summaryrefslogtreecommitdiffstats
path: root/contrib/libio/outfloat.c
diff options
context:
space:
mode:
Diffstat (limited to 'contrib/libio/outfloat.c')
-rw-r--r--contrib/libio/outfloat.c211
1 files changed, 211 insertions, 0 deletions
diff --git a/contrib/libio/outfloat.c b/contrib/libio/outfloat.c
new file mode 100644
index 0000000..5cbb6c9
--- /dev/null
+++ b/contrib/libio/outfloat.c
@@ -0,0 +1,211 @@
+/*
+Copyright (C) 1993 Free Software Foundation
+
+This file is part of the GNU IO Library. This library is free
+software; you can redistribute it and/or modify it under the
+terms of the GNU General Public License as published by the
+Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+This library is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this library; see the file COPYING. If not, write to the Free
+Software Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+As a special exception, if you link this library with files
+compiled with a GNU compiler to produce an executable, this does not cause
+the resulting executable to be covered by the GNU General Public License.
+This exception does not however invalidate any other reasons why
+the executable file might be covered by the GNU General Public License. */
+
+#include "libioP.h"
+
+#ifdef _IO_USE_DTOA
+/* Format floating-point number and print them.
+ Return number of chars printed, or EOF on error.
+
+ sign_mode == '+' : print "-" or "+"
+ sign_mode == ' ' : print "-" or " "
+ sign_mode == '\0' : print "-' or ""
+*/
+
+int
+_IO_outfloat (value, sb, type, width, precision, flags, sign_mode, fill)
+ double value;
+ _IO_FILE *sb;
+ int type;
+ int width;
+ int precision;
+ int flags;
+ int sign_mode;
+ int fill;
+{
+ int count = 0;
+#define PUT(x) do {if (_IO_putc(x, sb) < 0) goto error; count++;} while (0)
+#define PUTN(p, n) \
+ do {int _n=n; count+=_n; if (_IO_sputn(sb, p,_n) != _n) goto error;} while(0)
+#define PADN(fill, n) \
+ do {int _n = n; count+=_n; if (_IO_padn(sb, fill, _n) != _n) goto error;} while (0)
+ int pad_kind = flags & (_IO_LEFT|_IO_RIGHT|_IO_INTERNAL);
+ int skip_zeroes = 0;
+ int show_dot = (flags & _IO_SHOWPOINT) != 0;
+ int decpt;
+ int sign;
+ int mode;
+ int exponent_size;
+ int print_sign;
+ int trailing_zeroes, useful_digits;
+ int padding, unpadded_width;
+ char *p;
+ char *exponent_start;
+ register int i;
+#define EBUF_SIZE 12
+#define EBUF_END &ebuf[EBUF_SIZE]
+ char ebuf[EBUF_SIZE];
+ char *end;
+ int exp = 0;
+ switch (type)
+ {
+ case 'f':
+ mode = 3;
+ break;
+ case 'e':
+ case 'E':
+ exp = type;
+ mode = 2;
+ if (precision != 999)
+ precision++; /* Add one to include digit before decimal point. */
+ break;
+ case 'g':
+ case 'G':
+ exp = type == 'g' ? 'e' : 'E';
+ if (precision == 0) precision = 1;
+ if (!(flags & _IO_SHOWPOINT))
+ skip_zeroes = 1;
+ type = 'g';
+ mode = 2;
+ break;
+ }
+ /* Do the actual convension */
+ if (precision == 999 && mode != 3)
+ mode = 0;
+ p = _IO_dtoa(value, mode, precision, &decpt, &sign, &end);
+ useful_digits = end-p;
+ exponent_start = EBUF_END;
+ if (mode == 0)
+ precision = useful_digits;
+ /* Check if we need to emit an exponent. */
+ if (mode != 3 && decpt != 9999)
+ {
+ i = decpt - 1;
+ if ((type != 'g' && type != 'F') || i < -4 || i >= precision)
+ {
+ /* Print the exponent into ebuf.
+ We write ebuf in reverse order (right-to-left). */
+ char sign;
+ if (i >= 0)
+ sign = '+';
+ else
+ sign = '-', i = -i;
+ /* Note: ANSI requires at least 2 exponent digits. */
+ do {
+ *--exponent_start = (i % 10) + '0';
+ i /= 10;
+ } while (i >= 10);
+ *--exponent_start = i + '0';
+ *--exponent_start = sign;
+ *--exponent_start = exp;
+ }
+ }
+ exponent_size = EBUF_END - exponent_start;
+ if (mode == 1)
+ precision = 1;
+ /* If we print an exponent, always show just one digit before point. */
+ if (exponent_size)
+ decpt = 1;
+ if (decpt == 9999)
+ { /* Infinity or NaN */
+ decpt = useful_digits;
+ precision = 0;
+ show_dot = 0;
+ }
+
+ /* dtoa truncates trailing zeroes. Set the variable trailing_zeroes to
+ the number of 0's we have to add (after the decimal point). */
+ if (skip_zeroes)
+ trailing_zeroes = 0;
+ else if (type == 'f')
+ trailing_zeroes = useful_digits <= decpt ? precision
+ : precision-(useful_digits-decpt);
+ else if (exponent_size) /* 'e' 'E' or 'g' format using exponential notation*/
+ trailing_zeroes = precision - useful_digits;
+ else /* 'g' format not using exponential notation. */
+ trailing_zeroes = useful_digits <= decpt ? precision - decpt
+ : precision-useful_digits;
+ if (trailing_zeroes < 0) trailing_zeroes = 0;
+
+ if (trailing_zeroes != 0 || useful_digits > decpt)
+ show_dot = 1;
+ if (sign_mode == 0)
+ print_sign = sign ? '-' : 0;
+ else if (sign_mode == '+')
+ print_sign = sign ? '-' : '+';
+ else /* if (sign_mode == ' ') */
+ print_sign = sign ? '-' : ' ';
+
+ /* Calculate the width (before padding). */
+ unpadded_width =
+ (print_sign != 0) + trailing_zeroes + exponent_size + show_dot
+ + useful_digits
+ + (decpt > useful_digits ? decpt - useful_digits
+ : decpt > 0 ? 0 : 1 - decpt);
+
+ padding = width > unpadded_width ? width - unpadded_width : 0;
+ if (padding > 0 && pad_kind != _IO_LEFT && pad_kind != _IO_INTERNAL)
+ PADN(fill, padding); /* Default (right) adjust */
+ if (print_sign)
+ PUT(print_sign);
+ if (pad_kind == _IO_INTERNAL && padding > 0)
+ PADN(fill, padding);
+ if (decpt > 0)
+ {
+ if (useful_digits >= decpt)
+ PUTN(p, decpt);
+ else
+ {
+ PUTN(p, useful_digits);
+ PADN('0', decpt-useful_digits);
+ }
+ if (show_dot)
+ {
+ PUT('.');
+ /* Print digits after the decimal point. */
+ if (useful_digits > decpt)
+ PUTN(p + decpt, useful_digits-decpt);
+ }
+ }
+ else
+ {
+ PUT('0');
+ if (show_dot)
+ {
+ PUT('.');
+ PADN('0', -decpt);
+ /* Print digits after the decimal point. */
+ PUTN(p, useful_digits);
+ }
+ }
+ PADN('0', trailing_zeroes);
+ if (exponent_size)
+ PUTN(exponent_start, exponent_size);
+ if (pad_kind == _IO_LEFT && padding > 0) /* Left adjustment*/
+ PADN(fill, padding);
+ return count;
+ error:
+ return EOF;
+}
+#endif
OpenPOWER on IntegriCloud