scanw, wscanw, mvscanw, mvwscanw, vwscanw,
vw_scanw - convert formatted input from a curses window
int scanw(const char *fmt, ...);
int wscanw(WINDOW *win, const char *fmt,
int mvscanw(int y, int x, const char
int mvwscanw(WINDOW *win, int y, int
x , const char *fmt, ...);
int vw_scanw(WINDOW *win, const char
*fmt , va_list varglist);
/* obsolete */
The scanw, wscanw and mvscanw routines are analogous to
scanf [see scanf(3)]. The effect of these routines is as though
wgetstr were called on the window, and the resulting line used as input
for sscanf(3). Fields which do not map to a variable in the fmt
field are lost.
int vwscanw(WINDOW *win, const char *fmt,
The vwscanw and vw_scanw routines are analogous to
vscanf(3). They perform a wscanw using a variable argument
list. The third argument is a va_list, a pointer to a list of
arguments, as defined in <stdarg.h>.
vwscanw returns ERR on failure and an integer equal to the number
of fields scanned on success.
Applications may use the return value from the scanw,
wscanw, mvscanw and mvwscanw routines to determine the
number of fields which were mapped in the call.
Functions with a “mv” prefix first perform a cursor
movement using wmove, and return an error if the position is outside
the window, or if the window pointer is null.
While scanw was implemented in 4BSD, none of the BSD releases used it
until 4.4BSD (in a game). That early version of curses was before the ANSI C
standard. It did not use <varargs.h>, though that was available. In 1991
(a couple of years after SVr4 was generally available, and after the C
standard was published), other developers updated the library, using
<stdarg.h> internally in 4.4BSD curses. Even with this improvement, BSD
curses did not use function prototypes (or even declare functions) in the
<curses.h> header until 1992.
SVr2 documented scanw, wscanw tersely as
“scanf through stdscr” and tersely as “scanf
through win”, respectively.
SVr3 added mvscanw, and mvwscanw, with a three-line
summary saying that they were analogous to scanf(3), explaining that
the string which would be output from scanf(3) would instead be
output using waddstr on the given window. SVr3 also added
vwscanw, saying that the third parameter is a va_list, defined
in <varargs.h>, and referring the reader to the manual pages for
varargs and vprintf for detailed descriptions. (Because the
SVr3 documentation does not mention vscanf, that reference to
vprintf may not be an error).
SVr4 added no new variations of scanw, but provided for
using <varargs.h> or <stdarg.h> to define the va_list
X/Open Curses added vw_scanw to replace vwscanw,
stating that its va_list definition requires <stdarg.h>.
In this implementation, vw_scanw and vwscanw are equivalent, to
support legacy applications. However, the latter (vwscanw) is obsolete:
- The XSI Curses standard, Issue 4 described these functions, noting that
the function vwscanw is marked TO BE WITHDRAWN, and is to be
replaced by a function vw_scanw using the <stdarg.h>
- The Single Unix Specification, Version 2 states that vw_scanw is
preferred to vwscanw since the latter requires including
<varargs.h>, which cannot be used in the same file as
<stdarg.h>. This implementation uses <stdarg.h>
for both, because that header is included in <curses.h>.
- X/Open Curses, Issue 5 (December 2007) marked vwscanw (along with
vwprintw and the termcap interface) as withdrawn.
Both XSI and The Single Unix Specification, Version 2 state that
these functions return ERR or OK.
- Since the underlying scanf(3) can return the number of items
scanned, and the SVr4 code was documented to use this feature, this is
probably an editing error which was introduced in XSI, rather than being
- This implementation returns the number of items scanned, for compatibility
with SVr4 curses. As of 2018, NetBSD curses also returns the number of
items scanned. Both ncurses and NetBSD curses call vsscanf to scan
the string, which returns EOF on error.
- Portable applications should only test if the return value is ERR,
since the OK value (zero) is likely to be misleading.
- One possible way to get useful results would be to use a "%n"
conversion at the end of the format string to ensure that something was