Add a method for normalizing path separators on Windows.

Add a method for normalizing path separators on Windows.

TEST=new unittests

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=50822

[tony, 2010-06-24 08:19:43]

I considered making FilePath::operator== just do the right thing, but there's a
test case that expects foo\bar != foo/bar on windows.  Is this intentional or
should I just fix operator== for Windows?

[Mark Mentovai, 2010-06-24 14:51:31]

LG

http://codereview.chromium.org/2831029/diff/2001/3002
File base/file_path.h (right):

http://codereview.chromium.org/2831029/diff/2001/3002#newcode307
base/file_path.h:307: #if defined(FILE_PATH_USES_WIN_SEPARATORS)
I don’t know where you’re planning on using this. Is it worth defining it on
non-Windows as a simple identity operation?

http://codereview.chromium.org/2831029/diff/2001/3003
File base/file_path_unittest.cc (right):

http://codereview.chromium.org/2831029/diff/2001/3003#newcode664
base/file_path_unittest.cc:664: { { FPL("\\foo/bar"),     FPL("\\foo\\bar") },  
     false},
Oops.

http://codereview.chromium.org/2831029/diff/2001/3003#newcode1015
base/file_path_unittest.cc:1015: { FPL("foo"), FPL("foo") },
Give me a couple cases with leading and trailing \s and /s too.

[Mark Mentovai, 2010-06-24 14:52:12]

I would leave operator== alone.

[brettw, 2010-06-24 18:56:13]

Would "NormalizeWindowsPathSeparators" be a better name?

[tony, 2010-06-25 01:35:11]

I renamed the method to NormalizeWindowsPathSeparators.  I'll probably go ahead
and land this.  Let me know if you think I should add the method to posix as
well.

http://codereview.chromium.org/2831029/diff/2001/3002
File base/file_path.h (right):

http://codereview.chromium.org/2831029/diff/2001/3002#newcode307
base/file_path.h:307: #if defined(FILE_PATH_USES_WIN_SEPARATORS)
On 2010/06/24 14:51:31, Mark Mentovai wrote:
> I don’t know where you’re planning on using this. Is it worth defining it on
> non-Windows as a simple identity operation?

I considered it, but figured it would be better to just put the code that uses
it behind an #ifdef.  It seems silly to have to copy the path on Mac/Linux
(although I guess the compiler might optimize that out).  Anyway, I don't feel
strongly either way about this.

http://codereview.chromium.org/2831029/diff/2001/3003
File base/file_path_unittest.cc (right):

http://codereview.chromium.org/2831029/diff/2001/3003#newcode1015
base/file_path_unittest.cc:1015: { FPL("foo"), FPL("foo") },
On 2010/06/24 14:51:31, Mark Mentovai wrote:
> Give me a couple cases with leading and trailing \s and /s too.

Done.

[Mark Mentovai, 2010-06-25 01:50:09]

LGTM with a couple more test cases.

http://codereview.chromium.org/2831029/diff/11001/12003
File base/file_path_unittest.cc (right):

http://codereview.chromium.org/2831029/diff/11001/12003#newcode1022
base/file_path_unittest.cc:1022: { FPL("foo\\bar\\"), FPL("foo\\bar") },
You added trailing slashes, but none with a leading slash.

[tony, 2010-06-25 03:02:59]

I added a bunch of tests including:
- leading slashes
- double slashes
- mixing adjacent slashes

[Mark Mentovai, 2010-06-25 03:24:00]

Hot tests. LGTM.

[brettw, 2010-06-25 03:52:49]

LGTM