X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f
X-Recipient: djgpp-workers AT delorie DOT com
Date: Sat, 22 Dec 2012 10:27:30 +0200
From: Eli Zaretskii <eliz AT gnu DOT org>
Subject: Re: Implementation of %n$ conversion specifier for scanf family of functions.
In-reply-to: <50D4FBCF.2060706@gmx.de>
X-012-Sender: halo1 AT inter DOT net DOT il
To: djgpp-workers AT delorie DOT com
Message-id: <8338yyqyod.fsf@gnu.org>
References: <50D4FBCF DOT 2060706 AT gmx DOT de>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> Date: Sat, 22 Dec 2012 01:16:15 +0100
> From: Juan Manuel Guerrero <juan DOT guerrero AT gmx DOT de>
> 
> The patch below will implement the C99 numeric conversion specifier aka %n$ for
> the scanf family of functions.

Thanks.

> +@findex _doscan AT r{, and @acronym{C99} numeric conversion specifiers}
> +@findex scanf AT r{, and @acronym{C99} numeric conversion specifiers}
> +The @code{%n$} numeric conversion specifiers are now supported
> +by @code{_doscan} and the @code{scanf} family of functions.

The use of "n" in "%n$" may mislead the reader to think it's a literal
"n" meant here.  I suggest to use @var{n} instead, or even something
like @var{param-no}.

> +@samp{[} also skip leading whitespace automatically.
> +Conversions can be applied to the nth argument after the format in the
> +argument list, rather than to the next unused argument.  In this case,
> +the conversion specifier character @code{%} is replaced by the sequence
> +@code{%n$}, where @code{n} is a decimal integer in the range [@code{1}, @code{number of the last argument}].

          ^          ^^^^^^^^
@var{n} and @code{@var{n}}, respectively.  And again, maybe don't use
"n" at all, but something more descriptive.

> +the format string more than once.  The format can contain either form of
> +a conversion specification - that is, @code{%} or @code{%n$} - but the two

Don't use a single "-" character in Texinfo -- that typesets as a
minus sign, which is not what you want.  For em-dash, use "---", 3
dashes in a row, which will produce what you want both in Info and in
the printed manual.

> +forms cannot be mixed within a single format string.  The only exception
> +to this is that @code{%%} or @code{%*} can be mixed with the @code{%n$}
> +form.  When numbered argument specifications are used, specifying the
> +Nth argument requires that all the leading arguments, from the first
> +to the (N-1)th, are pointers.

Here, again, use @var{n} in all instances, including in "(N-1)th".  Do
NOT up-case "n", since @var will do that for you automatically.

> +/* Test the n$ numeric conversion specifier.  */
> +
> +#include <stdio.h>
> +
> +int main(void)
> +{
> +  const char *buffer[3] = {
> +    "Fecha y hora es: 21 de deciembre de 2012, 01:02:03",
> +    "Datum und Uhrzeit ist: 21. Dezember 2012, 01:02:03",
> +    "Date and hour is: december 21 2012, 01:02:03"
> +  };
> +  const char *format[3] = {
> +    "%*[A-Za-z :] %3$d %*[a-z] %2$9c %*[a-z] %1$d %*[,] %4$d%*[:] %5$d%*[:] %6$d%*",
> +    "%*[A-Za-z :] %3$d %*[.] %2$8c %1$d %*[,] %4$d%*[:] %5$d%*[:] %6$d%*",
> +    "%*[A-Za-z ]%*[:] %2$8c %3$d %1$d %*[,] %4$d%*[:] %5$d%*[:] %6$d%*"
> +  };
> +  const char *language[3] = {
> +    "spanish",
> +    "german",
> +    "english"
> +  };
> +  char month[10];
> +  int year, day, hour, min, sec, i;
> +
> +
> +  for (i = 0; i < (sizeof buffer / sizeof buffer[0]); i++)
> +  {
> +    sscanf(buffer[i], format[i], &year, month, &day, &hour, &min, &sec);
> +    if (i == 0)
> +      month[9] = '\0';
> +    else
> +      month[8] = '\0';
> +    printf("Result of scanf using %s format\n\"%s\":\n"
> +           "  year:  %d\n"
> +           "  month: %s\n"
> +           "  day:   %d\n"
> +           "  hour:  %d\n"
> +           "  min:   %d\n"
> +           "  sec:   %d\n", language[i], format[i], year, month, day, hour, min, sec);
> +  }
> +
> +  return 0;
> +}

It would be better to include the expected results in the program,
instead of relying on a human eye to detect errors.

Thanks again for working on this.