On Jul 23, 2012, at 17:34, Luis Lavena wrote:
> On Mon, Jul 23, 2012 at 9:30 PM, Eric Hodel <drbrain / segment7.net> =
wrote:
>>=20
>> On Jul 23, 2012, at 16:32, luislavena / ruby-lang.org wrote:
>>> luislavena    2012-07-24 08:32:29 +0900 (Tue, 24 Jul 2012)
>>>=20
>>> New Revision: 36509
>>>=20
>>> http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=3Drev&revision=3D3650=
9
>>=20
>> This change causes a failure on OS X 10.7.4:
>>=20
>> [12/13] TestDir_M17N#test_filename_utf8_raw_windows_1252_name =3D =
0.02 s
>>  1) Failure:
>> test_filename_utf8_raw_windows_1252_name(TestDir_M17N) =
[/Users/drbrain/Work/svn/ruby/trunk/test/ruby/test_dir_m17n.rb:23]:
>> character 199 in UTF-8 (#<Process::Status: pid 59046 exit 1>)
>=20
> Hmn, this test passed on my Snow Leopard box and also Ubuntu 12.04
>=20
>> When I run the script by hand I get a filename of "=C7" (That's =
U+0043 latin capital letter C, U+0327 combining cedilla). I'm unsure =
what the expected result is.
>=20
> I think Windows-1252 and Windows-1251 tests can be skipped on
> non-Windows platforms, but is weird because it worked on this side of
> the world.

Odd, today it passes without issue (and without my patch applied), and I =
asked Aaron to reproduce it as well, and he has no issue. =20

Perhaps the commits by usa fixed the issue.

>> PS: I changed the assertions slightly to be more informative about =
failures:
>>=20
>=20
> Can you tell me what the assertions say on your side?

Before my failure says:

#<Process::Status: pid 59046 exit 1>

Which is not very useful.  After my failure says:

character 199 in UTF-8 (#<Process::Status: pid 59046 exit 1>)

(which is 0xc7)=