[pypy-commit] pypy rpython-unicode-formatting: correctly support u'%s' % my_unicode_string

[antocuni]

Author: Antonio Cuni <anto.c...@gmail.com>
Branch: rpython-unicode-formatting
Changeset: r56135:02b40b853dc6
Date: 2012-07-18 14:09 +0200
http://bitbucket.org/pypy/pypy/changeset/02b40b853dc6/

Log:    correctly support u'%s' % my_unicode_string

diff --git a/pypy/rpython/lltypesystem/rstr.py 
b/pypy/rpython/lltypesystem/rstr.py
--- a/pypy/rpython/lltypesystem/rstr.py
+++ b/pypy/rpython/lltypesystem/rstr.py
@@ -170,6 +170,13 @@
         return result
 
     @jit.elidable
+    def ll_unicode(self, s):
+        if s:
+            return s
+        else:
+            return self.convert_const(u'None')
+
+    @jit.elidable
     def ll_encode_latin1(self, s):
         length = len(s.chars)
         result = mallocstr(length)
@@ -985,8 +992,13 @@
                 vitem, r_arg = argsiter.next()
                 if not hasattr(r_arg, 'll_str'):
                     raise TyperError("ll_str unsupported for: %r" % r_arg)
-                if code == 's' or (code == 'r' and isinstance(r_arg, 
InstanceRepr)):
-                    # XXX: if it's unicode we don't want to call ll_str
+                if code == 's':
+                    if is_unicode:
+                        # only UniCharRepr and UnicodeRepr has it so far
+                        vchunk = hop.gendirectcall(r_arg.ll_unicode, vitem)
+                    else:
+                        vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
+                elif code == 'r' and isinstance(r_arg, InstanceRepr):
                     vchunk = hop.gendirectcall(r_arg.ll_str, vitem)
                 elif code == 'd':
                     assert isinstance(r_arg, IntegerRepr)
diff --git a/pypy/rpython/rstr.py b/pypy/rpython/rstr.py
--- a/pypy/rpython/rstr.py
+++ b/pypy/rpython/rstr.py
@@ -483,6 +483,8 @@
         # xxx suboptimal, maybe
         return str(unicode(ch))
 
+    def ll_unicode(self, ch):
+        return unicode(ch)
 
 class __extend__(AbstractCharRepr,
                  AbstractUniCharRepr):
diff --git a/pypy/rpython/test/test_runicode.py 
b/pypy/rpython/test/test_runicode.py
--- a/pypy/rpython/test/test_runicode.py
+++ b/pypy/rpython/test/test_runicode.py
@@ -1,3 +1,4 @@
+# -*- encoding: utf-8 -*-
 
 from pypy.rpython.lltypesystem.lltype import malloc
 from pypy.rpython.lltypesystem.rstr import LLHelpers, UNICODE
@@ -194,6 +195,15 @@
         assert self.interpret(fn, [u'(']) == False
         assert self.interpret(fn, [u'\u1058']) == False
         assert self.interpret(fn, [u'X']) == True
+
+    def test_strformat_unicode_arg(self):
+        const = self.const
+        def percentS(s):
+            return const("before %s after") % (s,)
+
+        res = self.interpret(percentS, [const(u'&#224;')])
+        assert self.ll_to_string(res) == const(u'before &#224; after')
+        
     
     def unsupported(self):
         py.test.skip("not supported")
_______________________________________________
pypy-commit mailing list
pypy-commit@python.org
http://mail.python.org/mailman/listinfo/pypy-commit