1) Add a new MULTIPROCESS_TEST_MAIN macro to store child process names...

1) Add a new MULTIPROCESS_TEST_MAIN macro to store child process names
in a lookup table.

Previously we were using different mechanisms on each platform to look up
child process names at runtime.  This broke on OS X where we strip the symbol
table on release executables.

2) Enable process_util_unittest on OS X.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=4165

[jeremy, 2008-10-28 23:59:18]

Needed to kill the previous review because for some reason it wasn't seeing new
patches.

In this version I've moved the files around as we discussed, this should fix the
dependency issues I ran into previously.

[Mark Mentovai, 2008-10-29 13:38:13]

LGTM

http://codereview.chromium.org/8864/diff/15/26
File testing/multiprocess_func_list.h (right):

http://codereview.chromium.org/8864/diff/15/26#newcode1
Line 1: // Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
Just 2008 on new files.  Same on the other addition.

http://codereview.chromium.org/8864/diff/15/26#newcode5
Line 5: #ifndef TESTING_MULTIPROCESS_FUNC_LIST_H__
New style is to use just one trailing _.

http://codereview.chromium.org/8864/diff/15/26#newcode9
Line 9: #include <wchar.h>
What's this for?  I don't see anything in here that would require it.

http://codereview.chromium.org/8864/diff/15/26#newcode11
Line 11: namespace MultiProcessFunctionList {
Style sez namespaces_should_be_named_this_way AndNotThisWay.

http://codereview.chromium.org/8864/diff/15/26#newcode30
Line 30: // Example usage:
I think the bit with the example usage should move to the top of the header,
with a brief overview of what this all does and why.

http://codereview.chromium.org/8864/diff/15/26#newcode37
Line 37: // who's constructor does the work of adding the global mapping.
whose constructor

http://codereview.chromium.org/8864/diff/15/26#newcode46
Line 46: 
Nuke extra lines