[pypy-commit] pypy refactor-buffer-api: unicodeobject has old buffer interface

[bdkearns]

Author: Brian Kearns <bdkearns at gmail.com>
Branch: refactor-buffer-api
Changeset: r70923:bd46658a05eb
Date: 2014-04-24 03:38 -0400
http://bitbucket.org/pypy/pypy/changeset/bd46658a05eb/

Log:	unicodeobject has old buffer interface

diff --git a/pypy/module/__pypy__/test/test_bytebuffer.py b/pypy/module/__pypy__/test/test_bytebuffer.py
--- a/pypy/module/__pypy__/test/test_bytebuffer.py
+++ b/pypy/module/__pypy__/test/test_bytebuffer.py
@@ -25,3 +25,5 @@
         assert str(b) == "\x00xy" + "\x00" * 7
         b[4:8:2] = 'zw'
         assert str(b) == "\x00xy\x00z\x00w" + "\x00" * 3
+        b[6:10] = u'#'
+        assert str(b) == "\x00xy\x00z\x00#" + "\x00" * 3
diff --git a/pypy/module/_io/test/test_bufferedio.py b/pypy/module/_io/test/test_bufferedio.py
--- a/pypy/module/_io/test/test_bufferedio.py
+++ b/pypy/module/_io/test/test_bufferedio.py
@@ -139,6 +139,8 @@
         raw = _io.FileIO(self.tmpfile)
         f = _io.BufferedReader(raw)
         assert f.readinto(a) == 5
+        exc = raises(TypeError, f.readinto, u"hello")
+        assert str(exc.value) == "cannot use unicode as modifiable buffer"
         exc = raises(TypeError, f.readinto, buffer(b"hello"))
         assert str(exc.value) == "must be read-write buffer, not buffer"
         exc = raises(TypeError, f.readinto, memoryview(b"hello"))
@@ -239,7 +241,8 @@
         import _io
         raw = _io.FileIO(self.tmpfile, 'w')
         f = _io.BufferedWriter(raw)
-        f.write("abcd")
+        f.write("ab")
+        f.write(u"cd")
         f.close()
         assert self.readfile() == "abcd"
 
diff --git a/pypy/module/_io/test/test_bytesio.py b/pypy/module/_io/test/test_bytesio.py
--- a/pypy/module/_io/test/test_bytesio.py
+++ b/pypy/module/_io/test/test_bytesio.py
@@ -38,6 +38,8 @@
         f = _io.BytesIO()
         assert f.write("") == 0
         assert f.write("hello") == 5
+        exc = raises(TypeError, f.write, u"lo")
+        assert str(exc.value) == "'unicode' does not have the buffer interface"
         import gc; gc.collect()
         assert f.getvalue() == "hello"
         f.close()
@@ -97,6 +99,8 @@
         a2 = bytearray('testing')
         assert b.readinto(a1) == 1
         assert b.readinto(a2) == 4
+        exc = raises(TypeError, b.readinto, u"hello")
+        assert str(exc.value) == "cannot use unicode as modifiable buffer"
         exc = raises(TypeError, b.readinto, buffer(b"hello"))
         assert str(exc.value) == "must be read-write buffer, not buffer"
         exc = raises(TypeError, b.readinto, memoryview(b"hello"))
diff --git a/pypy/module/_io/test/test_fileio.py b/pypy/module/_io/test/test_fileio.py
--- a/pypy/module/_io/test/test_fileio.py
+++ b/pypy/module/_io/test/test_fileio.py
@@ -82,7 +82,8 @@
         import _io
         filename = self.tmpfile + '_w'
         f = _io.FileIO(filename, 'wb')
-        f.write("test")
+        f.write("te")
+        f.write(u"st")
         # try without flushing
         f2 = _io.FileIO(filename, 'rb')
         assert f2.read() == "test"
@@ -135,6 +136,8 @@
         a = bytearray('x' * 10)
         f = _io.FileIO(self.tmpfile, 'r+')
         assert f.readinto(a) == 10
+        exc = raises(TypeError, f.readinto, u"hello")
+        assert str(exc.value) == "cannot use unicode as modifiable buffer"
         exc = raises(TypeError, f.readinto, buffer(b"hello"))
         assert str(exc.value) == "must be read-write buffer, not buffer"
         exc = raises(TypeError, f.readinto, memoryview(b"hello"))
diff --git a/pypy/objspace/std/bufferobject.py b/pypy/objspace/std/bufferobject.py
--- a/pypy/objspace/std/bufferobject.py
+++ b/pypy/objspace/std/bufferobject.py
@@ -37,18 +37,7 @@
     @staticmethod
     @unwrap_spec(offset=int, size=int)
     def descr_new_buffer(space, w_subtype, w_object, offset=0, size=-1):
-        if space.isinstance_w(w_object, space.w_unicode):
-            # unicode objects support the old buffer interface
-            # but not the new buffer interface (change in python 2.7)
-            from rpython.rlib.rstruct.unichar import pack_unichar, UNICODE_SIZE
-            unistr = space.unicode_w(w_object)
-            builder = StringBuilder(len(unistr) * UNICODE_SIZE)
-            for unich in unistr:
-                pack_unichar(unich, builder)
-            buf = StringBuffer(builder.build())
-        else:
-            buf = space.readbuf_w(w_object)
-
+        buf = space.readbuf_w(w_object)
         if offset == 0 and size == -1:
             return W_Buffer(buf)
         # handle buffer slices
diff --git a/pypy/objspace/std/unicodeobject.py b/pypy/objspace/std/unicodeobject.py
--- a/pypy/objspace/std/unicodeobject.py
+++ b/pypy/objspace/std/unicodeobject.py
@@ -2,7 +2,8 @@
 
 from rpython.rlib.objectmodel import (
     compute_hash, compute_unique_id, import_from_mixin)
-from rpython.rlib.rstring import UnicodeBuilder
+from rpython.rlib.buffer import StringBuffer
+from rpython.rlib.rstring import StringBuilder, UnicodeBuilder
 from rpython.rlib.runicode import (
     make_unicode_escape_function, str_decode_ascii, str_decode_utf_8,
     unicode_encode_ascii, unicode_encode_utf_8)
@@ -64,6 +65,17 @@
     def unicode_w(self, space):
         return self._value
 
+    def readbuf_w(self, space):
+        from rpython.rlib.rstruct.unichar import pack_unichar, UNICODE_SIZE
+        builder = StringBuilder(len(self._value) * UNICODE_SIZE)
+        for unich in self._value:
+            pack_unichar(unich, builder)
+        return StringBuffer(builder.build())
+
+    def writebuf_w(self, space):
+        raise OperationError(space.w_TypeError, space.wrap(
+            "cannot use unicode as modifiable buffer"))
+
     def listview_unicode(w_self):
         return _create_list_from_unicode(w_self._value)