/* PSPP - a program for statistical analysis.
- Copyright (C) 1997-9, 2000, 2006 Free Software Foundation, Inc.
+ Copyright (C) 1997-9, 2000, 2006, 2009, 2011 Free Software Foundation, Inc.
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
#include <config.h>
-#include "data-out.h"
+#include "data/data-out.h"
#include <ctype.h>
#include <float.h>
#include <stdint.h>
#include <stdlib.h>
#include <time.h>
-
-#include "calendar.h"
-#include "format.h"
-#include "settings.h"
-#include "value.h"
-
-#include <libpspp/assertion.h>
-#include <libpspp/float-format.h>
-#include <libpspp/integer-format.h>
-#include <libpspp/message.h>
-#include <libpspp/misc.h>
-#include <libpspp/misc.h>
-#include <libpspp/str.h>
-
-#include "minmax.h"
+#include <unistr.h>
+
+#include "data/calendar.h"
+#include "data/format.h"
+#include "data/settings.h"
+#include "data/value.h"
+#include "libpspp/assertion.h"
+#include "libpspp/cast.h"
+#include "libpspp/float-format.h"
+#include "libpspp/i18n.h"
+#include "libpspp/integer-format.h"
+#include "libpspp/message.h"
+#include "libpspp/misc.h"
+#include "libpspp/pool.h"
+#include "libpspp/str.h"
+
+#include "gl/minmax.h"
#include "gettext.h"
#define _(msgid) gettext (msgid)
static void rounder_format (const struct rounder *, int decimals,
char *output);
\f
-/* Format of integers in output (SET WIB). */
-static enum integer_format output_integer_format = INTEGER_NATIVE;
-
-/* Format of reals in output (SET WRB). */
-static enum float_format output_float_format = FLOAT_NATIVE_DOUBLE;
-
typedef void data_out_converter_func (const union value *,
const struct fmt_spec *,
char *);
char *);
static void output_hex (const void *, size_t bytes, char *);
\f
-/* Converts the INPUT value into printable form in the exactly
- FORMAT->W characters in OUTPUT according to format
- specification FORMAT. No null terminator is appended to the
- buffer. */
-void
-data_out (const union value *input, const struct fmt_spec *format,
- char *output)
-{
- static data_out_converter_func *const converters[FMT_NUMBER_OF_FORMATS] =
+
+static data_out_converter_func *const converters[FMT_NUMBER_OF_FORMATS] =
{
#define FMT(NAME, METHOD, IMIN, OMIN, IO, CATEGORY) output_##METHOD,
#include "format.def"
};
- assert (fmt_check_output (format));
+/* Converts the INPUT value, encoded in INPUT_ENCODING, according to format
+ specification FORMAT, appending the output to OUTPUT in OUTPUT_ENCODING.
+ However, binary formats (FMT_P, FMT_PK, FMT_IB, FMT_PIB, FMT_RB) yield the
+ binary results, which may not be properly encoded for OUTPUT_ENCODING.
- converters[format->type] (input, format, output);
-}
+ VALUE must be the correct width for FORMAT, that is, its width must be
+ fmt_var_width(FORMAT).
-/* Returns the current output integer format. */
-enum integer_format
-data_out_get_integer_format (void)
+ INPUT_ENCODING can normally be obtained by calling dict_get_encoding() on
+ the dictionary with which INPUT is associated. ENCODING is only important
+ when FORMAT's type is FMT_A. */
+void
+data_out_recode (const union value *input, const char *input_encoding,
+ const struct fmt_spec *format,
+ struct string *output, const char *output_encoding)
{
- return output_integer_format;
+ assert (fmt_check_output (format));
+ if (format->type == FMT_A)
+ {
+ char *in = CHAR_CAST (char *, value_str (input, format->w));
+ char *out = recode_string (output_encoding, input_encoding,
+ in, format->w);
+ ds_put_cstr (output, out);
+ free (out);
+ }
+ else if (fmt_get_category (format->type) == FMT_CAT_BINARY)
+ converters[format->type] (input, format,
+ ds_put_uninit (output, format->w));
+ else
+ {
+ char *utf8_encoded = data_out (input, input_encoding, format);
+ char *output_encoded = recode_string (output_encoding, UTF8,
+ utf8_encoded, -1);
+ ds_put_cstr (output, output_encoded);
+ free (output_encoded);
+ free (utf8_encoded);
+ }
}
-/* Sets the output integer format to INTEGER_FORMAT. */
-void
-data_out_set_integer_format (enum integer_format integer_format)
+static char *
+binary_to_utf8 (const char *in, struct pool *pool)
{
- output_integer_format = integer_format;
+ uint8_t *out = pool_alloc_unaligned (pool, strlen (in) * 2 + 1);
+ uint8_t *p = out;
+
+ while (*in != '\0')
+ {
+ uint8_t byte = *in++;
+ int mblen = u8_uctomb (p, byte, 2);
+ assert (mblen > 0);
+ p += mblen;
+ }
+ *p = '\0';
+
+ return CHAR_CAST (char *, out);
}
-/* Returns the current output float format. */
-enum float_format
-data_out_get_float_format (void)
+/* Converts the INPUT value into a UTF-8 encoded string, according to format
+ specification FORMAT.
+
+ VALUE must be the correct width for FORMAT, that is, its width must be
+ fmt_var_width(FORMAT).
+
+ ENCODING must be the encoding of INPUT. Normally this can be obtained by
+ calling dict_get_encoding() on the dictionary with which INPUT is
+ associated. ENCODING is only important when FORMAT's type is FMT_A.
+
+ The return value is dynamically allocated, and must be freed by the caller.
+ If POOL is non-null, then the return value is allocated on that pool. */
+char *
+data_out_pool (const union value *input, const char *encoding,
+ const struct fmt_spec *format, struct pool *pool)
{
- return output_float_format;
+ assert (fmt_check_output (format));
+ if (format->type == FMT_A)
+ {
+ char *in = CHAR_CAST (char *, value_str (input, format->w));
+ return recode_string_pool (UTF8, encoding, in, format->w, pool);
+ }
+ else if (fmt_get_category (format->type) == FMT_CAT_BINARY)
+ {
+ char tmp[16];
+
+ assert (format->w + 1 <= sizeof tmp);
+ converters[format->type] (input, format, tmp);
+ return binary_to_utf8 (tmp, pool);
+ }
+ else
+ {
+ const struct fmt_number_style *style = settings_get_style (format->type);
+ size_t size = format->w + style->extra_bytes + 1;
+ char *output;
+
+ output = pool_alloc_unaligned (pool, size);
+ converters[format->type] (input, format, output);
+ return output;
+ }
}
-/* Sets the output float format to FLOAT_FORMAT. */
-void
-data_out_set_float_format (enum float_format float_format)
+char *
+data_out (const union value *input, const char *encoding, const struct fmt_spec *format)
{
- output_float_format = float_format;
+ return data_out_pool (input, encoding, format, NULL);
}
+
\f
/* Main conversion functions. */
else
output_overflow (format, output);
}
+
+ output[format->w] = '\0';
}
/* Outputs Z format. */
char buf[128];
if (input->f == SYSMIS)
output_missing (format, output);
- else if (fabs (number) >= power10 (format->w)
- || sprintf (buf, "%0*.0f", format->w,
- fabs (round (number))) != format->w)
- output_overflow (format, output);
- else
+ else if (fabs (number) < power10 (format->w)
+ && sprintf (buf, "%0*.0f", format->w,
+ fabs (round (number))) == format->w)
{
if (number < 0 && strspn (buf, "0") < format->w)
{
*p = "}JKLMNOPQR"[*p - '0'];
}
memcpy (output, buf, format->w);
+ output[format->w] = '\0';
}
+ else
+ output_overflow (format, output);
}
/* Outputs P format. */
uint64_t integer = fabs (number);
if (number < 0)
integer = -integer;
- output_binary_integer (integer, format->w, output_integer_format,
+ output_binary_integer (integer, format->w,
+ settings_get_output_integer_format (),
output);
}
+
+ output[format->w] = '\0';
}
/* Outputs PIB format. */
|| number < 0 || number >= power256 (format->w))
memset (output, 0, format->w);
else
- output_binary_integer (number, format->w, output_integer_format, output);
+ output_binary_integer (number, format->w,
+ settings_get_output_integer_format (), output);
+
+ output[format->w] = '\0';
}
/* Outputs PIBHEX format. */
output_binary_integer (number, format->w / 2, INTEGER_MSB_FIRST, tmp);
output_hex (tmp, format->w / 2, output);
}
+
}
/* Outputs RB format. */
{
double d = input->f;
memcpy (output, &d, format->w);
+
+ output[format->w] = '\0';
}
/* Outputs RBHEX format. */
char *output)
{
double d = input->f;
+
output_hex (&d, format->w / 2, output);
}
p += sprintf (p, "%02d", month);
else
{
- static const char *months[12] =
+ static const char *const months[12] =
{
"JAN", "FEB", "MAR", "APR", "MAY", "JUN",
"JUL", "AUG", "SEP", "OCT", "NOV", "DEC",
}
else
{
- int offset = year - get_epoch ();
+ int epoch = settings_get_epoch ();
+ int offset = year - epoch;
if (offset < 0 || offset > 99)
goto overflow;
p += sprintf (p, "%02d", abs (year) % 100);
int d = MIN (format->d, excess_width - 4);
int w = d + 3;
sprintf (p, ":%0*.*f", w, d, number);
- if (fmt_decimal_char (FMT_F) != '.')
+ if (settings_get_decimal_char (FMT_F) != '.')
{
char *cp = strchr (p, '.');
if (cp != NULL)
- *cp = fmt_decimal_char (FMT_F);
+ *cp = settings_get_decimal_char (FMT_F);
}
p += strlen (p);
}
}
}
- buf_copy_lpad (output, format->w, tmp, p - tmp);
+ buf_copy_lpad (output, format->w, tmp, p - tmp, ' ');
+ output[format->w] = '\0';
return;
overflow:
output_WKDAY (const union value *input, const struct fmt_spec *format,
char *output)
{
- static const char *weekdays[7] =
+ static const char *const weekdays[7] =
{
"SUNDAY", "MONDAY", "TUESDAY", "WEDNESDAY",
"THURSDAY", "FRIDAY", "SATURDAY",
};
if (input->f >= 1 && input->f < 8)
- buf_copy_str_rpad (output, format->w, weekdays[(int) input->f - 1]);
+ {
+ buf_copy_str_rpad (output, format->w,
+ weekdays[(int) input->f - 1], ' ');
+ output[format->w] = '\0';
+ }
else
{
if (input->f != SYSMIS)
msg (ME, _("Weekday number %f is not between 1 and 7."), input->f);
output_missing (format, output);
}
+
}
/* Outputs MONTH format. */
output_MONTH (const union value *input, const struct fmt_spec *format,
char *output)
{
- static const char *months[12] =
+ static const char *const months[12] =
{
"JANUARY", "FEBRUARY", "MARCH", "APRIL", "MAY", "JUNE",
"JULY", "AUGUST", "SEPTEMBER", "OCTOBER", "NOVEMBER", "DECEMBER",
};
if (input->f >= 1 && input->f < 13)
- buf_copy_str_rpad (output, format->w, months[(int) input->f - 1]);
+ {
+ buf_copy_str_rpad (output, format->w, months[(int) input->f - 1], ' ');
+ output[format->w] = '\0';
+ }
else
{
if (input->f != SYSMIS)
msg (ME, _("Month number %f is not between 1 and 12."), input->f);
output_missing (format, output);
}
+
}
/* Outputs A format. */
static void
-output_A (const union value *input, const struct fmt_spec *format,
- char *output)
+output_A (const union value *input UNUSED,
+ const struct fmt_spec *format UNUSED, char *output UNUSED)
{
- memcpy (output, input->s, format->w);
+ NOT_REACHED ();
}
/* Outputs AHEX format. */
output_AHEX (const union value *input, const struct fmt_spec *format,
char *output)
{
- output_hex (input->s, format->w, output);
+ output_hex (value_str (input, format->w), format->w / 2, output);
}
\f
/* Decimal and scientific formatting. */
output_decimal (const struct rounder *r, const struct fmt_spec *format,
bool require_affixes, char *output)
{
- const struct fmt_number_style *style = fmt_get_style (format->type);
+ const struct fmt_number_style *style =
+ settings_get_style (format->type);
+
int decimals;
for (decimals = format->d; decimals >= 0; decimals--)
the negative suffix, plus (if negative) the negative
prefix. */
width = rounder_width (r, decimals, &integer_digits, &add_neg_prefix);
- width += ss_length (style->neg_suffix);
+ width += style->neg_suffix.width;
if (add_neg_prefix)
- width += ss_length (style->neg_prefix);
+ width += style->neg_prefix.width;
if (width > format->w)
continue;
if (format->w > width)
p = mempset (p, ' ', format->w - width);
if (add_neg_prefix)
- p = mempcpy (p, ss_data (style->neg_prefix),
- ss_length (style->neg_prefix));
+ p = stpcpy (p, style->neg_prefix.s);
if (add_affixes)
- p = mempcpy (p, ss_data (style->prefix), ss_length (style->prefix));
+ p = stpcpy (p, style->prefix.s);
if (!add_grouping)
p = mempcpy (p, magnitude, integer_digits);
else
p = mempcpy (p, &magnitude[integer_digits + 1], decimals);
}
if (add_affixes)
- p = mempcpy (p, ss_data (style->suffix), ss_length (style->suffix));
+ p = stpcpy (p, style->suffix.s);
if (add_neg_prefix)
- p = mempcpy (p, ss_data (style->neg_suffix),
- ss_length (style->neg_suffix));
+ p = stpcpy (p, style->neg_suffix.s);
else
- p = mempset (p, ' ', ss_length (style->neg_suffix));
- assert (p == output + format->w);
+ p = mempset (p, ' ', style->neg_suffix.width);
+
+ assert (p >= output + format->w);
+ assert (p <= output + format->w + style->extra_bytes);
+ *p = '\0';
return true;
}
output_scientific (double number, const struct fmt_spec *format,
bool require_affixes, char *output)
{
- const struct fmt_number_style *style = fmt_get_style (format->type);
+ const struct fmt_number_style *style =
+ settings_get_style (format->type);
int width;
int fraction_width;
bool add_affixes;
- char buf[64], *p;
+ char *p;
/* Allocate minimum required space. */
- width = 6 + ss_length (style->neg_suffix);
+ width = 6 + style->neg_suffix.width;
if (number < 0)
- width += ss_length (style->neg_prefix);
+ width += style->neg_prefix.width;
if (width > format->w)
return false;
width += fraction_width;
/* Format (except suffix). */
- p = buf;
+ p = output;
if (width < format->w)
p = mempset (p, ' ', format->w - width);
if (number < 0)
- p = mempcpy (p, ss_data (style->neg_prefix),
- ss_length (style->neg_prefix));
+ p = stpcpy (p, style->neg_prefix.s);
if (add_affixes)
- p = mempcpy (p, ss_data (style->prefix), ss_length (style->prefix));
+ p = stpcpy (p, style->prefix.s);
if (fraction_width > 0)
sprintf (p, "%#.*E", fraction_width - 1, fabs (number));
else
/* Add suffixes. */
p = strchr (p, '\0');
if (add_affixes)
- p = mempcpy (p, ss_data (style->suffix), ss_length (style->suffix));
+ p = stpcpy (p, style->suffix.s);
if (number < 0)
- p = mempcpy (p, ss_data (style->neg_suffix),
- ss_length (style->neg_suffix));
+ p = stpcpy (p, style->neg_suffix.s);
else
- p = mempset (p, ' ', ss_length (style->neg_suffix));
+ p = mempset (p, ' ', style->neg_suffix.width);
- assert (p == buf + format->w);
- memcpy (output, buf, format->w);
+ assert (p >= output + format->w);
+ assert (p <= output + format->w + style->extra_bytes);
+ *p = '\0';
return true;
}
else
s = "Unknown";
- buf_copy_str_lpad (output, format->w, s);
+ buf_copy_str_lpad (output, format->w, s, ' ');
}
else
output_overflow (format, output);
+
+ output[format->w] = '\0';
}
/* Formats OUTPUT as a missing value for the given FORMAT. */
}
else
output[format->w - 1] = '.';
+
+ output[format->w] = '\0';
}
/* Formats OUTPUT for overflow given FORMAT. */
output_overflow (const struct fmt_spec *format, char *output)
{
memset (output, '*', format->w);
+ output[format->w] = '\0';
}
/* Converts the integer part of NUMBER to a packed BCD number
char decimal[64];
assert (digits < sizeof decimal);
+
+ output[DIV_RND_UP (digits, 2)] = '\0';
if (number != SYSMIS
&& number >= 0.
&& number < power10 (digits)
*output++ = hex_digits[data[i] >> 4];
*output++ = hex_digits[data[i] & 15];
}
+ *output = '\0';
}
projects / pspp-builds.git / blobdiff