[pypy-commit] pypy default: Fix this CPython test, and comment about why I think that PyPy's

[arigo]

Author: Armin Rigo <arigo at tunes.org>
Branch: 
Changeset: r49279:2f05272d1f77
Date: 2011-11-10 15:08 +0100
http://bitbucket.org/pypy/pypy/changeset/2f05272d1f77/

Log:	Fix this CPython test, and comment about why I think that PyPy's
	behavior is better (although it's all open to discussion of course)

diff --git a/lib-python/modified-2.7/test/test_repr.py b/lib-python/modified-2.7/test/test_repr.py
--- a/lib-python/modified-2.7/test/test_repr.py
+++ b/lib-python/modified-2.7/test/test_repr.py
@@ -254,8 +254,14 @@
         eq = self.assertEqual
         touch(os.path.join(self.subpkgname, self.pkgname + os.extsep + 'py'))
         from areallylongpackageandmodulenametotestreprtruncation.areallylongpackageandmodulenametotestreprtruncation import areallylongpackageandmodulenametotestreprtruncation
-        eq(repr(areallylongpackageandmodulenametotestreprtruncation),
-           "<module '%s' from '%s'>" % (areallylongpackageandmodulenametotestreprtruncation.__name__, areallylongpackageandmodulenametotestreprtruncation.__file__))
+        # On PyPy, we use %r to format the file name; on CPython it is done
+        # with '%s'.  It seems to me that %r is safer <arigo>.
+        if '__pypy__' in sys.builtin_module_names:
+            eq(repr(areallylongpackageandmodulenametotestreprtruncation),
+               "<module %r from %r>" % (areallylongpackageandmodulenametotestreprtruncation.__name__, areallylongpackageandmodulenametotestreprtruncation.__file__))
+        else:
+            eq(repr(areallylongpackageandmodulenametotestreprtruncation),
+               "<module '%s' from '%s'>" % (areallylongpackageandmodulenametotestreprtruncation.__name__, areallylongpackageandmodulenametotestreprtruncation.__file__))
         eq(repr(sys), "<module 'sys' (built-in)>")
 
     def test_type(self):