Add --root flag used for deriving header guard CPP variable.

Add --root flag used for deriving header guard CPP variable. 

By default, the header guard CPP variable is calculated as the relative 
path to the directory that contains .git, .hg, or .svn. When this flag 
is specified, the relative path is calculated from the specified 
directory.

[mazda, 2013-04-12 07:02:10]

Hello Elliot,

Could you review this CL?

[erg, 2013-04-12 19:17:52]

I patched this in to my local client, and the unit tests don't pass.

======================================================================
FAIL: testBuildHeaderGuard (__main__.CpplintTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "./cpplint_unittest.py", line 2419, in testBuildHeaderGuard
    self.assertEqual('CPPLINT_CPPLINT_TEST_HEADER_H_', expected_guard)
AssertionError: 'CPPLINT_CPPLINT_TEST_HEADER_H_' !=
'_USR_LOCAL_GOOGLE_HOME_ERG_GOOGLE_STYLEGUIDE_CPPLINT_CPPLINT_CPPLINT_TEST_HEADER_H_'

======================================================================
FAIL: testBuildHeaderGuardWithRoot (__main__.CpplintTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "./cpplint_unittest.py", line 2596, in testBuildHeaderGuardWithRoot
    cpplint.GetHeaderGuardCPPVariable(file_path))
AssertionError: 'CPPLINT_CPPLINT_TEST_HEADER_H_' !=
'_USR_LOCAL_GOOGLE_HOME_ERG_GOOGLE_STYLEGUIDE_CPPLINT_CPPLINT_CPPLINT_TEST_HEADER_H_'

[mazda, 2013-04-15 04:39:53]

Does your local client have .svn directory?
The new test verifies the behavior of a function that relies on the existence of
.svn directory, so it will fail without .svn directory.
Should I avoid this assumption?

[erg, 2013-04-15 17:34:48]

On 2013/04/15 04:39:53, mazda wrote:
> Does your local client have .svn directory?
> The new test verifies the behavior of a function that relies on the existence
of
> .svn directory, so it will fail without .svn directory.
> Should I avoid this assumption?

There is an .svn directory in the cpplint folder, but some people do check this
out w/ git-svn, so I don't think you should make that assumption.

Also, it looks like it works if the unittests are run from the top of the
repository. I'd like to see that made more robust.

[mazda, 2013-04-16 09:24:56]

On 2013/04/15 17:34:48, erg wrote:
> On 2013/04/15 04:39:53, mazda wrote:
> > Does your local client have .svn directory?
> > The new test verifies the behavior of a function that relies on the
existence
> of
> > .svn directory, so it will fail without .svn directory.
> > Should I avoid this assumption?
> 
> There is an .svn directory in the cpplint folder, but some people do check
this
> out w/ git-svn, so I don't think you should make that assumption.
> 
> Also, it looks like it works if the unittests are run from the top of the
> repository. I'd like to see that made more robust.

Thank you for the comments.
I fixed the tests to work when it runs from the directories other than top of
the repository.
FileInfo.RepositoryName in cpplint.py checks the existence of .svn, .git, or .hg
to determine the top directory of the repository, so git-svn is OK as well. 'svn
export' and 'git archive' are problematic cases. I modified the test at least
not to fail in such cases (just skip the test).
Please take another look.

[erg, 2013-04-16 21:05:59]

On 2013/04/16 09:24:56, mazda wrote:
> On 2013/04/15 17:34:48, erg wrote:
> > On 2013/04/15 04:39:53, mazda wrote:
> > > Does your local client have .svn directory?
> > > The new test verifies the behavior of a function that relies on the
> existence
> > of
> > > .svn directory, so it will fail without .svn directory.
> > > Should I avoid this assumption?
> > 
> > There is an .svn directory in the cpplint folder, but some people do check
> this
> > out w/ git-svn, so I don't think you should make that assumption.
> > 
> > Also, it looks like it works if the unittests are run from the top of the
> > repository. I'd like to see that made more robust.
> 
> Thank you for the comments.
> I fixed the tests to work when it runs from the directories other than top of
> the repository.
> FileInfo.RepositoryName in cpplint.py checks the existence of .svn, .git, or
.hg
> to determine the top directory of the repository, so git-svn is OK as well.
'svn
> export' and 'git archive' are problematic cases. I modified the test at least
> not to fail in such cases (just skip the test).
> Please take another look.

lgtm. I'll commit this. Thanks.

[erg, 2013-04-16 21:07:24]

On 2013/04/16 21:05:59, erg wrote:
> On 2013/04/16 09:24:56, mazda wrote:
> > On 2013/04/15 17:34:48, erg wrote:
> > > On 2013/04/15 04:39:53, mazda wrote:
> > > > Does your local client have .svn directory?
> > > > The new test verifies the behavior of a function that relies on the
> > existence
> > > of
> > > > .svn directory, so it will fail without .svn directory.
> > > > Should I avoid this assumption?
> > > 
> > > There is an .svn directory in the cpplint folder, but some people do check
> > this
> > > out w/ git-svn, so I don't think you should make that assumption.
> > > 
> > > Also, it looks like it works if the unittests are run from the top of the
> > > repository. I'd like to see that made more robust.
> > 
> > Thank you for the comments.
> > I fixed the tests to work when it runs from the directories other than top
of
> > the repository.
> > FileInfo.RepositoryName in cpplint.py checks the existence of .svn, .git, or
> .hg
> > to determine the top directory of the repository, so git-svn is OK as well.
> 'svn
> > export' and 'git archive' are problematic cases. I modified the test at
least
> > not to fail in such cases (just skip the test).
> > Please take another look.
> 
> lgtm. I'll commit this. Thanks.

https://code.google.com/p/google-styleguide/source/detail?r=103