Merged revisions 60481,60485,60489-60492,60494-60496,60498-60499,60501-60503,60505-60506,60508-60509,60523-60524,60532,60543,60545,60547-60548,60552,60554,60556-60559,60561-60562,60569,60571-60572,60574,60576-60583,60585-60586,60589,60591,60594-60595,60597-60598,60600-60601,60606-60612,60615,60617,60619-60621,60623-60625,60627-60629,60631,60633,60635,60647,60650,60652,60654,60656,60658-60659,60664-60666,60668-60670,60672,60676,60678,60680-60683,60685-60686,60688,60690,60692-60694,60697-60700,60705-60706,60708,60711,60714,60720,60724-60730,60732,60736,60742,60744,60746,60748,60750-60751,60753,60756-60757,60759-60761,60763-60764,60766,60769-60770,60774-60784,60787-60789,60793,60796,60799-60809,60812-60813,60815-60821,60823-60826,60828-60829,60831-60834,60836,60838-60839,60846-60849,60852-60854,60856-60859,60861-60870,60874-60875,60880-60881,60886,60888-60890,60892,60894-60898,60900-60931,60933-60958 via svnmerge from

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r60901 | eric.smith | 2008-02-19 14:21:56 +0100 (Tue, 19 Feb 2008) | 1 line

  Added PEP 3101.
........
  r60907 | georg.brandl | 2008-02-20 20:12:36 +0100 (Wed, 20 Feb 2008) | 2 lines

  Fixes contributed by Ori Avtalion.
........
  r60909 | eric.smith | 2008-02-21 00:34:22 +0100 (Thu, 21 Feb 2008) | 1 line

  Trim leading zeros from a floating point exponent, per C99.  See issue 1600.  As far as I know, this only affects Windows.  Add float type 'n' to PyOS_ascii_formatd (see PEP 3101 for 'n' description).
........
  r60910 | eric.smith | 2008-02-21 00:39:28 +0100 (Thu, 21 Feb 2008) | 1 line

  Now that PyOS_ascii_formatd supports the 'n' format, simplify the float formatting code to just call it.
........
  r60918 | andrew.kuchling | 2008-02-21 15:23:38 +0100 (Thu, 21 Feb 2008) | 2 lines

  Close manifest file.
  This change doesn't make any difference to CPython, but is a necessary fix for Jython.
........
  r60921 | guido.van.rossum | 2008-02-21 18:46:16 +0100 (Thu, 21 Feb 2008) | 2 lines

  Remove news about float repr() -- issue 1580 is still in limbo.
........
  r60923 | guido.van.rossum | 2008-02-21 19:18:37 +0100 (Thu, 21 Feb 2008) | 5 lines

  Removed uses of dict.has_key() from distutils, and uses of
  callable() from copy_reg.py, so the interpreter now starts up
  without warnings when '-3' is given.  More work like this needs to
  be done in the rest of the stdlib.
........
  r60924 | thomas.heller | 2008-02-21 19:28:48 +0100 (Thu, 21 Feb 2008) | 4 lines

  configure.ac: Remove the configure check for _Bool, it is already done in the
  top-level Python configure script.
  configure, fficonfig.h.in: regenerated.
........
  r60925 | thomas.heller | 2008-02-21 19:52:20 +0100 (Thu, 21 Feb 2008) | 3 lines

  Replace 'has_key()' with 'in'.
  Replace 'raise Error, stuff' with 'raise Error(stuff)'.
........
  r60927 | raymond.hettinger | 2008-02-21 20:24:53 +0100 (Thu, 21 Feb 2008) | 1 line

  Update more instances of has_key().
........
  r60928 | guido.van.rossum | 2008-02-21 20:46:35 +0100 (Thu, 21 Feb 2008) | 3 lines

  Fix a few typos and layout glitches (more work is needed).
  Move 2.5 news to Misc/HISTORY.
........
  r60936 | georg.brandl | 2008-02-21 21:33:38 +0100 (Thu, 21 Feb 2008) | 2 lines

  #2079: typo in userdict docs.
........
  r60938 | georg.brandl | 2008-02-21 21:38:13 +0100 (Thu, 21 Feb 2008) | 2 lines

  Part of #2154: minimal syntax fixes in doc example snippets.
........
  r60942 | raymond.hettinger | 2008-02-22 04:16:42 +0100 (Fri, 22 Feb 2008) | 1 line

  First draft for itertools.product().  Docs and other updates forthcoming.
........
  r60955 | nick.coghlan | 2008-02-22 11:54:06 +0100 (Fri, 22 Feb 2008) | 1 line

  Try to make command line error messages from runpy easier to understand (and suppress traceback cruft from the implicitly invoked runpy machinery)
........
  r60956 | georg.brandl | 2008-02-22 13:31:45 +0100 (Fri, 22 Feb 2008) | 2 lines

  A lot more typo fixes by Ori Avtalion.
........
  r60957 | georg.brandl | 2008-02-22 13:56:34 +0100 (Fri, 22 Feb 2008) | 2 lines

  Don't reference pyshell.
........
  r60958 | georg.brandl | 2008-02-22 13:57:05 +0100 (Fri, 22 Feb 2008) | 2 lines

  Another fix.
........

1 files changed, 127 insertions, 27 deletions

diff --git a/Python/pystrtod.c b/Python/pystrtod.c
index 14fb84b..16efa9d 100644
--- a/Python/pystrtod.c
+++ b/Python/pystrtod.c
@@ -186,6 +186,15 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
 }
 
 
+/* From the C99 standard, section 7.19.6:
+The exponent always contains at least two digits, and only as many more digits
+as necessary to represent the exponent.
+*/
+#define MIN_EXPONENT_DIGITS 2
+
+/* see FORMATBUFLEN in unicodeobject.c */
+#define FLOAT_FORMATBUFLEN 120
+
 /**
  * PyOS_ascii_formatd:
  * @buffer: A buffer to place the resulting string in
@@ -197,8 +206,10 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
  * Converts a #gdouble to a string, using the '.' as
  * decimal point. To format the number you pass in
  * a printf()-style format string. Allowed conversion
- * specifiers are 'e', 'E', 'f', 'F', 'g' and 'G'. 
+ * specifiers are 'e', 'E', 'f', 'F', 'g', 'G', and 'n'.
  * 
+ * 'n' is the same as 'g', except it uses the current locale.
+ *
  * Return value: The pointer to the buffer with the converted string.
  **/
 char *
@@ -207,17 +218,23 @@ PyOS_ascii_formatd(char       *buffer,
 		   const char *format, 
 		   double      d)
 {
-	struct lconv *locale_data;
-	const char *decimal_point;
-	size_t decimal_point_len, rest_len;
 	char *p;
 	char format_char;
+	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. */
+	char tmp_format[FLOAT_FORMATBUFLEN];
 
 /* 	g_return_val_if_fail (buffer != NULL, NULL); */
 /* 	g_return_val_if_fail (format[0] == '%', NULL); */
 /* 	g_return_val_if_fail (strpbrk (format + 1, "'l%") == NULL, NULL); */
 
-	format_char = format[strlen(format) - 1];
+	/* The last character in the format string must be the format char */
+	format_char = format[format_len - 1];
 
 /* 	g_return_val_if_fail (format_char == 'e' || format_char == 'E' || */
 /* 			      format_char == 'f' || format_char == 'F' || */
@@ -227,43 +244,126 @@ PyOS_ascii_formatd(char       *buffer,
 	if (format[0] != '%')
 		return NULL;
 
+	/* I'm not sure why this test is here.  It's ensuring that the format
+	   string after the first character doesn't have a single quote, a
+	   lowercase l, or a percent. This is the reverse of the commented-out
+	   test about 10 lines ago. */
 	if (strpbrk(format + 1, "'l%"))
 		return NULL;
 
 	if (!(format_char == 'e' || format_char == 'E' || 
 	      format_char == 'f' || format_char == 'F' || 
-	      format_char == 'g' || format_char == 'G'))
+	      format_char == 'g' || format_char == 'G' ||
+	      format_char == 'n'))
 		return NULL;
 
+	/* Map 'n' format_char to 'g', by copying the format string and
+	   replacing the final 'n' with a 'g' */
+	if (format_char == 'n') {
+		if (format_len + 1 >= sizeof(tmp_format)) {
+			/* The format won't fit in our copy.  Error out.  In
+			   practice, this will never happen and will be detected
+			   by returning NULL */
+			return NULL;
+		}
+		strcpy(tmp_format, format);
+		tmp_format[format_len - 1] = 'g';
+		format = tmp_format;
+	}
 
+	/* Have PyOS_snprintf do the hard work */
 	PyOS_snprintf(buffer, buf_len, format, d);
 
-	locale_data = localeconv();
-	decimal_point = locale_data->decimal_point;
-	decimal_point_len = strlen(decimal_point);
+	/* Get the current local, and find the decimal point character (or
+	   string?).  Convert that string back to a dot.  Do not do this if
+	   using the 'n' (number) format code. */
+	if (format_char != 'n') {
+		struct lconv *locale_data = localeconv();
+		const char *decimal_point = locale_data->decimal_point;
+		size_t decimal_point_len = strlen(decimal_point);
+		size_t rest_len;
 
-	assert(decimal_point_len != 0);
+		assert(decimal_point_len != 0);
 
-	if (decimal_point[0] != '.' || 
-	    decimal_point[1] != 0)
-	{
-		p = buffer;
+		if (decimal_point[0] != '.' || decimal_point[1] != 0) {
+			p = buffer;
 
-		if (*p == '+' || *p == '-')
-			p++;
+			if (*p == '+' || *p == '-')
+				p++;
 
-		while (isdigit((unsigned char)*p))
-			p++;
+			while (isdigit(Py_CHARMASK(*p)))
+				p++;
 
-		if (strncmp(p, decimal_point, decimal_point_len) == 0)
-		{
-			*p = '.';
-			p++;
-			if (decimal_point_len > 1) {
-				rest_len = strlen(p + (decimal_point_len - 1));
-				memmove(p, p + (decimal_point_len - 1), 
-					rest_len);
-				p[rest_len] = 0;
+			if (strncmp(p, decimal_point, decimal_point_len) == 0) {
+				*p = '.';
+				p++;
+				if (decimal_point_len > 1) {
+					rest_len = strlen(p +
+						      (decimal_point_len - 1));
+					memmove(p, p + (decimal_point_len - 1),
+						rest_len);
+					p[rest_len] = 0;
+				}
+			}
+		}
+	}
+
+	/* If an exponent exists, ensure that the exponent is at least
+	   MIN_EXPONENT_DIGITS digits, providing the buffer is large enough
+	   for the extra zeros.  Also, if there are more than
+	   MIN_EXPONENT_DIGITS, remove as many zeros as possible until we get
+	   back to MIN_EXPONENT_DIGITS */
+	p = strpbrk(buffer, "eE");
+	if (p && (*(p + 1) == '-' || *(p + 1) == '+')) {
+		char *start = p + 2;
+		int exponent_digit_cnt = 0;
+		int leading_zero_cnt = 0;
+		int in_leading_zeros = 1;
+		int significant_digit_cnt;
+
+		p += 2;
+		while (*p && isdigit(Py_CHARMASK(*p))) {
+			if (in_leading_zeros && *p == '0')
+				++leading_zero_cnt;
+			if (*p != '0')
+				in_leading_zeros = 0;
+			++p;
+			++exponent_digit_cnt;
+		}
+
+		significant_digit_cnt = exponent_digit_cnt - leading_zero_cnt;
+		if (exponent_digit_cnt == MIN_EXPONENT_DIGITS) {
+			/* If there are 2 exactly digits, we're done,
+			   regardless of what they contain */
+		}
+		else if (exponent_digit_cnt > MIN_EXPONENT_DIGITS) {
+			int extra_zeros_cnt;
+
+			/* There are more than 2 digits in the exponent.  See
+			   if we can delete some of the leading zeros */
+			if (significant_digit_cnt < MIN_EXPONENT_DIGITS)
+				significant_digit_cnt = MIN_EXPONENT_DIGITS;
+			extra_zeros_cnt = exponent_digit_cnt - significant_digit_cnt;
+
+			/* Delete extra_zeros_cnt worth of characters from the
+			   front of the exponent */
+			assert(extra_zeros_cnt >= 0);
+
+			/* Add one to significant_digit_cnt to copy the
+			   trailing 0 byte, thus setting the length */
+			memmove(start,
+				start + extra_zeros_cnt,
+				significant_digit_cnt + 1);
+		}
+		else {
+			/* If there are fewer than 2 digits, add zeros
+			   until there are 2, if there's enough room */
+			int zeros = MIN_EXPONENT_DIGITS - exponent_digit_cnt;
+			if (start + zeros + exponent_digit_cnt + 1
+			      < buffer + buf_len) {
+				memmove(start + zeros, start,
+					exponent_digit_cnt + 1);
+				memset(start, '0', zeros);
 			}
 		}
 	}