[MERGE/RFC] Better unicode symlinks handling

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


...


>     jam> +def load_tests(basic_tests, module, loader):
>     jam> +    suite = loader.suiteClass()
>     jam> +    dir_reader_tests, remaining_tests = tests.split_suite_by_condition(
>     jam> +        basic_tests, tests.condition_isinstance(TestCaseWithDirState))
>     jam> +    tests.multiply_tests(dir_reader_tests,
>     jam> +                         test_osutils.dir_reader_scenarios(), suite)
>     jam> +    suite.addTest(remaining_tests)
>     jam> +    return suite
>     jam> +
>     jam> +
> 
>     jam> ^- this sure looks like a second test multiplication over the
>     jam> same tests.
>     jam> (everything subclassing TestCaseWithDirState). Perhaps not, though.
> 
> I don't understand the 'second' here :-/
> 
> - multiply_tests parametrize TestCaseWithDirState tests against the dir readers.
> 


=== modified file 'bzrlib/tests/test__dirstate_helpers.py'

...
+def load_tests(basic_tests, module, loader):
...
+    dir_reader_tests, remaining_tests = tests.split_suite_by_condition(
+        remaining_tests, tests.condition_isinstance(
+            test_dirstate.TestCaseWithDirState))
+    tests.multiply_tests(dir_reader_tests, dir_reader_scenarios, suite)
+    suite.addTest(remaining_tests)
+
+    return suite
+
+

- - one

=== modified file 'bzrlib/tests/test_dirstate.py'
- --- bzrlib/tests/test_dirstate.py	2009-04-17 14:03:32 +0000
+++ bzrlib/tests/test_dirstate.py	2009-04-17 14:59:33 +0000
@@ -44,9 +44,39 @@
 # set_path_id  setting id when state is in memory modified


+def load_tests(basic_tests, module, loader):
+    suite = loader.suiteClass()
+    dir_reader_tests, remaining_tests = tests.split_suite_by_condition(
+        basic_tests, tests.condition_isinstance(TestCaseWithDirState))
+    tests.multiply_tests(dir_reader_tests,
+                         test_osutils.dir_reader_scenarios(), suite)
+    suite.addTest(remaining_tests)
+    return suite
+

- - two

I'm guessing they are multiplying different tests, and both sides are
using TestCaseWithDirstate as a convenient base. But the condition at
least looks identical.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkoAcUkACgkQJdeBCYSNAAP3awCdGh1WW5Q4Kvv4wx7CVd+uWvxG
w1EAoMHmBC+arn+QcVXUnPWYbZnXqPNG
=ccHB
-----END PGP SIGNATURE-----