Merged revisions 61431,61433-61436,61439,61444,61449-61450,61453,61458,61465,61468,61471-61474,61480,61483-61484,61488,61495-61496,61498,61503-61504,61507,61509-61510,61515-61518 via svnmerge from

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

........
  r61431 | vinay.sajip | 2008-03-16 22:35:58 +0100 (So, 16 Mär 2008) | 1 line

  Clarified documentation on use of shutdown().
........
  r61433 | mark.summerfield | 2008-03-17 09:28:15 +0100 (Mo, 17 Mär 2008) | 5 lines

  Added a footnote to each pointing out that for XML output if an encoding
  string is given it should conform to the appropriate XML standards---for
  example, "UTF-8" is okay, but "UTF8" is not.
........
  r61434 | eric.smith | 2008-03-17 12:01:01 +0100 (Mo, 17 Mär 2008) | 7 lines

  Issue 2264: empty float presentation type needs to have at least one digit past the decimal point.

  Added "Z" format_char to PyOS_ascii_formatd to support empty float presentation type.
  Renamed buf_size in PyOS_ascii_formatd to more accurately reflect it's meaning.
  Modified format.__float__ to use the new "Z" format as the default.
  Added test cases.
........
  r61435 | eric.smith | 2008-03-17 13:14:29 +0100 (Mo, 17 Mär 2008) | 2 lines

  Reformated lines > 79 chars.
  Deleted unused macro ISXDIGIT.
........
  r61436 | jeffrey.yasskin | 2008-03-17 15:40:53 +0100 (Mo, 17 Mär 2008) | 13 lines

  Allow Gnu gcc's to build python on OSX by removing -Wno-long-double,
  -no-cpp-precomp, and -mno-fused-madd from configure.
   * r22183 added -no-cpp-precomp, which
     http://gcc.gnu.org/ml/gcc/2005-12/msg00368.html claims hasn't been
     needed since gcc-3.1.
   * r25607 added -Wno-long-double to avoid a warning in
     Include/objimpl.h (issue 525481). The long double is still there,
     but OSX 10.4's gcc no longer warns about it.
   * r33666 fixed issue 775892 on OSX 10.3 by adding -mno-fused-madd,
     which changed the sign of some float 0s. Tim Peters said it wasn't
     a real issue anyway, and it no longer causes test failures.
  Fixes issue #1779871.
........
  r61439 | martin.v.loewis | 2008-03-17 17:31:57 +0100 (Mo, 17 Mär 2008) | 2 lines

  Add Trent Nelson.
........
  r61444 | travis.oliphant | 2008-03-17 18:36:12 +0100 (Mo, 17 Mär 2008) | 1 line

  Add necessary headers to back-port new buffer protocol to Python 2.6
........
  r61449 | gregory.p.smith | 2008-03-17 19:48:05 +0100 (Mo, 17 Mär 2008) | 8 lines

  Force zlib.crc32 and zlib.adler32 to return a signed integer on all platforms
  regardless of the native sizeof(long) used in the integer object.

  This somewhat odd behavior of returning a signed is maintained in 2.x for
  compatibility reasons of always returning an integer rather than a long object.

  Fixes Issue1202 for Python 2.6
........
  r61450 | neal.norwitz | 2008-03-17 20:02:45 +0100 (Mo, 17 Mär 2008) | 3 lines

  Use a buffer large enough to ensure we don't overrun, even if the value
  is outside the range we expect.
........
  r61453 | steven.bethard | 2008-03-17 20:33:11 +0100 (Mo, 17 Mär 2008) | 1 line

  Document unicode.isnumeric() and unicode.isdecimal() (issue2326)
........
  r61458 | neal.norwitz | 2008-03-17 21:22:43 +0100 (Mo, 17 Mär 2008) | 5 lines

  Issue 2321: reduce memory usage (increase the memory that is returned
  to the system) by using pymalloc for the data of unicode objects.

  Will backport.
........
  r61465 | martin.v.loewis | 2008-03-17 22:55:30 +0100 (Mo, 17 Mär 2008) | 2 lines

  Add David Wolever.
........
  r61468 | gregory.p.smith | 2008-03-18 01:20:01 +0100 (Di, 18 Mär 2008) | 3 lines

  Fix the IOError message text when opening a file with an invalid filename.
  Error reported by Ilan Schnell.
........
  r61471 | brett.cannon | 2008-03-18 02:00:07 +0100 (Di, 18 Mär 2008) | 2 lines

  Convert test_strftime, test_getargs, and test_pep247 to use unittest.
........
  r61472 | jeffrey.yasskin | 2008-03-18 02:09:59 +0100 (Di, 18 Mär 2008) | 2 lines

  Fix build on platforms that don't have intptr_t. Patch by Joseph Armbruster.
........
  r61473 | brett.cannon | 2008-03-18 02:50:25 +0100 (Di, 18 Mär 2008) | 2 lines

  Convert test_dummy_threading and test_dbm to unittest.
........
  r61474 | brett.cannon | 2008-03-18 02:58:56 +0100 (Di, 18 Mär 2008) | 2 lines

  Move test_extcall to doctest.
........
  r61480 | brett.cannon | 2008-03-18 04:46:22 +0100 (Di, 18 Mär 2008) | 2 lines

  test_errno was a no-op test; now it actually tests things and uses unittest.
........
  r61483 | brett.cannon | 2008-03-18 05:09:00 +0100 (Di, 18 Mär 2008) | 3 lines

  Remove our implementation of memmove() and strerror(); both are in the C89
  standard library.
........
  r61484 | brett.cannon | 2008-03-18 05:16:06 +0100 (Di, 18 Mär 2008) | 2 lines

  The output directory for tests that compare against stdout is now gone!
........
  r61488 | jeffrey.yasskin | 2008-03-18 05:29:35 +0100 (Di, 18 Mär 2008) | 2 lines

  Block the "socket.ssl() is deprecated" warning from test_socket_ssl.
........
  r61495 | jeffrey.yasskin | 2008-03-18 05:56:06 +0100 (Di, 18 Mär 2008) | 4 lines

  Speed test_thread up from 51.328s to 0.081s by reducing its sleep times. We
  still sleep at all to make it likely that all threads are active at the same
  time.
........
  r61496 | jeffrey.yasskin | 2008-03-18 06:12:41 +0100 (Di, 18 Mär 2008) | 4 lines

  Speed up test_dict by about 10x by only checking selected dict literal sizes,
  instead of every integer from 0 to 400. Exhaustive testing wastes time without
  providing enough more assurance that the code is correct.
........
  r61498 | neal.norwitz | 2008-03-18 06:20:29 +0100 (Di, 18 Mär 2008) | 1 line

  Try increasing the timeout to reduce the flakiness of this test.
........
  r61503 | brett.cannon | 2008-03-18 06:43:04 +0100 (Di, 18 Mär 2008) | 2 lines

  Improve the error message for a test that failed on the S-390 Debian buildbot.
........
  r61504 | jeffrey.yasskin | 2008-03-18 06:45:40 +0100 (Di, 18 Mär 2008) | 3 lines

  Add a -S/--slow flag to regrtest to have it print the 10 slowest tests with
  their times.
........
  r61507 | neal.norwitz | 2008-03-18 07:03:46 +0100 (Di, 18 Mär 2008) | 1 line

  Add some info to the failure messages
........
  r61509 | trent.nelson | 2008-03-18 08:02:12 +0100 (Di, 18 Mär 2008) | 1 line

  Issue 2286: bump up the stack size of the 64-bit debug python_d.exe to 2100000.  The default value of 200000 causes a stack overflow at 1965 iterations of r_object() in marshal.c, 35 iterations before the 2000 limit enforced by MAX_MARSHAL_STACK_DEPTH.
........
  r61510 | trent.nelson | 2008-03-18 08:32:47 +0100 (Di, 18 Mär 2008) | 5 lines

  The behaviour of winsound.Beep() seems to differ between different versions of Windows when there's either:
      a) no sound card entirely
      b) legacy beep driver has been disabled
      c) the legacy beep driver has been uninstalled
  Sometimes RuntimeErrors are raised, sometimes they're not.  If _have_soundcard() returns False, don't expect winsound.Beep() to raise a RuntimeError, as this clearly isn't the case, as demonstrated by the various Win32 XP buildbots.
........
  r61515 | martin.v.loewis | 2008-03-18 13:20:15 +0100 (Di, 18 Mär 2008) | 2 lines

  norwitz-amd64 (gentoo) has EREMOTEIO.
........
  r61516 | martin.v.loewis | 2008-03-18 13:45:37 +0100 (Di, 18 Mär 2008) | 2 lines

  Add more Linux error codes.
........
  r61517 | martin.v.loewis | 2008-03-18 14:05:03 +0100 (Di, 18 Mär 2008) | 2 lines

  Add WSA errors.
........
  r61518 | martin.v.loewis | 2008-03-18 14:16:05 +0100 (Di, 18 Mär 2008) | 2 lines

  Note that the stderr output of the test is intentional.
........

3 files changed, 73 insertions, 69 deletions

diff --git a/Python/memmove.c b/Python/memmove.c
deleted file mode 100644
index 6fb0dad..0000000
--- a/Python/memmove.c
+++ /dev/null
@@ -1,25 +0,0 @@
-
-/* A perhaps slow but I hope correct implementation of memmove */
-
-extern char *memcpy(char *, char *, int);
-
-char *
-memmove(char *dst, char *src, int n)
-{
-	char *realdst = dst;
-	if (n <= 0)
-		return dst;
-	if (src >= dst+n || dst >= src+n)
-		return memcpy(dst, src, n);
-	if (src > dst) {
-		while (--n >= 0)
-			*dst++ = *src++;
-	}
-	else if (src < dst) {
-		src += n;
-		dst += n;
-		while (--n >= 0)
-			*--dst = *--src;
-	}
-	return realdst;
-}
diff --git a/Python/pystrtod.c b/Python/pystrtod.c
index 16efa9d..2ca8402 100644
--- a/Python/pystrtod.c
+++ b/Python/pystrtod.c
@@ -7,7 +7,6 @@
 #define ISSPACE(c)  ((c) == ' ' || (c) == '\f' || (c) == '\n' || \
                      (c) == '\r' || (c) == '\t' || (c) == '\v')
 #define ISDIGIT(c)  ((c) >= '0' && (c) <= '9')
-#define ISXDIGIT(c) (ISDIGIT(c) || ((c) >= 'a' && (c) <= 'f') || ((c) >= 'A' && (c) <= 'F'))
 
 
 /**
@@ -123,7 +122,8 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
 			errno = EINVAL;
 			return val;
 		}
-		/* For the other cases, we need not convert the decimal point */
+		/* For the other cases, we need not convert the decimal
+		   point */
 	}
 
 	/* Set errno to zero, so that we can distinguish zero results
@@ -134,7 +134,8 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
 	{
 		char *copy, *c;
 
-		/* We need to convert the '.' to the locale specific decimal point */
+		/* We need to convert the '.' to the locale specific decimal
+		   point */
 		copy = (char *)PyMem_MALLOC(end - digits_pos +
 					    1 + decimal_point_len);
 		if (copy == NULL) {
@@ -149,7 +150,8 @@ PyOS_ascii_strtod(const char *nptr, char **endptr)
 		c += decimal_point_pos - digits_pos;
 		memcpy(c, decimal_point, decimal_point_len);
 		c += decimal_point_len;
-		memcpy(c, decimal_point_pos + 1, end - (decimal_point_pos + 1));
+		memcpy(c, decimal_point_pos + 1,
+		       end - (decimal_point_pos + 1));
 		c += end - (decimal_point_pos + 1);
 		*c = 0;
 
@@ -198,7 +200,7 @@ as necessary to represent the exponent.
 /**
  * PyOS_ascii_formatd:
  * @buffer: A buffer to place the resulting string in
- * @buf_len: The length of the buffer.
+ * @buf_size: The length of the buffer.
  * @format: The printf()-style format to use for the
  *          code to use for converting. 
  * @d: The #gdouble to convert
@@ -209,12 +211,14 @@ as necessary to represent the exponent.
  * specifiers are 'e', 'E', 'f', 'F', 'g', 'G', and 'n'.
  * 
  * '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
+ *     at least one digit after the decimal.
  *
  * Return value: The pointer to the buffer with the converted string.
  **/
 char *
 PyOS_ascii_formatd(char       *buffer, 
-		   size_t      buf_len, 
+		   size_t      buf_size, 
 		   const char *format, 
 		   double      d)
 {
@@ -227,20 +231,13 @@ PyOS_ascii_formatd(char       *buffer,
 	   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
+	   also with at least one character past the decimal. */
 	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); */
-
 	/* 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' || */
-/* 			      format_char == 'g' || format_char == 'G', */
-/* 			      NULL); */
-
 	if (format[0] != '%')
 		return NULL;
 
@@ -251,19 +248,24 @@ PyOS_ascii_formatd(char       *buffer,
 	if (strpbrk(format + 1, "'l%"))
 		return NULL;
 
+	/* Also curious about this function is that it accepts format strings
+	   like "%xg", which are invalid for floats.  In general, the
+	   interface to this function is not very good, but changing it is
+	   difficult because it's a public API. */
+
 	if (!(format_char == 'e' || format_char == 'E' || 
 	      format_char == 'f' || format_char == 'F' || 
 	      format_char == 'g' || format_char == 'G' ||
-	      format_char == 'n'))
+	      format_char == 'n' || format_char == 'Z'))
 		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') {
+	/* Map 'n' or 'Z' format_char to 'g', by copying the format string and
+	   replacing the final char with a 'g' */
+	if (format_char == 'n' || format_char == 'Z') {
 		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 */
+			   practice, this will never happen and will be
+			   detected by returning NULL */
 			return NULL;
 		}
 		strcpy(tmp_format, format);
@@ -271,8 +273,9 @@ PyOS_ascii_formatd(char       *buffer,
 		format = tmp_format;
 	}
 
+
 	/* Have PyOS_snprintf do the hard work */
-	PyOS_snprintf(buffer, buf_len, format, d);
+	PyOS_snprintf(buffer, buf_size, format, d);
 
 	/* Get the current local, and find the decimal point character (or
 	   string?).  Convert that string back to a dot.  Do not do this if
@@ -294,7 +297,8 @@ PyOS_ascii_formatd(char       *buffer,
 			while (isdigit(Py_CHARMASK(*p)))
 				p++;
 
-			if (strncmp(p, decimal_point, decimal_point_len) == 0) {
+			if (strncmp(p, decimal_point,
+				    decimal_point_len) == 0) {
 				*p = '.';
 				p++;
 				if (decimal_point_len > 1) {
@@ -343,7 +347,8 @@ PyOS_ascii_formatd(char       *buffer,
 			   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;
+			extra_zeros_cnt = exponent_digit_cnt -
+				significant_digit_cnt;
 
 			/* Delete extra_zeros_cnt worth of characters from the
 			   front of the exponent */
@@ -360,7 +365,7 @@ PyOS_ascii_formatd(char       *buffer,
 			   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) {
+			      < buffer + buf_size) {
 				memmove(start + zeros, start,
 					exponent_digit_cnt + 1);
 				memset(start, '0', zeros);
@@ -368,6 +373,49 @@ PyOS_ascii_formatd(char       *buffer,
 		}
 	}
 
+	/* If format_char is 'Z', make sure we have at least one character
+	   after the decimal point (and make sure we have a decimal point). */
+	if (format_char == 'Z') {
+		int insert_count = 0;
+		char* chars_to_insert;
+
+		/* search for the first non-digit character */
+		p = buffer;
+		while (*p && isdigit(Py_CHARMASK(*p)))
+			++p;
+
+		if (*p == '.') {
+			if (isdigit(Py_CHARMASK(*(p+1)))) {
+				/* Nothing to do, we already have a decimal
+				   point and a digit after it */
+			}
+			else {
+				/* We have a decimal point, but no following
+				   digit.  Insert a zero after the decimal. */
+				++p;
+				chars_to_insert = "0";
+				insert_count = 1;
+			}
+		}
+		else {
+			chars_to_insert = ".0";
+			insert_count = 2;
+		}
+		if (insert_count) {
+			size_t buf_len = strlen(buffer);
+			if (buf_len + insert_count + 1 >= buf_size) {
+				/* If there is not enough room in the buffer
+				   for the additional text, just skip it.  It's
+				   not worth generating an error over. */
+			}
+			else {
+				memmove(p + insert_count, p,
+					buffer + strlen(buffer) - p + 1);
+				memcpy(p, chars_to_insert, insert_count);
+			}
+		}
+	}
+
 	return buffer;
 }
 
diff --git a/Python/strerror.c b/Python/strerror.c
deleted file mode 100644
index 55f8342..0000000
--- a/Python/strerror.c
+++ /dev/null
@@ -1,19 +0,0 @@
-
-/* PD implementation of strerror() for systems that don't have it.
-   Author: Guido van Rossum, CWI Amsterdam, Oct. 1990, <guido@cwi.nl>. */
-
-#include <stdio.h>
-#include "Python.h"
-
-extern int sys_nerr;
-extern char *sys_errlist[];
-
-char *
-strerror(int err)
-{
-	static char buf[20];
-	if (err >= 0 && err < sys_nerr)
-		return sys_errlist[err];
-	PyOS_snprintf(buf, sizeof(buf), "Unknown errno %d", err);
-	return buf;
-}