be passed in a succession of
+ calls to <STRONG>waddch</STRONG>. The other implementations do not do this; a call to
+ <STRONG>waddch</STRONG> passes exactly one character which may be rendered as one or
+ more cells on the screen depending on whether it is printable.
+
+ Depending on the locale settings, ncurses will inspect the byte passed
+ in each call to <STRONG>waddch</STRONG>, and check if the latest call will continue a
+ multibyte sequence. When a character is <EM>complete</EM>, ncurses displays the
+ character and moves to the next position in the screen.
+
+ If the calling application interrupts the succession of bytes in a
+ multibyte character by moving the current location (e.g., using <STRONG>wmove</STRONG>),
+ ncurses discards the partially built character, starting over again.
+
+ For portability to other implementations, do not rely upon this
+ behavior:
+
+ <STRONG>o</STRONG> check if a character can be represented as a single byte in the
+ current locale before attempting call <STRONG>waddch</STRONG>, and
+
+ <STRONG>o</STRONG> call <STRONG>wadd_wch</STRONG> for characters which cannot be handled by <STRONG>waddch</STRONG>.
+
+
+</PRE><H3><a name="h3-TABSIZE">TABSIZE</a></H3><PRE>
+ The <STRONG>TABSIZE</STRONG> variable is implemented in SVr4 and other versions of
+ curses, but is not part of X/Open curses (see <STRONG><A HREF="curs_variables.3x.html">curs_variables(3x)</A></STRONG> for
+ more details).
+
+ If <EM>ch</EM> is a carriage return, the cursor is moved to the beginning of the
+ current row of the window. This is true of other implementations, but
+ is not documented.
+
+
+</PRE><H2><a name="h2-SEE-ALSO">SEE ALSO</a></H2><PRE>
+ <STRONG><A HREF="ncurses.3x.html">curses(3x)</A></STRONG>, <STRONG><A HREF="curs_attr.3x.html">curs_attr(3x)</A></STRONG>, <STRONG><A HREF="curs_clear.3x.html">curs_clear(3x)</A></STRONG>, <STRONG><A HREF="curs_inch.3x.html">curs_inch(3x)</A></STRONG>,
+ <STRONG><A HREF="curs_outopts.3x.html">curs_outopts(3x)</A></STRONG>, <STRONG><A HREF="curs_refresh.3x.html">curs_refresh(3x)</A></STRONG>, <STRONG><A HREF="curs_variables.3x.html">curs_variables(3x)</A></STRONG>, <STRONG>putc(3)</STRONG>.
+
+ Comparable functions in the wide-character (ncursesw) library are
+ described in <STRONG><A HREF="curs_add_wch.3x.html">curs_add_wch(3x)</A></STRONG>.