Move all revision parsing code to _EnforceRevisions.

Move all revision parsing code to _EnforceRevisions.
Make -r invalid@rev a warning and ignore the argument.
Make -r rev use the Nth solution.
Fix a bug in FakeRepos.assertTree().

TEST=updated the smoke test to reflect the new behavior which should be the most compatible with the current buildbot behavior
BUG=chromium-os:3465
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=48607

[M-A Ruel, 2010-05-31 13:44:36]

Whoever comes first.

[Nicolas Sylvain, 2010-05-31 18:15:42]

lg, few nits below

http://codereview.chromium.org/2430001/diff/4001/5001
File gclient.py (right):

http://codereview.chromium.org/2430001/diff/4001/5001#newcode518
gclient.py:518: self._options.revisions.append('%s@%s' % (s['name'], rev))
not clear to me if --revision takes over safesync or not, but I think it should.

http://codereview.chromium.org/2430001/diff/4001/5001#newcode948
gclient.py:948: "Nth solution will be used. Can be used multiple "
i don't understand why you say "Nth".

http://codereview.chromium.org/2430001/diff/4001/5002
File tests/fake_repos.py (right):

http://codereview.chromium.org/2430001/diff/4001/5002#newcode11
tests/fake_repos.py:11: import pprint
why?

[M-A Ruel, 2010-05-31 18:40:04]

http://codereview.chromium.org/2430001/diff/4001/5001
File gclient.py (right):

http://codereview.chromium.org/2430001/diff/4001/5001#newcode518
gclient.py:518: self._options.revisions.append('%s@%s' % (s['name'], rev))
On 2010/05/31 18:15:42, Nicolas Sylvain wrote:
> not clear to me if --revision takes over safesync or not, but I think it
should.

Added a comment.

http://codereview.chromium.org/2430001/diff/4001/5001#newcode948
gclient.py:948: "Nth solution will be used. Can be used multiple "
On 2010/05/31 18:15:42, Nicolas Sylvain wrote:
> i don't understand why you say "Nth". 

Improved doc (hopefully)

http://codereview.chromium.org/2430001/diff/4001/5002
File tests/fake_repos.py (right):

http://codereview.chromium.org/2430001/diff/4001/5002#newcode11
tests/fake_repos.py:11: import pprint
On 2010/05/31 18:15:42, Nicolas Sylvain wrote:
> why?

Because it's used, I forgot to add the import when I moved code from
gclient_smoketest.py to this file but this code path was only triggered on test
failure so I didn't realize until I broke a test. :)

[Nicolas Sylvain, 2010-05-31 21:56:02]

lgtm