[Python-3000-checkins] r58276 - in python/branches/py3k: Doc/library/stdtypes.rst Lib/test/string_tests.py Lib/test/test_descr.py Lib/test/test_unicode.py Objects/unicodeobject.c

[guido.van.rossum]

Author: guido.van.rossum
Date: Thu Sep 27 20:01:22 2007
New Revision: 58276

Modified:
   python/branches/py3k/Doc/library/stdtypes.rst
   python/branches/py3k/Lib/test/string_tests.py
   python/branches/py3k/Lib/test/test_descr.py
   python/branches/py3k/Lib/test/test_unicode.py
   python/branches/py3k/Objects/unicodeobject.c
Log:
Patch # 1145 by Thomas Lee:

str.join(...) now applies str() to the sequence elements if they're
not strings alraedy, except for bytes, which still raise TypeError
(for the same reasons why ""==b"" raises it).


Modified: python/branches/py3k/Doc/library/stdtypes.rst
==============================================================================
--- python/branches/py3k/Doc/library/stdtypes.rst	(original)
+++ python/branches/py3k/Doc/library/stdtypes.rst	Thu Sep 27 20:01:22 2007
@@ -789,8 +789,11 @@
 
 .. method:: str.join(seq)
 
-   Return a string which is the concatenation of the strings in the sequence *seq*.
-   The separator between elements is the string providing this method.
+   Return a string which is the concatenation of the values in the sequence
+   *seq*. Non-string values in *seq* will be converted to a string using their
+   respective ``str()`` value. If there are any :class:`bytes` objects in
+   *seq*, a :exc:`TypeError` will be raised. The separator between elements is
+   the string providing this method.
 
 
 .. method:: str.ljust(width[, fillchar])

Modified: python/branches/py3k/Lib/test/string_tests.py
==============================================================================
--- python/branches/py3k/Lib/test/string_tests.py	(original)
+++ python/branches/py3k/Lib/test/string_tests.py	Thu Sep 27 20:01:22 2007
@@ -13,6 +13,7 @@
 
 class BadSeq1(Sequence):
     def __init__(self): self.seq = [7, 'hello', 123]
+    def __str__(self): return '{0} {1} {2}'.format(*self.seq)
 
 class BadSeq2(Sequence):
     def __init__(self): self.seq = ['a', 'b', 'c']
@@ -987,19 +988,19 @@
         self.checkequal('abc', 'a', 'join', ('abc',))
         self.checkequal('z', 'a', 'join', UserList(['z']))
         self.checkequal('a.b.c', '.', 'join', ['a', 'b', 'c'])
-        self.checkraises(TypeError, '.', 'join', ['a', 'b', 3])
+        self.checkequal('a.b.3', '.', 'join', ['a', 'b', 3])
         for i in [5, 25, 125]:
             self.checkequal(((('a' * i) + '-') * i)[:-1], '-', 'join',
                  ['a' * i] * i)
             self.checkequal(((('a' * i) + '-') * i)[:-1], '-', 'join',
                  ('a' * i,) * i)
 
-        self.checkraises(TypeError, ' ', 'join', BadSeq1())
+        self.checkequal(str(BadSeq1()), ' ', 'join', BadSeq1())
         self.checkequal('a b c', ' ', 'join', BadSeq2())
 
         self.checkraises(TypeError, ' ', 'join')
         self.checkraises(TypeError, ' ', 'join', 7)
-        self.checkraises(TypeError, ' ', 'join', Sequence([7, 'hello', 123]))
+        self.checkraises(TypeError, ' ', 'join', [1, 2, bytes()])
         try:
             def f():
                 yield 4 + ""

Modified: python/branches/py3k/Lib/test/test_descr.py
==============================================================================
--- python/branches/py3k/Lib/test/test_descr.py	(original)
+++ python/branches/py3k/Lib/test/test_descr.py	Thu Sep 27 20:01:22 2007
@@ -3238,10 +3238,6 @@
     except ValueError: pass
     else: raise TestFailed("''.split('') doesn't raise ValueError")
 
-    try: ''.join([0])
-    except TypeError: pass
-    else: raise TestFailed("''.join([0]) doesn't raise TypeError")
-
     try: ''.rindex('5')
     except ValueError: pass
     else: raise TestFailed("''.rindex('5') doesn't raise ValueError")

Modified: python/branches/py3k/Lib/test/test_unicode.py
==============================================================================
--- python/branches/py3k/Lib/test/test_unicode.py	(original)
+++ python/branches/py3k/Lib/test/test_unicode.py	Thu Sep 27 20:01:22 2007
@@ -178,6 +178,10 @@
     def test_join(self):
         string_tests.MixinStrUnicodeUserStringTest.test_join(self)
 
+        class MyWrapper:
+            def __init__(self, sval): self.sval = sval
+            def __str__(self): return self.sval
+
         # mixed arguments
         self.checkequalnofix('a b c d', ' ', 'join', ['a', 'b', 'c', 'd'])
         self.checkequalnofix('abcd', '', 'join', ('a', 'b', 'c', 'd'))
@@ -186,6 +190,8 @@
         self.checkequalnofix('a b c d', ' ', 'join', ['a', 'b', 'c', 'd'])
         self.checkequalnofix('abcd', '', 'join', ('a', 'b', 'c', 'd'))
         self.checkequalnofix('w x y z', ' ', 'join', string_tests.Sequence('wxyz'))
+        self.checkequalnofix('1 2 foo', ' ', 'join', [1, 2, MyWrapper('foo')])
+        self.checkraises(TypeError, ' ', 'join', [1, 2, 3, bytes()])
 
     def test_replace(self):
         string_tests.CommonTest.test_replace(self)

Modified: python/branches/py3k/Objects/unicodeobject.c
==============================================================================
--- python/branches/py3k/Objects/unicodeobject.c	(original)
+++ python/branches/py3k/Objects/unicodeobject.c	Thu Sep 27 20:01:22 2007
@@ -5412,14 +5412,20 @@
 
 	item = PySequence_Fast_GET_ITEM(fseq, i);
 	/* Convert item to Unicode. */
-	if (! PyUnicode_Check(item) && ! PyString_Check(item)) {
-	    PyErr_Format(PyExc_TypeError,
-			 "sequence item %zd: expected string or Unicode,"
-			 " %.80s found",
-			 i, Py_Type(item)->tp_name);
-	    goto onError;
+	if (!PyString_Check(item) && !PyUnicode_Check(item))
+	{
+		if (PyBytes_Check(item))
+		{
+			PyErr_Format(PyExc_TypeError,
+                            "sequence item %d: join() will not operate on "
+                            "bytes objects", i);
+			goto onError;
+		}
+		item = PyObject_Unicode(item);
 	}
-	item = PyUnicode_FromObject(item);
+	else
+		item = PyUnicode_FromObject(item);
+
 	if (item == NULL)
 	    goto onError;
 	/* We own a reference to item from here on. */