mirror of
https://github.com/python/cpython.git
synced 2025-07-28 21:55:21 +00:00
Backport of some of the work in r71665 to trunk. This reworks much of
int, long, and float __format__(), and it keeps their implementation in sync with py3k. Also added PyOS_double_to_string. This is the "fallback" version that's also available in trunk, and should be kept in sync with that code. I'll add an issue to document PyOS_double_to_string in the C API. There are many internal cleanups. Externally visible changes include: - Implement PEP 378, Format Specifier for Thousands Separator, for floats, ints, and longs. - Issue #5515: 'n' formatting for ints, longs, and floats handles leading zero formatting poorly. - Issue #5772: For float.__format__, don't add a trailing ".0" if we're using no type code and we have an exponent.
This commit is contained in:
parent
cbb5308723
commit
aca19e6a74
9 changed files with 894 additions and 496 deletions
|
@ -10,6 +10,25 @@ PyAPI_FUNC(double) PyOS_ascii_strtod(const char *str, char **ptr);
|
||||||
PyAPI_FUNC(double) PyOS_ascii_atof(const char *str);
|
PyAPI_FUNC(double) PyOS_ascii_atof(const char *str);
|
||||||
PyAPI_FUNC(char *) PyOS_ascii_formatd(char *buffer, size_t buf_len, const char *format, double d);
|
PyAPI_FUNC(char *) PyOS_ascii_formatd(char *buffer, size_t buf_len, const char *format, double d);
|
||||||
|
|
||||||
|
/* The caller is responsible for calling PyMem_Free to free the buffer
|
||||||
|
that's is returned. */
|
||||||
|
PyAPI_FUNC(char *) PyOS_double_to_string(double val,
|
||||||
|
char format_code,
|
||||||
|
int precision,
|
||||||
|
int flags,
|
||||||
|
int *type);
|
||||||
|
|
||||||
|
|
||||||
|
/* PyOS_double_to_string's "flags" parameter can be set to 0 or more of: */
|
||||||
|
#define Py_DTSF_SIGN 0x01 /* always add the sign */
|
||||||
|
#define Py_DTSF_ADD_DOT_0 0x02 /* if the result is an integer add ".0" */
|
||||||
|
#define Py_DTSF_ALT 0x04 /* "alternate" formatting. it's format_code
|
||||||
|
specific */
|
||||||
|
|
||||||
|
/* PyOS_double_to_string's "type", if non-NULL, will be set to one of: */
|
||||||
|
#define Py_DTST_FINITE 0
|
||||||
|
#define Py_DTST_INFINITE 1
|
||||||
|
#define Py_DTST_NAN 2
|
||||||
|
|
||||||
#ifdef __cplusplus
|
#ifdef __cplusplus
|
||||||
}
|
}
|
||||||
|
|
|
@ -177,16 +177,26 @@ PyAPI_FUNC(int) PyString_AsStringAndSize(
|
||||||
strings) */
|
strings) */
|
||||||
);
|
);
|
||||||
|
|
||||||
|
|
||||||
/* Using the current locale, insert the thousands grouping
|
/* Using the current locale, insert the thousands grouping
|
||||||
into the string pointed to by buffer. For the argument descriptions,
|
into the string pointed to by buffer. For the argument descriptions,
|
||||||
see Objects/stringlib/localeutil.h */
|
see Objects/stringlib/localeutil.h */
|
||||||
|
PyAPI_FUNC(Py_ssize_t) _PyString_InsertThousandsGroupingLocale(char *buffer,
|
||||||
|
Py_ssize_t n_buffer,
|
||||||
|
char *digits,
|
||||||
|
Py_ssize_t n_digits,
|
||||||
|
Py_ssize_t min_width);
|
||||||
|
|
||||||
PyAPI_FUNC(int) _PyString_InsertThousandsGrouping(char *buffer,
|
/* Using explicit passed-in values, insert the thousands grouping
|
||||||
Py_ssize_t n_buffer,
|
into the string pointed to by buffer. For the argument descriptions,
|
||||||
Py_ssize_t n_digits,
|
see Objects/stringlib/localeutil.h */
|
||||||
Py_ssize_t buf_size,
|
PyAPI_FUNC(Py_ssize_t) _PyString_InsertThousandsGrouping(char *buffer,
|
||||||
Py_ssize_t *count,
|
Py_ssize_t n_buffer,
|
||||||
int append_zero_char);
|
char *digits,
|
||||||
|
Py_ssize_t n_digits,
|
||||||
|
Py_ssize_t min_width,
|
||||||
|
const char *grouping,
|
||||||
|
const char *thousands_sep);
|
||||||
|
|
||||||
/* Format the object based on the format_spec, as defined in PEP 3101
|
/* Format the object based on the format_spec, as defined in PEP 3101
|
||||||
(Advanced String Formatting). */
|
(Advanced String Formatting). */
|
||||||
|
|
|
@ -232,6 +232,10 @@ class FormatTest(unittest.TestCase):
|
||||||
testboth("%o", -042L, "-42")
|
testboth("%o", -042L, "-42")
|
||||||
testboth("%o", float(042), "42")
|
testboth("%o", float(042), "42")
|
||||||
|
|
||||||
|
# alternate float formatting
|
||||||
|
testformat('%g', 1.1, '1.1')
|
||||||
|
testformat('%#g', 1.1, '1.10000')
|
||||||
|
|
||||||
# Test exception for unknown format characters
|
# Test exception for unknown format characters
|
||||||
if verbose:
|
if verbose:
|
||||||
print 'Testing exceptions'
|
print 'Testing exceptions'
|
||||||
|
|
|
@ -113,6 +113,9 @@ class TypesTests(unittest.TestCase):
|
||||||
self.assertEqual(1.5e-101.__format__('e'), '1.500000e-101')
|
self.assertEqual(1.5e-101.__format__('e'), '1.500000e-101')
|
||||||
self.assertEqual('%e' % 1.5e-101, '1.500000e-101')
|
self.assertEqual('%e' % 1.5e-101, '1.500000e-101')
|
||||||
|
|
||||||
|
self.assertEqual('%g' % 1.0, '1')
|
||||||
|
self.assertEqual('%#g' % 1.0, '1.00000')
|
||||||
|
|
||||||
def test_normal_integers(self):
|
def test_normal_integers(self):
|
||||||
# Ensure the first 256 integers are shared
|
# Ensure the first 256 integers are shared
|
||||||
a = 256
|
a = 256
|
||||||
|
@ -412,6 +415,9 @@ class TypesTests(unittest.TestCase):
|
||||||
self.assertRaises(TypeError, 3 .__format__, None)
|
self.assertRaises(TypeError, 3 .__format__, None)
|
||||||
self.assertRaises(TypeError, 3 .__format__, 0)
|
self.assertRaises(TypeError, 3 .__format__, 0)
|
||||||
|
|
||||||
|
# can't have ',' with 'c'
|
||||||
|
self.assertRaises(ValueError, 3 .__format__, ",c")
|
||||||
|
|
||||||
# ensure that only int and float type specifiers work
|
# ensure that only int and float type specifiers work
|
||||||
for format_spec in ([chr(x) for x in range(ord('a'), ord('z')+1)] +
|
for format_spec in ([chr(x) for x in range(ord('a'), ord('z')+1)] +
|
||||||
[chr(x) for x in range(ord('A'), ord('Z')+1)]):
|
[chr(x) for x in range(ord('A'), ord('Z')+1)]):
|
||||||
|
@ -609,11 +615,37 @@ class TypesTests(unittest.TestCase):
|
||||||
# a totaly empty format specifier means something else.
|
# a totaly empty format specifier means something else.
|
||||||
# So, just use a sign flag
|
# So, just use a sign flag
|
||||||
test(1e200, '+g', '+1e+200')
|
test(1e200, '+g', '+1e+200')
|
||||||
test(1e200, '+', '+1.0e+200')
|
test(1e200, '+', '+1e+200')
|
||||||
test(1.1e200, '+g', '+1.1e+200')
|
test(1.1e200, '+g', '+1.1e+200')
|
||||||
test(1.1e200, '+', '+1.1e+200')
|
test(1.1e200, '+', '+1.1e+200')
|
||||||
|
|
||||||
# % formatting
|
test(1.1e200, '+g', '+1.1e+200')
|
||||||
|
test(1.1e200, '+', '+1.1e+200')
|
||||||
|
|
||||||
|
# 0 padding
|
||||||
|
test(1234., '010f', '1234.000000')
|
||||||
|
test(1234., '011f', '1234.000000')
|
||||||
|
test(1234., '012f', '01234.000000')
|
||||||
|
test(-1234., '011f', '-1234.000000')
|
||||||
|
test(-1234., '012f', '-1234.000000')
|
||||||
|
test(-1234., '013f', '-01234.000000')
|
||||||
|
test(-1234.12341234, '013f', '-01234.123412')
|
||||||
|
test(-123456.12341234, '011.2f', '-0123456.12')
|
||||||
|
|
||||||
|
# 0 padding with commas
|
||||||
|
test(1234., '011,f', '1,234.000000')
|
||||||
|
test(1234., '012,f', '1,234.000000')
|
||||||
|
test(1234., '013,f', '01,234.000000')
|
||||||
|
test(-1234., '012,f', '-1,234.000000')
|
||||||
|
test(-1234., '013,f', '-1,234.000000')
|
||||||
|
test(-1234., '014,f', '-01,234.000000')
|
||||||
|
test(-12345., '015,f', '-012,345.000000')
|
||||||
|
test(-123456., '016,f', '-0,123,456.000000')
|
||||||
|
test(-123456., '017,f', '-0,123,456.000000')
|
||||||
|
test(-123456.12341234, '017,f', '-0,123,456.123412')
|
||||||
|
test(-123456.12341234, '013,.2f', '-0,123,456.12')
|
||||||
|
|
||||||
|
# % formatting
|
||||||
test(-1.0, '%', '-100.000000%')
|
test(-1.0, '%', '-100.000000%')
|
||||||
|
|
||||||
# format spec must be string
|
# format spec must be string
|
||||||
|
@ -637,6 +669,24 @@ class TypesTests(unittest.TestCase):
|
||||||
self.assertRaises(ValueError, format, 0.0, '#')
|
self.assertRaises(ValueError, format, 0.0, '#')
|
||||||
self.assertRaises(ValueError, format, 0.0, '#20f')
|
self.assertRaises(ValueError, format, 0.0, '#20f')
|
||||||
|
|
||||||
|
def test_format_spec_errors(self):
|
||||||
|
# int, float, and string all share the same format spec
|
||||||
|
# mini-language parser.
|
||||||
|
|
||||||
|
# Check that we can't ask for too many digits. This is
|
||||||
|
# probably a CPython specific test. It tries to put the width
|
||||||
|
# into a C long.
|
||||||
|
self.assertRaises(ValueError, format, 0, '1'*10000 + 'd')
|
||||||
|
|
||||||
|
# Similar with the precision.
|
||||||
|
self.assertRaises(ValueError, format, 0, '.' + '1'*10000 + 'd')
|
||||||
|
|
||||||
|
# And may as well test both.
|
||||||
|
self.assertRaises(ValueError, format, 0, '1'*1000 + '.' + '1'*10000 + 'd')
|
||||||
|
|
||||||
|
# Make sure commas aren't allowed with various type codes
|
||||||
|
for code in 'xXobns':
|
||||||
|
self.assertRaises(ValueError, format, 0, ',' + code)
|
||||||
|
|
||||||
def test_main():
|
def test_main():
|
||||||
run_unittest(TypesTests)
|
run_unittest(TypesTests)
|
||||||
|
|
|
@ -12,6 +12,15 @@ What's New in Python 2.7 alpha 1
|
||||||
Core and Builtins
|
Core and Builtins
|
||||||
-----------------
|
-----------------
|
||||||
|
|
||||||
|
- Implement PEP 378, Format Specifier for Thousands Separator, for
|
||||||
|
floats, ints, and longs.
|
||||||
|
|
||||||
|
- Issue #5515: 'n' formatting for ints, longs, and floats handles
|
||||||
|
leading zero formatting poorly.
|
||||||
|
|
||||||
|
- Issue #5772: For float.__format__, don't add a trailing ".0" if
|
||||||
|
we're using no type code and we have an exponent.
|
||||||
|
|
||||||
- Issue #3166: Make long -> float (and int -> float) conversions
|
- Issue #3166: Make long -> float (and int -> float) conversions
|
||||||
correctly rounded.
|
correctly rounded.
|
||||||
|
|
||||||
|
|
File diff suppressed because it is too large
Load diff
|
@ -5,126 +5,208 @@
|
||||||
|
|
||||||
#include <locale.h>
|
#include <locale.h>
|
||||||
|
|
||||||
|
#define MAX(x, y) ((x) < (y) ? (y) : (x))
|
||||||
|
#define MIN(x, y) ((x) < (y) ? (x) : (y))
|
||||||
|
|
||||||
|
typedef struct {
|
||||||
|
const char *grouping;
|
||||||
|
char previous;
|
||||||
|
Py_ssize_t i; /* Where we're currently pointing in grouping. */
|
||||||
|
} GroupGenerator;
|
||||||
|
|
||||||
|
static void
|
||||||
|
_GroupGenerator_init(GroupGenerator *self, const char *grouping)
|
||||||
|
{
|
||||||
|
self->grouping = grouping;
|
||||||
|
self->i = 0;
|
||||||
|
self->previous = 0;
|
||||||
|
}
|
||||||
|
|
||||||
|
/* Returns the next grouping, or 0 to signify end. */
|
||||||
|
static Py_ssize_t
|
||||||
|
_GroupGenerator_next(GroupGenerator *self)
|
||||||
|
{
|
||||||
|
/* Note that we don't really do much error checking here. If a
|
||||||
|
grouping string contains just CHAR_MAX, for example, then just
|
||||||
|
terminate the generator. That shouldn't happen, but at least we
|
||||||
|
fail gracefully. */
|
||||||
|
switch (self->grouping[self->i]) {
|
||||||
|
case 0:
|
||||||
|
return self->previous;
|
||||||
|
case CHAR_MAX:
|
||||||
|
/* Stop the generator. */
|
||||||
|
return 0;
|
||||||
|
default: {
|
||||||
|
char ch = self->grouping[self->i];
|
||||||
|
self->previous = ch;
|
||||||
|
self->i++;
|
||||||
|
return (Py_ssize_t)ch;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
/* Fill in some digits, leading zeros, and thousands separator. All
|
||||||
|
are optional, depending on when we're called. */
|
||||||
|
static void
|
||||||
|
fill(STRINGLIB_CHAR **digits_end, STRINGLIB_CHAR **buffer_end,
|
||||||
|
Py_ssize_t n_chars, Py_ssize_t n_zeros, const char* thousands_sep,
|
||||||
|
Py_ssize_t thousands_sep_len)
|
||||||
|
{
|
||||||
|
#if STRINGLIB_IS_UNICODE
|
||||||
|
Py_ssize_t i;
|
||||||
|
#endif
|
||||||
|
|
||||||
|
if (thousands_sep) {
|
||||||
|
*buffer_end -= thousands_sep_len;
|
||||||
|
|
||||||
|
/* Copy the thousands_sep chars into the buffer. */
|
||||||
|
#if STRINGLIB_IS_UNICODE
|
||||||
|
/* Convert from the char's of the thousands_sep from
|
||||||
|
the locale into unicode. */
|
||||||
|
for (i = 0; i < thousands_sep_len; ++i)
|
||||||
|
(*buffer_end)[i] = thousands_sep[i];
|
||||||
|
#else
|
||||||
|
/* No conversion, just memcpy the thousands_sep. */
|
||||||
|
memcpy(*buffer_end, thousands_sep, thousands_sep_len);
|
||||||
|
#endif
|
||||||
|
}
|
||||||
|
|
||||||
|
*buffer_end -= n_chars;
|
||||||
|
*digits_end -= n_chars;
|
||||||
|
memcpy(*buffer_end, *digits_end, n_chars * sizeof(STRINGLIB_CHAR));
|
||||||
|
|
||||||
|
*buffer_end -= n_zeros;
|
||||||
|
STRINGLIB_FILL(*buffer_end, '0', n_zeros);
|
||||||
|
}
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* _Py_InsertThousandsGrouping:
|
* _Py_InsertThousandsGrouping:
|
||||||
* @buffer: A pointer to the start of a string.
|
* @buffer: A pointer to the start of a string.
|
||||||
* @n_buffer: The length of the string.
|
* @n_buffer: Number of characters in @buffer.
|
||||||
|
* @digits: A pointer to the digits we're reading from. If count
|
||||||
|
* is non-NULL, this is unused.
|
||||||
* @n_digits: The number of digits in the string, in which we want
|
* @n_digits: The number of digits in the string, in which we want
|
||||||
* to put the grouping chars.
|
* to put the grouping chars.
|
||||||
* @buf_size: The maximum size of the buffer pointed to by buffer.
|
* @min_width: The minimum width of the digits in the output string.
|
||||||
* @count: If non-NULL, points to a variable that will receive the
|
* Output will be zero-padded on the left to fill.
|
||||||
* number of characters we need to insert (and no formatting
|
* @grouping: see definition in localeconv().
|
||||||
* will actually occur).
|
* @thousands_sep: see definition in localeconv().
|
||||||
* @append_zero_char: If non-zero, put a trailing zero at the end of
|
|
||||||
* of the resulting string, if and only if we modified the
|
|
||||||
* string.
|
|
||||||
*
|
*
|
||||||
* Inserts thousand grouping characters (as defined in the current
|
* There are 2 modes: counting and filling. If @buffer is NULL,
|
||||||
* locale) into the string between buffer and buffer+n_digits. If
|
* we are in counting mode, else filling mode.
|
||||||
* count is non-NULL, don't do any formatting, just count the number
|
* If counting, the required buffer size is returned.
|
||||||
* of characters to insert. This is used by the caller to
|
* If filling, we know the buffer will be large enough, so we don't
|
||||||
* appropriately resize the buffer, if needed. If count is non-NULL,
|
* need to pass in the buffer size.
|
||||||
* buffer can be NULL (it is not dereferenced at all in that case).
|
* Inserts thousand grouping characters (as defined by grouping and
|
||||||
|
* thousands_sep) into the string between buffer and buffer+n_digits.
|
||||||
*
|
*
|
||||||
* Return value: 0 on error, else 1. Note that no error can occur if
|
* Return value: 0 on error, else 1. Note that no error can occur if
|
||||||
* count is non-NULL.
|
* count is non-NULL.
|
||||||
*
|
*
|
||||||
* This name won't be used, the includer of this file should define
|
* This name won't be used, the includer of this file should define
|
||||||
* it to be the actual function name, based on unicode or string.
|
* it to be the actual function name, based on unicode or string.
|
||||||
|
*
|
||||||
|
* As closely as possible, this code mimics the logic in decimal.py's
|
||||||
|
_insert_thousands_sep().
|
||||||
**/
|
**/
|
||||||
int
|
Py_ssize_t
|
||||||
_Py_InsertThousandsGrouping(STRINGLIB_CHAR *buffer,
|
_Py_InsertThousandsGrouping(STRINGLIB_CHAR *buffer,
|
||||||
Py_ssize_t n_buffer,
|
Py_ssize_t n_buffer,
|
||||||
Py_ssize_t n_digits,
|
STRINGLIB_CHAR *digits,
|
||||||
Py_ssize_t buf_size,
|
Py_ssize_t n_digits,
|
||||||
Py_ssize_t *count,
|
Py_ssize_t min_width,
|
||||||
int append_zero_char)
|
const char *grouping,
|
||||||
|
const char *thousands_sep)
|
||||||
{
|
{
|
||||||
struct lconv *locale_data = localeconv();
|
Py_ssize_t count = 0;
|
||||||
const char *grouping = locale_data->grouping;
|
Py_ssize_t n_zeros;
|
||||||
const char *thousands_sep = locale_data->thousands_sep;
|
int loop_broken = 0;
|
||||||
Py_ssize_t thousands_sep_len = strlen(thousands_sep);
|
int use_separator = 0; /* First time through, don't append the
|
||||||
STRINGLIB_CHAR *pend = NULL; /* current end of buffer */
|
separator. They only go between
|
||||||
STRINGLIB_CHAR *pmax = NULL; /* max of buffer */
|
groups. */
|
||||||
char current_grouping;
|
STRINGLIB_CHAR *buffer_end = NULL;
|
||||||
Py_ssize_t remaining = n_digits; /* Number of chars remaining to
|
STRINGLIB_CHAR *digits_end = NULL;
|
||||||
be looked at */
|
Py_ssize_t l;
|
||||||
|
Py_ssize_t n_chars;
|
||||||
|
Py_ssize_t thousands_sep_len = strlen(thousands_sep);
|
||||||
|
Py_ssize_t remaining = n_digits; /* Number of chars remaining to
|
||||||
|
be looked at */
|
||||||
|
/* A generator that returns all of the grouping widths, until it
|
||||||
|
returns 0. */
|
||||||
|
GroupGenerator groupgen;
|
||||||
|
_GroupGenerator_init(&groupgen, grouping);
|
||||||
|
|
||||||
/* Initialize the character count, if we're just counting. */
|
if (buffer) {
|
||||||
if (count)
|
buffer_end = buffer + n_buffer;
|
||||||
*count = 0;
|
digits_end = digits + n_digits;
|
||||||
else {
|
}
|
||||||
/* We're not just counting, we're modifying buffer */
|
|
||||||
pend = buffer + n_buffer;
|
|
||||||
pmax = buffer + buf_size;
|
|
||||||
}
|
|
||||||
|
|
||||||
/* Starting at the end and working right-to-left, keep track of
|
while ((l = _GroupGenerator_next(&groupgen)) > 0) {
|
||||||
what grouping needs to be added and insert that. */
|
l = MIN(l, MAX(MAX(remaining, min_width), 1));
|
||||||
current_grouping = *grouping++;
|
n_zeros = MAX(0, l - remaining);
|
||||||
|
n_chars = MAX(0, MIN(remaining, l));
|
||||||
|
|
||||||
/* If the first character is 0, perform no grouping at all. */
|
/* Use n_zero zero's and n_chars chars */
|
||||||
if (current_grouping == 0)
|
|
||||||
return 1;
|
|
||||||
|
|
||||||
while (remaining > current_grouping) {
|
/* Count only, don't do anything. */
|
||||||
/* Always leave buffer and pend valid at the end of this
|
count += (use_separator ? thousands_sep_len : 0) + n_zeros + n_chars;
|
||||||
loop, since we might leave with a return statement. */
|
|
||||||
|
|
||||||
remaining -= current_grouping;
|
if (buffer) {
|
||||||
if (count) {
|
/* Copy into the output buffer. */
|
||||||
/* We're only counting, not touching the memory. */
|
fill(&digits_end, &buffer_end, n_chars, n_zeros,
|
||||||
*count += thousands_sep_len;
|
use_separator ? thousands_sep : NULL, thousands_sep_len);
|
||||||
}
|
}
|
||||||
else {
|
|
||||||
/* Do the formatting. */
|
|
||||||
|
|
||||||
STRINGLIB_CHAR *plast = buffer + remaining;
|
/* Use a separator next time. */
|
||||||
|
use_separator = 1;
|
||||||
|
|
||||||
/* Is there room to insert thousands_sep_len chars? */
|
remaining -= n_chars;
|
||||||
if (pmax - pend < thousands_sep_len)
|
min_width -= l;
|
||||||
/* No room. */
|
|
||||||
return 0;
|
|
||||||
|
|
||||||
/* Move the rest of the string down. */
|
if (remaining <= 0 && min_width <= 0) {
|
||||||
memmove(plast + thousands_sep_len,
|
loop_broken = 1;
|
||||||
plast,
|
break;
|
||||||
(pend - plast) * sizeof(STRINGLIB_CHAR));
|
}
|
||||||
/* Copy the thousands_sep chars into the buffer. */
|
min_width -= thousands_sep_len;
|
||||||
#if STRINGLIB_IS_UNICODE
|
}
|
||||||
/* Convert from the char's of the thousands_sep from
|
if (!loop_broken) {
|
||||||
the locale into unicode. */
|
/* We left the loop without using a break statement. */
|
||||||
{
|
|
||||||
Py_ssize_t i;
|
|
||||||
for (i = 0; i < thousands_sep_len; ++i)
|
|
||||||
plast[i] = thousands_sep[i];
|
|
||||||
}
|
|
||||||
#else
|
|
||||||
/* No conversion, just memcpy the thousands_sep. */
|
|
||||||
memcpy(plast, thousands_sep, thousands_sep_len);
|
|
||||||
#endif
|
|
||||||
}
|
|
||||||
|
|
||||||
/* Adjust end pointer. */
|
l = MAX(MAX(remaining, min_width), 1);
|
||||||
pend += thousands_sep_len;
|
n_zeros = MAX(0, l - remaining);
|
||||||
|
n_chars = MAX(0, MIN(remaining, l));
|
||||||
|
|
||||||
/* Move to the next grouping character, unless we're
|
/* Use n_zero zero's and n_chars chars */
|
||||||
repeating (which is designated by a grouping of 0). */
|
count += (use_separator ? thousands_sep_len : 0) + n_zeros + n_chars;
|
||||||
if (*grouping != 0) {
|
if (buffer) {
|
||||||
current_grouping = *grouping++;
|
/* Copy into the output buffer. */
|
||||||
if (current_grouping == CHAR_MAX)
|
fill(&digits_end, &buffer_end, n_chars, n_zeros,
|
||||||
/* We're done. */
|
use_separator ? thousands_sep : NULL, thousands_sep_len);
|
||||||
break;
|
}
|
||||||
}
|
}
|
||||||
}
|
return count;
|
||||||
if (append_zero_char) {
|
}
|
||||||
/* Append a zero character to mark the end of the string,
|
|
||||||
if there's room. */
|
/**
|
||||||
if (pend - (buffer + remaining) < 1)
|
* _Py_InsertThousandsGroupingLocale:
|
||||||
/* No room, error. */
|
* @buffer: A pointer to the start of a string.
|
||||||
return 0;
|
* @n_digits: The number of digits in the string, in which we want
|
||||||
*pend = 0;
|
* to put the grouping chars.
|
||||||
}
|
*
|
||||||
return 1;
|
* Reads thee current locale and calls _Py_InsertThousandsGrouping().
|
||||||
|
**/
|
||||||
|
Py_ssize_t
|
||||||
|
_Py_InsertThousandsGroupingLocale(STRINGLIB_CHAR *buffer,
|
||||||
|
Py_ssize_t n_buffer,
|
||||||
|
STRINGLIB_CHAR *digits,
|
||||||
|
Py_ssize_t n_digits,
|
||||||
|
Py_ssize_t min_width)
|
||||||
|
{
|
||||||
|
struct lconv *locale_data = localeconv();
|
||||||
|
const char *grouping = locale_data->grouping;
|
||||||
|
const char *thousands_sep = locale_data->thousands_sep;
|
||||||
|
|
||||||
|
return _Py_InsertThousandsGrouping(buffer, n_buffer, digits, n_digits,
|
||||||
|
min_width, grouping, thousands_sep);
|
||||||
}
|
}
|
||||||
#endif /* STRINGLIB_LOCALEUTIL_H */
|
#endif /* STRINGLIB_LOCALEUTIL_H */
|
||||||
|
|
|
@ -6,6 +6,15 @@
|
||||||
compiled as unicode. */
|
compiled as unicode. */
|
||||||
#define STRINGLIB_IS_UNICODE 0
|
#define STRINGLIB_IS_UNICODE 0
|
||||||
|
|
||||||
|
/* _tolower and _toupper are defined by SUSv2, but they're not ISO C */
|
||||||
|
/* This needs to be cleaned up. See issue 5793. */
|
||||||
|
#ifndef _tolower
|
||||||
|
#define _tolower tolower
|
||||||
|
#endif
|
||||||
|
#ifndef _toupper
|
||||||
|
#define _toupper toupper
|
||||||
|
#endif
|
||||||
|
|
||||||
#define STRINGLIB_OBJECT PyStringObject
|
#define STRINGLIB_OBJECT PyStringObject
|
||||||
#define STRINGLIB_CHAR char
|
#define STRINGLIB_CHAR char
|
||||||
#define STRINGLIB_TYPE_NAME "string"
|
#define STRINGLIB_TYPE_NAME "string"
|
||||||
|
@ -13,8 +22,8 @@
|
||||||
#define STRINGLIB_EMPTY nullstring
|
#define STRINGLIB_EMPTY nullstring
|
||||||
#define STRINGLIB_ISDECIMAL(x) ((x >= '0') && (x <= '9'))
|
#define STRINGLIB_ISDECIMAL(x) ((x >= '0') && (x <= '9'))
|
||||||
#define STRINGLIB_TODECIMAL(x) (STRINGLIB_ISDECIMAL(x) ? (x - '0') : -1)
|
#define STRINGLIB_TODECIMAL(x) (STRINGLIB_ISDECIMAL(x) ? (x - '0') : -1)
|
||||||
#define STRINGLIB_TOUPPER toupper
|
#define STRINGLIB_TOUPPER(x) _toupper(Py_CHARMASK(x))
|
||||||
#define STRINGLIB_TOLOWER tolower
|
#define STRINGLIB_TOLOWER(x) _tolower(Py_CHARMASK(x))
|
||||||
#define STRINGLIB_FILL memset
|
#define STRINGLIB_FILL memset
|
||||||
#define STRINGLIB_STR PyString_AS_STRING
|
#define STRINGLIB_STR PyString_AS_STRING
|
||||||
#define STRINGLIB_LEN PyString_GET_SIZE
|
#define STRINGLIB_LEN PyString_GET_SIZE
|
||||||
|
@ -24,5 +33,6 @@
|
||||||
#define STRINGLIB_CMP memcmp
|
#define STRINGLIB_CMP memcmp
|
||||||
#define STRINGLIB_TOSTR PyObject_Str
|
#define STRINGLIB_TOSTR PyObject_Str
|
||||||
#define STRINGLIB_GROUPING _PyString_InsertThousandsGrouping
|
#define STRINGLIB_GROUPING _PyString_InsertThousandsGrouping
|
||||||
|
#define STRINGLIB_GROUPING_LOCALE _PyString_InsertThousandsGroupingLocale
|
||||||
|
|
||||||
#endif /* !STRINGLIB_STRINGDEFS_H */
|
#endif /* !STRINGLIB_STRINGDEFS_H */
|
||||||
|
|
|
@ -37,6 +37,15 @@
|
||||||
*
|
*
|
||||||
* Return value: the #gdouble value.
|
* Return value: the #gdouble value.
|
||||||
**/
|
**/
|
||||||
|
|
||||||
|
/*
|
||||||
|
Use system strtod; since strtod is locale aware, we may
|
||||||
|
have to first fix the decimal separator.
|
||||||
|
|
||||||
|
Note that unlike _Py_dg_strtod, the system strtod may not always give
|
||||||
|
correctly rounded results.
|
||||||
|
*/
|
||||||
|
|
||||||
double
|
double
|
||||||
PyOS_ascii_strtod(const char *nptr, char **endptr)
|
PyOS_ascii_strtod(const char *nptr, char **endptr)
|
||||||
{
|
{
|
||||||
|
@ -187,6 +196,13 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
|
||||||
return val;
|
return val;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
double
|
||||||
|
PyOS_ascii_atof(const char *nptr)
|
||||||
|
{
|
||||||
|
return PyOS_ascii_strtod(nptr, NULL);
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
/* Given a string that may have a decimal point in the current
|
/* Given a string that may have a decimal point in the current
|
||||||
locale, change it back to a dot. Since the string cannot get
|
locale, change it back to a dot. Since the string cannot get
|
||||||
longer, no need for a maximum buffer size parameter. */
|
longer, no need for a maximum buffer size parameter. */
|
||||||
|
@ -292,8 +308,9 @@ ensure_minumim_exponent_length(char* buffer, size_t buf_size)
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/* Ensure that buffer has a decimal point in it. The decimal point
|
/* Ensure that buffer has a decimal point in it. The decimal point will not
|
||||||
will not be in the current locale, it will always be '.' */
|
be in the current locale, it will always be '.'. Don't add a decimal if an
|
||||||
|
exponent is present. */
|
||||||
Py_LOCAL_INLINE(void)
|
Py_LOCAL_INLINE(void)
|
||||||
ensure_decimal_point(char* buffer, size_t buf_size)
|
ensure_decimal_point(char* buffer, size_t buf_size)
|
||||||
{
|
{
|
||||||
|
@ -322,7 +339,8 @@ ensure_decimal_point(char* buffer, size_t buf_size)
|
||||||
insert_count = 1;
|
insert_count = 1;
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
else {
|
else if (!(*p == 'e' || *p == 'E')) {
|
||||||
|
/* Don't add ".0" if we have an exponent. */
|
||||||
chars_to_insert = ".0";
|
chars_to_insert = ".0";
|
||||||
insert_count = 2;
|
insert_count = 2;
|
||||||
}
|
}
|
||||||
|
@ -341,37 +359,6 @@ ensure_decimal_point(char* buffer, size_t buf_size)
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/* Add the locale specific grouping characters to buffer. Note
|
|
||||||
that any decimal point (if it's present) in buffer is already
|
|
||||||
locale-specific. Return 0 on error, else 1. */
|
|
||||||
Py_LOCAL_INLINE(int)
|
|
||||||
add_thousands_grouping(char* buffer, size_t buf_size)
|
|
||||||
{
|
|
||||||
Py_ssize_t len = strlen(buffer);
|
|
||||||
struct lconv *locale_data = localeconv();
|
|
||||||
const char *decimal_point = locale_data->decimal_point;
|
|
||||||
|
|
||||||
/* Find the decimal point, if any. We're only concerned
|
|
||||||
about the characters to the left of the decimal when
|
|
||||||
adding grouping. */
|
|
||||||
char *p = strstr(buffer, decimal_point);
|
|
||||||
if (!p) {
|
|
||||||
/* No decimal, use the entire string. */
|
|
||||||
|
|
||||||
/* If any exponent, adjust p. */
|
|
||||||
p = strpbrk(buffer, "eE");
|
|
||||||
if (!p)
|
|
||||||
/* No exponent and no decimal. Use the entire
|
|
||||||
string. */
|
|
||||||
p = buffer + len;
|
|
||||||
}
|
|
||||||
/* At this point, p points just past the right-most character we
|
|
||||||
want to format. We need to add the grouping string for the
|
|
||||||
characters between buffer and p. */
|
|
||||||
return _PyString_InsertThousandsGrouping(buffer, len, p-buffer,
|
|
||||||
buf_size, NULL, 1);
|
|
||||||
}
|
|
||||||
|
|
||||||
/* see FORMATBUFLEN in unicodeobject.c */
|
/* see FORMATBUFLEN in unicodeobject.c */
|
||||||
#define FLOAT_FORMATBUFLEN 120
|
#define FLOAT_FORMATBUFLEN 120
|
||||||
|
|
||||||
|
@ -386,9 +373,8 @@ add_thousands_grouping(char* buffer, size_t buf_size)
|
||||||
* Converts a #gdouble to a string, using the '.' as
|
* Converts a #gdouble to a string, using the '.' as
|
||||||
* decimal point. To format the number you pass in
|
* decimal point. To format the number you pass in
|
||||||
* a printf()-style format string. Allowed conversion
|
* a printf()-style format string. Allowed conversion
|
||||||
* specifiers are 'e', 'E', 'f', 'F', 'g', 'G', and 'n'.
|
* specifiers are 'e', 'E', 'f', 'F', 'g', 'G', and 'Z'.
|
||||||
*
|
*
|
||||||
* 'n' is the same as 'g', except it uses the current locale.
|
|
||||||
* 'Z' is the same as 'g', except it always has a decimal and
|
* 'Z' is the same as 'g', except it always has a decimal and
|
||||||
* at least one digit after the decimal.
|
* at least one digit after the decimal.
|
||||||
*
|
*
|
||||||
|
@ -403,11 +389,6 @@ PyOS_ascii_formatd(char *buffer,
|
||||||
char format_char;
|
char format_char;
|
||||||
size_t format_len = strlen(format);
|
size_t format_len = strlen(format);
|
||||||
|
|
||||||
/* For type 'n', we need to make a copy of the format string, because
|
|
||||||
we're going to modify 'n' -> 'g', and format is const char*, so we
|
|
||||||
can't modify it directly. FLOAT_FORMATBUFLEN should be longer than
|
|
||||||
we ever need this to be. There's an upcoming check to ensure it's
|
|
||||||
big enough. */
|
|
||||||
/* Issue 2264: code 'Z' requires copying the format. 'Z' is 'g', but
|
/* Issue 2264: code 'Z' requires copying the format. 'Z' is 'g', but
|
||||||
also with at least one character past the decimal. */
|
also with at least one character past the decimal. */
|
||||||
char tmp_format[FLOAT_FORMATBUFLEN];
|
char tmp_format[FLOAT_FORMATBUFLEN];
|
||||||
|
@ -433,12 +414,12 @@ PyOS_ascii_formatd(char *buffer,
|
||||||
if (!(format_char == 'e' || format_char == 'E' ||
|
if (!(format_char == 'e' || format_char == 'E' ||
|
||||||
format_char == 'f' || format_char == 'F' ||
|
format_char == 'f' || format_char == 'F' ||
|
||||||
format_char == 'g' || format_char == 'G' ||
|
format_char == 'g' || format_char == 'G' ||
|
||||||
format_char == 'n' || format_char == 'Z'))
|
format_char == 'Z'))
|
||||||
return NULL;
|
return NULL;
|
||||||
|
|
||||||
/* Map 'n' or 'Z' format_char to 'g', by copying the format string and
|
/* Map 'Z' format_char to 'g', by copying the format string and
|
||||||
replacing the final char with a 'g' */
|
replacing the final char with a 'g' */
|
||||||
if (format_char == 'n' || format_char == 'Z') {
|
if (format_char == 'Z') {
|
||||||
if (format_len + 1 >= sizeof(tmp_format)) {
|
if (format_len + 1 >= sizeof(tmp_format)) {
|
||||||
/* The format won't fit in our copy. Error out. In
|
/* The format won't fit in our copy. Error out. In
|
||||||
practice, this will never happen and will be
|
practice, this will never happen and will be
|
||||||
|
@ -457,11 +438,8 @@ PyOS_ascii_formatd(char *buffer,
|
||||||
/* Do various fixups on the return string */
|
/* Do various fixups on the return string */
|
||||||
|
|
||||||
/* Get the current locale, and find the decimal point string.
|
/* Get the current locale, and find the decimal point string.
|
||||||
Convert that string back to a dot. Do not do this if using the
|
Convert that string back to a dot. */
|
||||||
'n' (number) format code, since we want to keep the localized
|
change_decimal_from_locale_to_dot(buffer);
|
||||||
decimal point in that case. */
|
|
||||||
if (format_char != 'n')
|
|
||||||
change_decimal_from_locale_to_dot(buffer);
|
|
||||||
|
|
||||||
/* If an exponent exists, ensure that the exponent is at least
|
/* If an exponent exists, ensure that the exponent is at least
|
||||||
MIN_EXPONENT_DIGITS digits, providing the buffer is large enough
|
MIN_EXPONENT_DIGITS digits, providing the buffer is large enough
|
||||||
|
@ -475,16 +453,111 @@ PyOS_ascii_formatd(char *buffer,
|
||||||
if (format_char == 'Z')
|
if (format_char == 'Z')
|
||||||
ensure_decimal_point(buffer, buf_size);
|
ensure_decimal_point(buffer, buf_size);
|
||||||
|
|
||||||
/* If format_char is 'n', add the thousands grouping. */
|
|
||||||
if (format_char == 'n')
|
|
||||||
if (!add_thousands_grouping(buffer, buf_size))
|
|
||||||
return NULL;
|
|
||||||
|
|
||||||
return buffer;
|
return buffer;
|
||||||
}
|
}
|
||||||
|
|
||||||
double
|
PyAPI_FUNC(char *) PyOS_double_to_string(double val,
|
||||||
PyOS_ascii_atof(const char *nptr)
|
char format_code,
|
||||||
|
int precision,
|
||||||
|
int flags,
|
||||||
|
int *type)
|
||||||
{
|
{
|
||||||
return PyOS_ascii_strtod(nptr, NULL);
|
char buf[128];
|
||||||
|
char format[32];
|
||||||
|
Py_ssize_t len;
|
||||||
|
char *result;
|
||||||
|
char *p;
|
||||||
|
int t;
|
||||||
|
int upper = 0;
|
||||||
|
|
||||||
|
/* Validate format_code, and map upper and lower case */
|
||||||
|
switch (format_code) {
|
||||||
|
case 'e': /* exponent */
|
||||||
|
case 'f': /* fixed */
|
||||||
|
case 'g': /* general */
|
||||||
|
break;
|
||||||
|
case 'E':
|
||||||
|
upper = 1;
|
||||||
|
format_code = 'e';
|
||||||
|
break;
|
||||||
|
case 'F':
|
||||||
|
upper = 1;
|
||||||
|
format_code = 'f';
|
||||||
|
break;
|
||||||
|
case 'G':
|
||||||
|
upper = 1;
|
||||||
|
format_code = 'g';
|
||||||
|
break;
|
||||||
|
case 'r': /* repr format */
|
||||||
|
/* Supplied precision is unused, must be 0. */
|
||||||
|
if (precision != 0) {
|
||||||
|
PyErr_BadInternalCall();
|
||||||
|
return NULL;
|
||||||
|
}
|
||||||
|
precision = 17;
|
||||||
|
format_code = 'g';
|
||||||
|
break;
|
||||||
|
case 's': /* str format */
|
||||||
|
/* Supplied precision is unused, must be 0. */
|
||||||
|
if (precision != 0) {
|
||||||
|
PyErr_BadInternalCall();
|
||||||
|
return NULL;
|
||||||
|
}
|
||||||
|
precision = 12;
|
||||||
|
format_code = 'g';
|
||||||
|
break;
|
||||||
|
default:
|
||||||
|
PyErr_BadInternalCall();
|
||||||
|
return NULL;
|
||||||
|
}
|
||||||
|
|
||||||
|
/* Handle nan and inf. */
|
||||||
|
if (Py_IS_NAN(val)) {
|
||||||
|
strcpy(buf, "nan");
|
||||||
|
t = Py_DTST_NAN;
|
||||||
|
} else if (Py_IS_INFINITY(val)) {
|
||||||
|
if (copysign(1., val) == 1.)
|
||||||
|
strcpy(buf, "inf");
|
||||||
|
else
|
||||||
|
strcpy(buf, "-inf");
|
||||||
|
t = Py_DTST_INFINITE;
|
||||||
|
} else {
|
||||||
|
t = Py_DTST_FINITE;
|
||||||
|
|
||||||
|
|
||||||
|
if (flags & Py_DTSF_ADD_DOT_0)
|
||||||
|
format_code = 'Z';
|
||||||
|
|
||||||
|
PyOS_snprintf(format, 32, "%%%s.%i%c", (flags & Py_DTSF_ALT ? "#" : ""), precision, format_code);
|
||||||
|
PyOS_ascii_formatd(buf, sizeof(buf), format, val);
|
||||||
|
}
|
||||||
|
|
||||||
|
len = strlen(buf);
|
||||||
|
|
||||||
|
/* Add 1 for the trailing 0 byte.
|
||||||
|
Add 1 because we might need to make room for the sign.
|
||||||
|
*/
|
||||||
|
result = PyMem_Malloc(len + 2);
|
||||||
|
if (result == NULL) {
|
||||||
|
PyErr_NoMemory();
|
||||||
|
return NULL;
|
||||||
|
}
|
||||||
|
p = result;
|
||||||
|
|
||||||
|
/* Never add sign for nan/inf, even if asked. */
|
||||||
|
if (flags & Py_DTSF_SIGN && buf[0] != '-' && t == Py_DTST_FINITE)
|
||||||
|
*p++ = '+';
|
||||||
|
|
||||||
|
strcpy(p, buf);
|
||||||
|
|
||||||
|
if (upper) {
|
||||||
|
/* Convert to upper case. */
|
||||||
|
char *p1;
|
||||||
|
for (p1 = p; *p1; p1++)
|
||||||
|
*p1 = toupper(*p1);
|
||||||
|
}
|
||||||
|
|
||||||
|
if (type)
|
||||||
|
*type = t;
|
||||||
|
return result;
|
||||||
}
|
}
|
||||||
|
|
Loading…
Add table
Add a link
Reference in a new issue