Add macros for asserting/expecting equality for ObjC objects.

This adds adds four macros when compiling using GTEST_OS_MAC:
{ASSERT,EXPECT}_NS{NE,EQ}.  These test ObjC objects using |-isEqual:| and
prints failures using the |-description| selector.  This allows for better
debugging output with ObjC++ test cases.

This also makes changes to the xcconfig files because linking to the Foundation
framework on 10.4 while supporting 10.5 and i386/x64 requires special-casing.

BUG=http://code.google.com/p/googletest/issues/detail?id=303
TEST=Covered by unit tests.

[Robert Sesek (Google), 2010-08-04 21:47:39]

Requesting review for this patch.  Thanks.

[Mark Mentovai, 2010-08-04 22:04:09]

http://codereview.appspot.com/1743056/diff/5001/6002
File include/gtest/gtest-mac.h (right):

http://codereview.appspot.com/1743056/diff/5001/6002#newcode71
include/gtest/gtest-mac.h:71: 
Extra blank line.

http://codereview.appspot.com/1743056/diff/5001/6003
File src/gtest-mac.mm (right):

http://codereview.appspot.com/1743056/diff/5001/6003#newcode73
src/gtest-mac.mm:73: msg << "Expected: (" << expected_expression << ") != (" <<
actual_expression
I’m surprised that there’s no NeFailure since there’s an EqFailure. Want to
introduce it?

http://codereview.appspot.com/1743056/diff/5001/6001
File test/gtest_mac_test.mm (right):

http://codereview.appspot.com/1743056/diff/5001/6001#newcode1
test/gtest_mac_test.mm:1: // Copyright 2008, Google Inc.
2010. You got it right in the other files.

http://codereview.appspot.com/1743056/diff/5001/6001#newcode45
test/gtest_mac_test.mm:45: EXPECT_NSEQ(@"a", @"a");
@"a" == @"a" too (pointer-wise, as opposed to -isEqual:), so this isn’t
absolutely the best possible test. You can keep this, but you should also try it
in a case where you have NSString* s1, s2; s1 ≠ s2; and [s1 isEqual:s2] == YES.

[Robert Sesek (Google), 2010-08-04 23:56:26]

I've addressed all comments.  I added ::testing::internal::NeFailure() but I
have not converted all other sites that use the same type of message.  Since
it's not a logical part of this CL, I'll do that in a followup CL.

http://codereview.appspot.com/1743056/diff/5001/6003
File src/gtest-mac.mm (right):

http://codereview.appspot.com/1743056/diff/5001/6003#newcode73
src/gtest-mac.mm:73: msg << "Expected: (" << expected_expression << ") != (" <<
actual_expression
On 2010/08/04 22:04:09, Mark Mentovai wrote:
> I’m surprised that there’s no NeFailure since there’s an EqFailure. Want to
> introduce it?

I've added this and use it in gtest-mac.mm, but I did not convert all the other
instances of this type of code to use it.  I'll do so in a follow-up CL.

http://codereview.appspot.com/1743056/diff/5001/6001
File test/gtest_mac_test.mm (right):

http://codereview.appspot.com/1743056/diff/5001/6001#newcode45
test/gtest_mac_test.mm:45: EXPECT_NSEQ(@"a", @"a");
On 2010/08/04 22:04:09, Mark Mentovai wrote:
> @"a" == @"a" too (pointer-wise, as opposed to -isEqual:), so this isn’t
> absolutely the best possible test. You can keep this, but you should also try
it
> in a case where you have NSString* s1, s2; s1 ≠&nbsp;s2; and [s1 isEqual:s2]
== YES.

I've added another set of expectations to this test that takes care of that. 
Nice use of Unicode ;).

[Mark Mentovai, 2010-08-05 00:31:24]

LGTM!

You should also get buy-in from a gtester. I can’t check this in.

http://codereview.appspot.com/1743056/diff/20001/21004
File include/gtest/internal/gtest-internal.h (right):

http://codereview.appspot.com/1743056/diff/20001/21004#newcode301
include/gtest/internal/gtest-internal.h:301: // where foo is 5 and bar is 5, we
have:
we? Know your reviewer!

http://codereview.appspot.com/1743056/diff/20001/21004#newcode303
include/gtest/internal/gtest-internal.h:303: //   Expected: (foo) != (bar),
actual: 5 vs 5
I wonder if it ever makes sense to print the actual values twice. The existing
stuff does, I think, right?

http://codereview.appspot.com/1743056/diff/20001/21006
File src/gtest-mac.mm (right):

http://codereview.appspot.com/1743056/diff/20001/21006#newcode52
src/gtest-mac.mm:52: if ([expected isEqual:actual]) {
Not enough indent in this function.