[Python-checkins] r41996 - in python/trunk: Misc/NEWS Objects/unicodeobject.c

[neal.norwitz]

Author: neal.norwitz
Date: Tue Jan 10 07:03:13 2006
New Revision: 41996

Modified:
   python/trunk/Misc/NEWS
   python/trunk/Objects/unicodeobject.c
Log:
- Patch #1400181, fix unicode string formatting to not use the locale.
  This is how string objects work.  u'%f' could use , instead of .
  for the decimal point.  Now both strings and unicode always use periods.

This is the code that would break:

import locale
locale.setlocale(locale.LC_NUMERIC, 'de_DE')
u'%.1f' % 1.0
assert '1.0' == u'%.1f' % 1.0

I couldn't create a test case which fails, but this fixes the problem.

Will backport.


Modified: python/trunk/Misc/NEWS
==============================================================================
--- python/trunk/Misc/NEWS	(original)
+++ python/trunk/Misc/NEWS	Tue Jan 10 07:03:13 2006
@@ -12,6 +12,10 @@
 Core and builtins
 -----------------
 
+- Patch #1400181, fix unicode string formatting to not use the locale.
+  This is how string objects work.  u'%f' could use , instead of .
+  for the decimal point.  Now both strings and unicode always use periods.
+
 - Bug #1244610, #1392915, fix build problem on OpenBSD 3.7 and 3.8.
   configure would break checking curses.h.
 

Modified: python/trunk/Objects/unicodeobject.c
==============================================================================
--- python/trunk/Objects/unicodeobject.c	(original)
+++ python/trunk/Objects/unicodeobject.c	Tue Jan 10 07:03:13 2006
@@ -6579,26 +6579,31 @@
 #define F_ALT	(1<<3)
 #define F_ZERO	(1<<4)
 
-static
-int usprintf(register Py_UNICODE *buffer, char *format, ...)
+static int
+strtounicode(Py_UNICODE *buffer, const char *charbuffer)
 {
-    register int i;
-    int len;
-    va_list va;
-    char *charbuffer;
-    va_start(va, format);
-
-    /* First, format the string as char array, then expand to Py_UNICODE
-       array. */
-    charbuffer = (char *)buffer;
-    len = vsprintf(charbuffer, format, va);
+    register long i;
+    long len = strlen(charbuffer);
     for (i = len - 1; i >= 0; i--)
 	buffer[i] = (Py_UNICODE) charbuffer[i];
 
-    va_end(va);
     return len;
 }
 
+static int
+doubletounicode(Py_UNICODE *buffer, size_t len, const char *format, double x)
+{
+    PyOS_ascii_formatd((char *)buffer, len, format, x);
+    return strtounicode(buffer, (char *)buffer);
+}
+
+static int
+longtounicode(Py_UNICODE *buffer, size_t len, const char *format, long x)
+{
+    PyOS_snprintf((char *)buffer, len, format, x);
+    return strtounicode(buffer, (char *)buffer);
+}
+
 /* XXX To save some code duplication, formatfloat/long/int could have been
    shared with stringobject.c, converting from 8-bit to Unicode after the
    formatting is done. */
@@ -6648,7 +6653,7 @@
     PyOS_snprintf(fmt, sizeof(fmt), "%%%s.%d%c",
 		  (flags&F_ALT) ? "#" : "",
 		  prec, type);
-    return usprintf(buf, fmt, x);
+    return doubletounicode(buf, buflen, fmt, x);
 }
 
 static PyObject*
@@ -6740,9 +6745,9 @@
                       prec, type);
     }
     if (sign[0])
-        return usprintf(buf, fmt, -x);
+        return longtounicode(buf, buflen, fmt, -x);
     else
-        return usprintf(buf, fmt, x);
+        return longtounicode(buf, buflen, fmt, x);
 }
 
 static int