Split out the big chromite shell 'main.py' into lots of subfiles.

Split out the big chromite shell 'main.py' into lots of subfiles.

Main focus for review should be organization / naming.  Make sure
those are OK.  Since this is hard to follow, a high-level description
of things that happened here:

1. Almost everything moved out of 'main.py'.  The 'utils.py' became a
   bit of a dumping ground (for now) for functions needed by multiple
   modules.
2. All subcommands went into the 'subcmds' folder.  Mostly I have one
   command per file, except for the portage commands (which are all
   wrappers).
3. To make mox work better, we now import cros_build_lib directly (as
   cros_lib).  Now, we can mox out cros_lib.Die.
4. Cleaned up how the "--resume-state" gets handled for EnterChroot.
   It's now abstracted into its own function (so that it can be in the
   same file as EnterChroot).
5. Moved 'WrappedChrootCmd' into subcmd.py.
6. (Minor) Hack sys.argv[0] as 'chromite' so that help commands look
   better.

Change-Id: I19aad2a437864d364f4df70236ab84bfeba35376

BUG=chromium-os:10556
TEST=Ran unit tests; built mario; tried shell; tried workon command.

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=a9b1090

[diandersAtChromium, 2011-01-31 22:40:50]

PTAL.  See CL comments above for details.

[davidjames, 2011-02-02 01:15:18]

LGTM w/nits

http://codereview.chromium.org/6250058/diff/1/shell/main.py
File shell/main.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode39
shell/main.py:39: for cls in _COMMAND_HANDLERS      ]
No need for extra spaces at the end here.

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode104
shell/main.py:104: sys.argv[0] = 'chromite'
Can you add a TODO to remove this hack? We discussed a few solutions in person
that might be cleaner, so feel free to add your favorite one to the comment.

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode177
shell/main.py:177: cmd_cls = _COMMAND_HANDLERS[_COMMAND_STRS.index(cmd_str)]
Thank you! No eval please :)

http://codereview.chromium.org/6250058/diff/1/shell/utils.py
File shell/utils.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode374
shell/utils.py:374: sys.exit(0)
Can we move the sys.exit(0) out of here and into the caller? It's surprising to
see a utility function exit out of the program, so putting the logic in main()
will be more explicit and easier to understand.

[sosa, 2011-02-02 01:16:19]

Mostly nits

Your refactor got a lot of > 80 chars lines

http://codereview.chromium.org/6250058/diff/1/shell/main.py
File shell/main.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode16
shell/main.py:16: from chromite.lib import cros_build_lib as cros_lib
might be cleaner if you did:

import chromite.lib.cros_build_lib as cros_lib

http://codereview.chromium.org/6250058/diff/1/shell/subcmd.py
File shell/subcmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmd.py#newcode11
shell/subcmd.py:11: 
Don't need dbl lines here

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/build_cmd.py
File shell/subcmds/build_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/build_cmd.py#newc...
shell/subcmds/build_cmd.py:1: #!/usr/bin/python
Only directly shell-callable python should have this up top (i.e main /
unittests)

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/clean_cmd.py
File shell/subcmds/clean_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/clean_cmd.py#newc...
shell/subcmds/clean_cmd.py:1: #!/usr/bin/python
Only directly shell-callable python should have this up top (i.e main /
unittests)

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/portage_cmds.py
File shell/subcmds/portage_cmds.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/portage_cmds.py#n...
shell/subcmds/portage_cmds.py:2: #
same

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py
File shell/subcmds/shell_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py#newc...
shell/subcmds/shell_cmd.py:1: #!/usr/bin/python
same

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py#newc...
shell/subcmds/shell_cmd.py:125: cros_lib.RunCommand(argv, cwd=cwd, env=env,
error_ok=True, ignore_sigint=True)
80 char

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/workon_cmd.py
File shell/subcmds/workon_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/workon_cmd.py#new...
shell/subcmds/workon_cmd.py:1: #!/usr/bin/python
same

http://codereview.chromium.org/6250058/diff/1/shell/utils.py
File shell/utils.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode1
shell/utils.py:1: #!/usr/bin/python
same

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode194
shell/utils.py:194: cros_lib.Die("Couldn't find any matching %s specs for: %s" %
(spec_type, spec_name))
80 chars

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode305
shell/utils.py:305: cros_lib.Die('Chroot dir does not exist; try the "build
host" command.\n  %s.' %
80 chars

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode358
shell/utils.py:358: argv: The value of sys.argv
um why not just take sys.argv then?

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode374
shell/utils.py:374: sys.exit(0)
I don't really like this sys.exit here

[diandersAtChromium, 2011-02-02 01:49:10]

Got a verbal that sosa's review was 'LGTM w/ nits'.  I've fixed all nits and
will push.

http://codereview.chromium.org/6250058/diff/1/shell/main.py
File shell/main.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode16
shell/main.py:16: from chromite.lib import cros_build_lib as cros_lib
On 2011/02/02 01:16:19, sosa wrote:
> might be cleaner if you did:
> 
> import chromite.lib.cros_build_lib as cros_lib

Done.

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode39
shell/main.py:39: for cls in _COMMAND_HANDLERS      ]
On 2011/02/02 01:15:18, davidjames wrote:
> No need for extra spaces at the end here.

Done.

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode104
shell/main.py:104: sys.argv[0] = 'chromite'
On 2011/02/02 01:15:18, davidjames wrote:
> Can you add a TODO to remove this hack? We discussed a few solutions in person
> that might be cleaner, so feel free to add your favorite one to the comment.
> 

Done.

http://codereview.chromium.org/6250058/diff/1/shell/main.py#newcode177
shell/main.py:177: cmd_cls = _COMMAND_HANDLERS[_COMMAND_STRS.index(cmd_str)]
On 2011/02/02 01:15:18, davidjames wrote:
> Thank you! No eval please :)

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmd.py
File shell/subcmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmd.py#newcode11
shell/subcmd.py:11: 
On 2011/02/02 01:16:19, sosa wrote:
> Don't need dbl lines here

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/build_cmd.py
File shell/subcmds/build_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/build_cmd.py#newc...
shell/subcmds/build_cmd.py:1: #!/usr/bin/python
On 2011/02/02 01:16:19, sosa wrote:
> Only directly shell-callable python should have this up top (i.e main /
> unittests)

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/clean_cmd.py
File shell/subcmds/clean_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/clean_cmd.py#newc...
shell/subcmds/clean_cmd.py:1: #!/usr/bin/python
On 2011/02/02 01:16:19, sosa wrote:
> Only directly shell-callable python should have this up top (i.e main /
> unittests)

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/portage_cmds.py
File shell/subcmds/portage_cmds.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/portage_cmds.py#n...
shell/subcmds/portage_cmds.py:2: #
On 2011/02/02 01:16:19, sosa wrote:
> same

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py
File shell/subcmds/shell_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py#newc...
shell/subcmds/shell_cmd.py:1: #!/usr/bin/python
On 2011/02/02 01:16:19, sosa wrote:
> same

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/shell_cmd.py#newc...
shell/subcmds/shell_cmd.py:125: cros_lib.RunCommand(argv, cwd=cwd, env=env,
error_ok=True, ignore_sigint=True)
On 2011/02/02 01:16:19, sosa wrote:
> 80 char

Done.

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/workon_cmd.py
File shell/subcmds/workon_cmd.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/subcmds/workon_cmd.py#new...
shell/subcmds/workon_cmd.py:1: #!/usr/bin/python
On 2011/02/02 01:16:19, sosa wrote:
> same

Done.

http://codereview.chromium.org/6250058/diff/1/shell/utils.py
File shell/utils.py (right):

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode1
shell/utils.py:1: #!/usr/bin/python
On 2011/02/02 01:16:19, sosa wrote:
> same

Done.

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode194
shell/utils.py:194: cros_lib.Die("Couldn't find any matching %s specs for: %s" %
(spec_type, spec_name))
On 2011/02/02 01:16:19, sosa wrote:
> 80 chars

Done.

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode305
shell/utils.py:305: cros_lib.Die('Chroot dir does not exist; try the "build
host" command.\n  %s.' %
On 2011/02/02 01:16:19, sosa wrote:
> 80 chars

Done.

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode358
shell/utils.py:358: argv: The value of sys.argv
Wanted to be explicit that this works on sys.argv, so made caller pass it in.

On 2011/02/02 01:16:19, sosa wrote:
> um why not just take sys.argv then?

http://codereview.chromium.org/6250058/diff/1/shell/utils.py#newcode374
shell/utils.py:374: sys.exit(0)
OK, we now return True if we handled things.  It's up to main() to return then.