Add git-map and git-short-map to depot_tools.

Add git-map and git-map-branches to depot_tools.

git-map: Show your local repo's history in a pseudo-graphical format from the command line.

git-map-branches: Show the topology of all of your branches, and their upstream relationships.

git-nav-upstream: Navigate (checkout) to the upstream branch of the current branch.

git-nav-downstream: Navigate (checkout) to a downstream branch of the current branch. If there's more than one downstream branch, then present a menu to select which one you want.

R=agable@chromium.org, hinoka@chromium.org, stip@chromium.org, szager@chromium.org
BUG=

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=256384

[ghost stip (do not use), 2014-02-28 03:29:01]

these are used instead of the native git log because of windows I assume?

[iannucci, 2014-02-28 03:34:17]

On 2014/02/28 03:29:01, stip wrote:
> these are used instead of the native git log because of windows I assume?

git-map is like your showmethemoney, except that it highlights local branches in
cyan, highlights the current value of HEAD by making a solid blue background,
and some other things.

git-short-map just shows the upstream relationship of all of your local branches
(color-coded of course), with a * next to the current branch.

[Ryan Tseng, 2014-02-28 08:03:54]

I don't think they got uploaded correctly.  You'll need to clear out the last
uploaded hash and upload again I think.

[iannucci, 2014-02-28 08:29:14]

Should be good now

[agable, 2014-02-28 19:54:57]

Since this CL has python_git_runner in it, I'll leave my top-level comments that
apply to all of the CLs here.

1) All of the commands need more documentation. There are literally zero
docstrings across all of these. The one-liner bash ones need comments too.
1a) Command line documentation would be even better. When git finds a
'git-foo-bar' executable, it also aliases 'git help foo-bar' to 'man foo-bar'.
If we could automatically install man pages for these that would be really
really useful.

2) We're making depot_tools pretty cluttered. All of the top-level git
executables should stay here, but I think all of the python backends should move
into depot_tools/gitscripts.
2a) This will also make the gitscripts git config namespace make more sense.

3) You should file a CL that updates all of the current git extensions we have
to use the new python launcher.
3a) Does the new launcher mean that the .bat files are also unnecessary now? Or
do you need to create .bat files for all of the new ones too?

https://codereview.chromium.org/184113002/diff/20001/git-map
File git-map (right):

https://codereview.chromium.org/184113002/diff/20001/git-map#newcode1
git-map:1: #!/bin/bash
Yes, it is really nice that all of these tiny little bash scripts are just the
same piece of boilerplate due to python_git_runner. But they should point at
their python scripts directly in a comment so it is easy for someone figuring
out what they do to find it.

https://codereview.chromium.org/184113002/diff/20001/git_common.py
File git_common.py (right):

https://codereview.chromium.org/184113002/diff/20001/git_common.py#newcode30
git_common.py:30: class DEFAULT(object):
Why? This is so completely unnecessary. Maybe kinda elegant, but generally
obtuse and irrelevant.

https://codereview.chromium.org/184113002/diff/20001/git_common.py#newcode227
git_common.py:227: def config_list(option, default=DEFAULT(list)):
default=None
if default is None: default=[]

Let's stick with the classics, here. Especially since nothing actually calls
this method with a default= kwarg.

https://codereview.chromium.org/184113002/diff/20001/git_common.py#newcode274
git_common.py:274: def hash_multi(*reflike):
This should be "*reflikes" (cf. "*args").

https://codereview.chromium.org/184113002/diff/20001/git_map.py
File git_map.py (right):

https://codereview.chromium.org/184113002/diff/20001/git_map.py#newcode1
git_map.py:1: #!/usr/bin/env python
Needs a module level docstring. All of these do, in fact.

https://codereview.chromium.org/184113002/diff/20001/git_short_map.py
File git_short_map.py (right):

https://codereview.chromium.org/184113002/diff/20001/git_short_map.py#newcode1
git_short_map.py:1: #!/usr/bin/env python
I'd call this 'git-map-branches', not 'git-short-map'. Semantically it is a
subcommand of git-map (--format=minimalist :P).

https://codereview.chromium.org/184113002/diff/20001/python_git_runner.sh
File python_git_runner.sh (right):

https://codereview.chromium.org/184113002/diff/20001/python_git_runner.sh#new...
python_git_runner.sh:1: #!/bin/bash
Make 'git cl' and 'git cache' use this.

https://codereview.chromium.org/184113002/diff/20001/python_git_runner.sh#new...
python_git_runner.sh:7: ## form 'git-sub-command'. This script will then instead
invoke 'git_sub_command.py'
80 chars? (It's wrapping on Rietveld but I'm too lazy to actually count.)

[iannucci, 2014-03-06 00:18:39]

rebase + ptal

https://chromiumcodereview.appspot.com/184113002/diff/20001/git-map
File git-map (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/git-map#newcode1
git-map:1: #!/bin/bash
On 2014/02/28 19:54:58, agable wrote:
> Yes, it is really nice that all of these tiny little bash scripts are just the
> same piece of boilerplate due to python_git_runner. But they should point at
> their python scripts directly in a comment so it is easy for someone figuring
> out what they do to find it.

Really? What if I rename `python_git_runner.sh` ->
`run_python_git_wrapper_named_the_same_as_argv_0.sh`

It's obvious that this script does only one thing, which is forward to
python_git_runner. The docs in that script should make it obvious what's going
on, right?

I'll introduce this launcher refactor in a parent CL.

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py
File git_common.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py#new...
git_common.py:30: class DEFAULT(object):
On 2014/02/28 19:54:58, agable wrote:
> Why? This is so completely unnecessary. Maybe kinda elegant, but generally
> obtuse and irrelevant.

*sigh* going back to useless-mode. lcd for the win :(

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py#new...
git_common.py:227: def config_list(option, default=DEFAULT(list)):
On 2014/02/28 19:54:58, agable wrote:
> default=None
> if default is None: default=[]
> 
> Let's stick with the classics, here. Especially since nothing actually calls
> this method with a default= kwarg.

sometimes I hate python. The =None idiom is so uninformative :(.

I just got rid of default for the time being.

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py#new...
git_common.py:274: def hash_multi(*reflike):
On 2014/02/28 19:54:58, agable wrote:
> This should be "*reflikes" (cf. "*args").

Done.

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_map.py
File git_map.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_map.py#newcode1
git_map.py:1: #!/usr/bin/env python
On 2014/02/28 19:54:58, agable wrote:
> Needs a module level docstring. All of these do, in fact.

Right. Done.

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_short_map.py
File git_short_map.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_short_map.py#...
git_short_map.py:1: #!/usr/bin/env python
On 2014/02/28 19:54:58, agable wrote:
> I'd call this 'git-map-branches', not 'git-short-map'. Semantically it is a
> subcommand of git-map (--format=minimalist :P).

Done

https://chromiumcodereview.appspot.com/184113002/diff/20001/python_git_runner.sh
File python_git_runner.sh (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/python_git_runner...
python_git_runner.sh:7: ## form 'git-sub-command'. This script will then instead
invoke 'git_sub_command.py'
On 2014/02/28 19:54:58, agable wrote:
> 80 chars? (It's wrapping on Rietveld but I'm too lazy to actually count.)

Oh, weird. I think this is my bad. Done.

[agable, 2014-03-06 18:09:07]

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py
File git_common.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py#new...
git_common.py:227: def config_list(option, default=DEFAULT(list)):
On 2014/03/06 00:18:39, iannucci wrote:
> On 2014/02/28 19:54:58, agable wrote:
> > default=None
> > if default is None: default=[]
> > 
> > Let's stick with the classics, here. Especially since nothing actually calls
> > this method with a default= kwarg.
> 
> sometimes I hate python. The =None idiom is so uninformative :(.
> 
> I just got rid of default for the time being.

I know, and I totally agree. But I don't think this is the place to get fancy
and introduce new non-idioms.

https://chromiumcodereview.appspot.com/184113002/diff/40001/git-map
File git-map (right):

https://chromiumcodereview.appspot.com/184113002/diff/40001/git-map#newcode5
git-map:5: 
These don't have handy dandy pretty comments like the ones in
https://codereview.chromium.org/188383002/ do.

https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map.py
File git_map.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map.py#newcode1
git_map.py:1: #!/usr/bin/env python
Also move these into gitscripts/

https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map_branches.py
File git_map_branches.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map_branches....
git_map_branches.py:1: #!/usr/bin/env python
Update name of the CL to reflect new name of this file.

https://chromiumcodereview.appspot.com/184113002/diff/40001/python_git_runner.sh
File python_git_runner.sh (right):

https://chromiumcodereview.appspot.com/184113002/diff/40001/python_git_runner...
python_git_runner.sh:1: #!/bin/bash
This will vanish when this CL gets rebased on top of
https://codereview.chromium.org/188383002/ presumably.

[agable, 2014-03-07 21:41:19]

On 2014/03/06 18:09:07, agable wrote:
> https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py
> File git_common.py (right):
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/20001/git_common.py#new...
> git_common.py:227: def config_list(option, default=DEFAULT(list)):
> On 2014/03/06 00:18:39, iannucci wrote:
> > On 2014/02/28 19:54:58, agable wrote:
> > > default=None
> > > if default is None: default=[]
> > > 
> > > Let's stick with the classics, here. Especially since nothing actually
calls
> > > this method with a default= kwarg.
> > 
> > sometimes I hate python. The =None idiom is so uninformative :(.
> > 
> > I just got rid of default for the time being.
> 
> I know, and I totally agree. But I don't think this is the place to get fancy
> and introduce new non-idioms.
> 
> https://chromiumcodereview.appspot.com/184113002/diff/40001/git-map
> File git-map (right):
> 
> https://chromiumcodereview.appspot.com/184113002/diff/40001/git-map#newcode5
> git-map:5: 
> These don't have handy dandy pretty comments like the ones in
> https://codereview.chromium.org/188383002/ do.
> 
> https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map.py
> File git_map.py (right):
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map.py#newcode1
> git_map.py:1: #!/usr/bin/env python
> Also move these into gitscripts/
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map_branches.py
> File git_map_branches.py (right):
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/40001/git_map_branches....
> git_map_branches.py:1: #!/usr/bin/env python
> Update name of the CL to reflect new name of this file.
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/40001/python_git_runner.sh
> File python_git_runner.sh (right):
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/40001/python_git_runner...
> python_git_runner.sh:1: #!/bin/bash
> This will vanish when this CL gets rebased on top of
> https://codereview.chromium.org/188383002/ presumably.

I am seriously so sad that these don't work in gitscripts. LGTM.

[iannucci, 2014-03-12 00:12:21]

OK, ready for final review. Rebased onto HEAD, works on windows, tests added for
git_common changes.

https://chromiumcodereview.appspot.com/184113002/diff/210001/git_map.py
File git_map.py (right):

https://chromiumcodereview.appspot.com/184113002/diff/210001/git_map.py#newco...
git_map.py:31: 
oddly, this seems to be 'smart' and only emit colors if it thinks it's going to
a tty. However, less.exe seems to be able to translate the raw ANSI color codes
into cmd.exe compatible codes by itself.

[agable, 2014-03-12 00:46:15]

On 2014/03/12 00:12:21, iannucci wrote:
> OK, ready for final review. Rebased onto HEAD, works on windows, tests added
for
> git_common changes.
> 
> https://chromiumcodereview.appspot.com/184113002/diff/210001/git_map.py
> File git_map.py (right):
> 
>
https://chromiumcodereview.appspot.com/184113002/diff/210001/git_map.py#newco...
> git_map.py:31: 
> oddly, this seems to be 'smart' and only emit colors if it thinks it's going
to
> a tty. However, less.exe seems to be able to translate the raw ANSI color
codes
> into cmd.exe compatible codes by itself.

re-lgtm

[iannucci, 2014-03-12 00:50:35]

The CQ bit was checked by iannucci@chromium.org

[commit-bot: I haz the power, 2014-03-12 00:52:07]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/iannucci@chromium.org/184113002/250001

[commit-bot: I haz the power, 2014-03-12 01:07:47]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-12 01:07:48]

Presubmit check for 184113002-250001 failed and returned exit status -2001.
The presubmit check was hung. It took 900.0 seconds to execute and the time
limit is 900.0 seconds.

Running presubmit commit checks ...

[iannucci, 2014-03-12 01:13:43]

The CQ bit was checked by iannucci@chromium.org

[commit-bot: I haz the power, 2014-03-12 01:31:48]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/iannucci@chromium.org/184113002/250001

[commit-bot: I haz the power, 2014-03-12 01:36:19]

Change committed as 256384