summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorEric Smith <eric@trueblade.com>2009-04-03 14:45:06 (GMT)
committerEric Smith <eric@trueblade.com>2009-04-03 14:45:06 (GMT)
commita3b1ac8dca5afa01d9d542e125052014a9af56eb (patch)
treeb48929a3d3460d116182f81e7fb80dab5548211f
parentf8c8b6d3ea995d9ddce9470fe6fd2a9c956a5d5d (diff)
downloadcpython-a3b1ac8dca5afa01d9d542e125052014a9af56eb.zip
cpython-a3b1ac8dca5afa01d9d542e125052014a9af56eb.tar.gz
cpython-a3b1ac8dca5afa01d9d542e125052014a9af56eb.tar.bz2
Added ',' thousands grouping to int.__format__. See PEP 378.
This is incomplete, but I want to get some version into the next alpha. I am still working on: Documentation. More tests. Implement for floats. In addition, there's an existing bug with 'n' formatting that carries forward to thousands grouping (issue 5515).
-rw-r--r--Include/bytesobject.h14
-rw-r--r--Include/unicodeobject.h13
-rw-r--r--Lib/test/test_types.py11
-rw-r--r--Objects/bytesobject.c1
-rw-r--r--Objects/stringlib/formatter.h35
-rw-r--r--Objects/stringlib/localeutil.h199
-rw-r--r--Objects/stringlib/stringdefs.h1
-rw-r--r--Objects/stringlib/unicodedefs.h1
-rw-r--r--Objects/unicodeobject.c1
-rw-r--r--Python/pystrtod.c2
10 files changed, 190 insertions, 88 deletions
diff --git a/Include/bytesobject.h b/Include/bytesobject.h
index 3f275a8..f5a5085 100644
--- a/Include/bytesobject.h
+++ b/Include/bytesobject.h
@@ -91,13 +91,25 @@ PyAPI_FUNC(int) PyBytes_AsStringAndSize(
into the string pointed to by buffer. For the argument descriptions,
see Objects/stringlib/localeutil.h */
-PyAPI_FUNC(int) _PyBytes_InsertThousandsGrouping(char *buffer,
+PyAPI_FUNC(int) _PyBytes_InsertThousandsGroupingLocale(char *buffer,
Py_ssize_t n_buffer,
Py_ssize_t n_digits,
Py_ssize_t buf_size,
Py_ssize_t *count,
int append_zero_char);
+/* Using explicit passed-in values, insert the thousands grouping
+ into the string pointed to by buffer. For the argument descriptions,
+ see Objects/stringlib/localeutil.h */
+PyAPI_FUNC(int) _PyBytes_InsertThousandsGrouping(char *buffer,
+ Py_ssize_t n_buffer,
+ Py_ssize_t n_digits,
+ Py_ssize_t buf_size,
+ Py_ssize_t *count,
+ int append_zero_char,
+ const char *grouping,
+ const char *thousands_sep);
+
/* Flags used by string formatting */
#define F_LJUST (1<<0)
#define F_SIGN (1<<1)
diff --git a/Include/unicodeobject.h b/Include/unicodeobject.h
index 8259743..98c0372 100644
--- a/Include/unicodeobject.h
+++ b/Include/unicodeobject.h
@@ -1482,13 +1482,24 @@ PyAPI_FUNC(PyObject *) _PyUnicode_XStrip(
into the string pointed to by buffer. For the argument descriptions,
see Objects/stringlib/localeutil.h */
-PyAPI_FUNC(int) _PyUnicode_InsertThousandsGrouping(Py_UNICODE *buffer,
+PyAPI_FUNC(int) _PyUnicode_InsertThousandsGroupingLocale(Py_UNICODE *buffer,
Py_ssize_t n_buffer,
Py_ssize_t n_digits,
Py_ssize_t buf_size,
Py_ssize_t *count,
int append_zero_char);
+/* Using explicit passed-in values, insert the thousands grouping
+ into the string pointed to by buffer. For the argument descriptions,
+ see Objects/stringlib/localeutil.h */
+PyAPI_FUNC(int) _PyUnicode_InsertThousandsGrouping(Py_UNICODE *buffer,
+ Py_ssize_t n_buffer,
+ Py_ssize_t n_digits,
+ Py_ssize_t buf_size,
+ Py_ssize_t *count,
+ int append_zero_char,
+ const char *grouping,
+ const char *thousands_sep);
/* === Characters Type APIs =============================================== */
/* Helper array used by Py_UNICODE_ISSPACE(). */
diff --git a/Lib/test/test_types.py b/Lib/test/test_types.py
index c200e07..0d4d1b3 100644
--- a/Lib/test/test_types.py
+++ b/Lib/test/test_types.py
@@ -338,6 +338,15 @@ class TypesTests(unittest.TestCase):
test(123456, "#012X", '0X000001E240')
test(-123456, "#012X", '-0X00001E240')
+ test(123, ',', '123')
+ test(-123, ',', '-123')
+ test(1234, ',', '1,234')
+ test(-1234, ',', '-1,234')
+ test(123456, ',', '123,456')
+ test(-123456, ',', '-123,456')
+ test(1234567, ',', '1,234,567')
+ test(-1234567, ',', '-1,234,567')
+
# make sure these are errors
# precision disallowed
@@ -347,6 +356,8 @@ class TypesTests(unittest.TestCase):
# format spec must be string
self.assertRaises(TypeError, 3 .__format__, None)
self.assertRaises(TypeError, 3 .__format__, 0)
+ # can't have ',' with 'n'
+ self.assertRaises(ValueError, 3 .__format__, ",n")
# ensure that only int and float type specifiers work
for format_spec in ([chr(x) for x in range(ord('a'), ord('z')+1)] +
diff --git a/Objects/bytesobject.c b/Objects/bytesobject.c
index d3b598e..d5c2bea 100644
--- a/Objects/bytesobject.c
+++ b/Objects/bytesobject.c
@@ -583,6 +583,7 @@ PyBytes_AsStringAndSize(register PyObject *obj,
#include "stringlib/transmogrify.h"
#define _Py_InsertThousandsGrouping _PyBytes_InsertThousandsGrouping
+#define _Py_InsertThousandsGroupingLocale _PyBytes_InsertThousandsGroupingLocale
#include "stringlib/localeutil.h"
PyObject *
diff --git a/Objects/stringlib/formatter.h b/Objects/stringlib/formatter.h
index c367ec5..57e5452 100644
--- a/Objects/stringlib/formatter.h
+++ b/Objects/stringlib/formatter.h
@@ -120,6 +120,7 @@ typedef struct {
int alternate;
STRINGLIB_CHAR sign;
Py_ssize_t width;
+ int thousands_separators;
Py_ssize_t precision;
STRINGLIB_CHAR type;
} InternalFormatSpec;
@@ -149,6 +150,7 @@ parse_internal_render_format_spec(STRINGLIB_CHAR *format_spec,
format->alternate = 0;
format->sign = '\0';
format->width = -1;
+ format->thousands_separators = 0;
format->precision = -1;
format->type = default_type;
@@ -201,6 +203,12 @@ parse_internal_render_format_spec(STRINGLIB_CHAR *format_spec,
format->width = -1;
}
+ /* Comma signifies add thousands separators */
+ if (end-ptr && ptr[0] == ',') {
+ format->thousands_separators = 1;
+ ++ptr;
+ }
+
/* Parse field precision */
if (end-ptr && ptr[0] == '.') {
++ptr;
@@ -230,6 +238,11 @@ parse_internal_render_format_spec(STRINGLIB_CHAR *format_spec,
++ptr;
}
+ if (format->type == 'n' && format->thousands_separators) {
+ PyErr_Format(PyExc_ValueError, "Cannot specify ',' with 'n'.");
+ return 0;
+ }
+
return 1;
}
@@ -630,8 +643,13 @@ format_int_or_long_internal(PyObject *value, const InternalFormatSpec *format,
if (format->type == 'n')
/* Compute how many additional chars we need to allocate
to hold the thousands grouping. */
- STRINGLIB_GROUPING(NULL, n_digits, n_digits,
+ STRINGLIB_GROUPING_LOCALE(NULL, n_digits, n_digits,
0, &n_grouping_chars, 0);
+ if (format->thousands_separators)
+ /* Compute how many additional chars we need to allocate
+ to hold the thousands grouping. */
+ STRINGLIB_GROUPING(NULL, n_digits, n_digits,
+ 0, &n_grouping_chars, 0, "\3", ",");
/* Calculate the widths of the various leading and trailing parts */
calc_number_widths(&spec, sign, n_prefix, n_digits + n_grouping_chars,
@@ -670,11 +688,22 @@ format_int_or_long_internal(PyObject *value, const InternalFormatSpec *format,
reserved enough space. */
STRINGLIB_CHAR *pstart = p + n_leading_chars;
#ifndef NDEBUG
- int r =
+ int r;
+#endif
+ if (format->type == 'n')
+#ifndef NDEBUG
+ r =
#endif
- STRINGLIB_GROUPING(pstart, n_digits, n_digits,
+ STRINGLIB_GROUPING_LOCALE(pstart, n_digits, n_digits,
spec.n_total+n_grouping_chars-n_leading_chars,
NULL, 0);
+ else
+#ifndef NDEBUG
+ r =
+ STRINGLIB_GROUPING(pstart, n_digits, n_digits,
+ spec.n_total+n_grouping_chars-n_leading_chars,
+ NULL, 0, "\3", ",");
+#endif
assert(r);
}
diff --git a/Objects/stringlib/localeutil.h b/Objects/stringlib/localeutil.h
index 1105609..9254c09 100644
--- a/Objects/stringlib/localeutil.h
+++ b/Objects/stringlib/localeutil.h
@@ -18,11 +18,13 @@
* @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.
+ * @grouping: see definition in localeconv().
+ * @thousands_sep: see definition in localeconv().
*
- * Inserts thousand grouping characters (as defined in the current
- * locale) into the string between buffer and buffer+n_digits. If
- * count is non-NULL, don't do any formatting, just count the number
- * of characters to insert. This is used by the caller to
+ * Inserts thousand grouping characters (as defined by grouping and
+ * thousands_sep) into the string between buffer and buffer+n_digits.
+ * If count is non-NULL, don't do any formatting, just count the
+ * number of characters to insert. This is used by the caller to
* appropriately resize the buffer, if needed. If count is non-NULL,
* buffer can be NULL (it is not dereferenced at all in that case).
*
@@ -34,97 +36,130 @@
**/
int
_Py_InsertThousandsGrouping(STRINGLIB_CHAR *buffer,
- Py_ssize_t n_buffer,
- Py_ssize_t n_digits,
- Py_ssize_t buf_size,
- Py_ssize_t *count,
- int append_zero_char)
+ Py_ssize_t n_buffer,
+ Py_ssize_t n_digits,
+ Py_ssize_t buf_size,
+ Py_ssize_t *count,
+ int append_zero_char,
+ const char *grouping,
+ const char *thousands_sep)
{
- struct lconv *locale_data = localeconv();
- const char *grouping = locale_data->grouping;
- const char *thousands_sep = locale_data->thousands_sep;
- Py_ssize_t thousands_sep_len = strlen(thousands_sep);
- STRINGLIB_CHAR *pend = NULL; /* current end of buffer */
- STRINGLIB_CHAR *pmax = NULL; /* max of buffer */
- char current_grouping;
- Py_ssize_t remaining = n_digits; /* Number of chars remaining to
- be looked at */
+ Py_ssize_t thousands_sep_len = strlen(thousands_sep);
+ STRINGLIB_CHAR *pend = NULL; /* current end of buffer */
+ STRINGLIB_CHAR *pmax = NULL; /* max of buffer */
+ char current_grouping;
+ Py_ssize_t remaining = n_digits; /* Number of chars remaining to
+ be looked at */
- /* Initialize the character count, if we're just counting. */
- if (count)
- *count = 0;
- else {
- /* We're not just counting, we're modifying buffer */
- pend = buffer + n_buffer;
- pmax = buffer + buf_size;
- }
+ /* Initialize the character count, if we're just counting. */
+ if (count)
+ *count = 0;
+ 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
- what grouping needs to be added and insert that. */
- current_grouping = *grouping++;
+ /* Starting at the end and working right-to-left, keep track of
+ what grouping needs to be added and insert that. */
+ current_grouping = *grouping++;
- /* If the first character is 0, perform no grouping at all. */
- if (current_grouping == 0)
- return 1;
+ /* If the first character is 0, perform no grouping at all. */
+ if (current_grouping == 0)
+ return 1;
- while (remaining > current_grouping) {
- /* Always leave buffer and pend valid at the end of this
- loop, since we might leave with a return statement. */
+ while (remaining > current_grouping) {
+ /* Always leave buffer and pend valid at the end of this
+ loop, since we might leave with a return statement. */
- remaining -= current_grouping;
- if (count) {
- /* We're only counting, not touching the memory. */
- *count += thousands_sep_len;
- }
- else {
- /* Do the formatting. */
+ remaining -= current_grouping;
+ if (count) {
+ /* We're only counting, not touching the memory. */
+ *count += thousands_sep_len;
+ }
+ else {
+ /* Do the formatting. */
- STRINGLIB_CHAR *plast = buffer + remaining;
+ STRINGLIB_CHAR *plast = buffer + remaining;
- /* Is there room to insert thousands_sep_len chars? */
- if (pmax - pend < thousands_sep_len)
- /* No room. */
- return 0;
+ /* Is there room to insert thousands_sep_len chars? */
+ if (pmax - pend < thousands_sep_len)
+ /* No room. */
+ return 0;
- /* Move the rest of the string down. */
- memmove(plast + thousands_sep_len,
- plast,
- (pend - plast) * sizeof(STRINGLIB_CHAR));
- /* Copy the thousands_sep chars into the buffer. */
+ /* Move the rest of the string down. */
+ memmove(plast + thousands_sep_len,
+ plast,
+ (pend - plast) * sizeof(STRINGLIB_CHAR));
+ /* 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. */
- {
- Py_ssize_t i;
- for (i = 0; i < thousands_sep_len; ++i)
- plast[i] = thousands_sep[i];
- }
+ /* Convert from the char's of the thousands_sep from
+ the locale into unicode. */
+ {
+ 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);
+ /* No conversion, just memcpy the thousands_sep. */
+ memcpy(plast, thousands_sep, thousands_sep_len);
#endif
- }
+ }
- /* Adjust end pointer. */
- pend += thousands_sep_len;
+ /* Adjust end pointer. */
+ pend += thousands_sep_len;
- /* Move to the next grouping character, unless we're
- repeating (which is designated by a grouping of 0). */
- if (*grouping != 0) {
- current_grouping = *grouping++;
- if (current_grouping == CHAR_MAX)
- /* We're done. */
- break;
- }
- }
- if (append_zero_char) {
- /* Append a zero character to mark the end of the string,
- if there's room. */
- if (pend - (buffer + remaining) < 1)
- /* No room, error. */
- return 0;
- *pend = 0;
- }
- return 1;
+ /* Move to the next grouping character, unless we're
+ repeating (which is designated by a grouping of 0). */
+ if (*grouping != 0) {
+ current_grouping = *grouping++;
+ if (current_grouping == CHAR_MAX)
+ /* We're done. */
+ break;
+ }
+ }
+ if (append_zero_char) {
+ /* Append a zero character to mark the end of the string,
+ if there's room. */
+ if (pend - (buffer + remaining) < 1)
+ /* No room, error. */
+ return 0;
+ *pend = 0;
+ }
+ return 1;
+}
+
+/**
+ * _Py_InsertThousandsGroupingLocale:
+ * @buffer: A pointer to the start of a string.
+ * @n_buffer: The length of the string.
+ * @n_digits: The number of digits in the string, in which we want
+ * to put the grouping chars.
+ * @buf_size: The maximum size of the buffer pointed to by buffer.
+ * @count: If non-NULL, points to a variable that will receive the
+ * number of characters we need to insert (and no formatting
+ * will actually occur).
+ * @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.
+ *
+ * Reads thee current locale and calls _Py_InsertThousandsGrouping().
+ **/
+int
+_Py_InsertThousandsGroupingLocale(STRINGLIB_CHAR *buffer,
+ Py_ssize_t n_buffer,
+ Py_ssize_t n_digits,
+ Py_ssize_t buf_size,
+ Py_ssize_t *count,
+ int append_zero_char)
+{
+ 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, n_digits,
+ buf_size, count,
+ append_zero_char, grouping,
+ thousands_sep);
}
#endif /* STRINGLIB_LOCALEUTIL_H */
diff --git a/Objects/stringlib/stringdefs.h b/Objects/stringlib/stringdefs.h
index 946de05..1659315 100644
--- a/Objects/stringlib/stringdefs.h
+++ b/Objects/stringlib/stringdefs.h
@@ -24,5 +24,6 @@
#define STRINGLIB_CMP memcmp
#define STRINGLIB_TOSTR PyObject_Str
#define STRINGLIB_GROUPING _PyBytes_InsertThousandsGrouping
+#define STRINGLIB_GROUPING_LOCALE _PyBytes_InsertThousandsGroupingLocale
#define STRINGLIB_TOASCII PyObject_Repr
#endif /* !STRINGLIB_STRINGDEFS_H */
diff --git a/Objects/stringlib/unicodedefs.h b/Objects/stringlib/unicodedefs.h
index d579c1c..366acfe 100644
--- a/Objects/stringlib/unicodedefs.h
+++ b/Objects/stringlib/unicodedefs.h
@@ -22,6 +22,7 @@
#define STRINGLIB_RESIZE PyUnicode_Resize
#define STRINGLIB_CHECK PyUnicode_Check
#define STRINGLIB_GROUPING _PyUnicode_InsertThousandsGrouping
+#define STRINGLIB_GROUPING_LOCALE _PyUnicode_InsertThousandsGroupingLocale
#if PY_VERSION_HEX < 0x03000000
#define STRINGLIB_TOSTR PyObject_Unicode
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index 4def537..d3e5283 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -5635,6 +5635,7 @@ int PyUnicode_EncodeDecimal(Py_UNICODE *s,
#include "stringlib/partition.h"
#define _Py_InsertThousandsGrouping _PyUnicode_InsertThousandsGrouping
+#define _Py_InsertThousandsGroupingLocale _PyUnicode_InsertThousandsGroupingLocale
#include "stringlib/localeutil.h"
/* helper macro to fixup start/end slice values */
diff --git a/Python/pystrtod.c b/Python/pystrtod.c
index b373852..b81abce 100644
--- a/Python/pystrtod.c
+++ b/Python/pystrtod.c
@@ -368,7 +368,7 @@ add_thousands_grouping(char* buffer, size_t buf_size)
/* 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 _PyBytes_InsertThousandsGrouping(buffer, len, p-buffer,
+ return _PyBytes_InsertThousandsGroupingLocale(buffer, len, p-buffer,
buf_size, NULL, 1);
}