Convert setup.py to use setuptools

Convert setup.py to use setuptools

Due to r1667, setup.py was left broken installing 'gyp' as a script, which happened to be the *nix invocation script wrapper.

By converting setup.py to use setuptools, when the gyp script is installed, the appropriate file name will be generated to match the local conventions on both *nix and Windows platforms.

This change required adding a proxy function in gyp/__init__.py, called 'script_main', which takes 0 arguments as required by setuptools that directly ends up calling 'main' passing along sys.argv[1:].  Other areas in gyp/__init__.py and gyp_main.py were modified to use this new function as well.

The MANIFEST file was deleted since setuptools automatically includes the correct files in source distributions.

Contributed by ericvw@gmail.com

BUG=None

[Shezan Baig (Bloomberg), 2013-08-26 13:38:24]

I tested this locally and it worked.  After agl's review, I can commit this on
Eric's behalf, since he doesn't have commit access yet.

[agl, 2013-08-26 14:16:03]

This seems reasonable, but there's no way that I'm the correct reviewer for
this. You probably want someone who works on gyp.

[Shezan Baig (Bloomberg), 2013-08-26 14:21:49]

On 2013/08/26 14:16:03, agl wrote:
> This seems reasonable, but there's no way that I'm the correct reviewer for
> this. You probably want someone who works on gyp.

Ahh sorry, I suggested Eric add you because I saw you committed the file
originally:
https://code.google.com/p/gyp/source/detail?r=767

But when looking at the codereview for that commit, it looks like sgk reviewed
it.  Will add sgk to review instead.  Thanks for looking at this!

[Shezan Baig (Bloomberg), 2013-08-26 14:22:22]

Switching reviewer to sgk

[ericvw, 2013-08-26 16:25:12]

Uploaded a new patch that removed 'scripts' from setup.py

[Shezan Baig (Bloomberg), 2013-09-04 14:16:50]

Hey Scott, can you take a look at this when you have time? Thanks!

[scottmg, 2013-09-04 15:48:52]

On 2013/09/04 14:16:50, Shezan Baig (Bloomberg) wrote:
> Hey Scott, can you take a look at this when you have time? Thanks!

I'm inclined to just delete setup.py, tbh. Chrome (and other projects) just
clone-and-run.

I don't think it's worth adding ez_setup.py at the very least.

[ericvw, 2013-09-04 16:16:12]

On 2013/09/04 15:48:52, scottmg wrote:
> On 2013/09/04 14:16:50, Shezan Baig (Bloomberg) wrote:
> > Hey Scott, can you take a look at this when you have time? Thanks!
> 
> I'm inclined to just delete setup.py, tbh. Chrome (and other projects) just
> clone-and-run.

I thought about just removing the setup.py as well.  However, the setup.py the
is useful for non clone-and-run environments (e.g., pip + virtualenv), where one
may want to have the gyp module and command line script installed.

> I don't think it's worth adding ez_setup.py at the very least.

I only added ez_setup.py in case setuptools wasn't installed.  I entirely agree
it is not necessary and would be happy remove it.

[scottmg, 2013-09-04 16:55:56]

On 2013/09/04 16:16:12, ericvw wrote:
> On 2013/09/04 15:48:52, scottmg wrote:
> > On 2013/09/04 14:16:50, Shezan Baig (Bloomberg) wrote:
> > > Hey Scott, can you take a look at this when you have time? Thanks!
> > 
> > I'm inclined to just delete setup.py, tbh. Chrome (and other projects) just
> > clone-and-run.
> 
> I thought about just removing the setup.py as well.  However, the setup.py the
> is useful for non clone-and-run environments (e.g., pip + virtualenv), where
one
> may want to have the gyp module and command line script installed.
> 
> > I don't think it's worth adding ez_setup.py at the very least.
> 
> I only added ez_setup.py in case setuptools wasn't installed.  I entirely
agree
> it is not necessary and would be happy remove it.

Ok, with ez_setup.py removed lgtm, with the caveat that it remains untested so
I'm not sure it's of much use.

[ericvw, 2013-09-04 17:43:38]

removed ez_setup.py as requested

[ericvw, 2013-09-04 17:47:56]

Patchset should be complete based on feedback and the description of the change
has been modified to match what is being introduced.  If someone could commit
for me, that would be great.  Thanks!

[Shezan Baig (Bloomberg), 2013-09-04 18:00:18]

Hmm, I just committed it using "git cl dcommit", and dcommit seemed to get stuck
in an infinite loop (kept prompting me for my username and password).  The
commit itself seems to have gone through:
https://code.google.com/p/gyp/source/detail?r=1713
But the issue is still open.  Scott, is there anything you can check on Rietveld
to see why the issue is still open?

[scottmg, 2013-09-05 19:14:50]

On 2013/09/04 18:00:18, Shezan Baig (Bloomberg) wrote:
> Hmm, I just committed it using "git cl dcommit", and dcommit seemed to get
stuck
> in an infinite loop (kept prompting me for my username and password).  The
> commit itself seems to have gone through:
> https://code.google.com/p/gyp/source/detail?r=1713
> But the issue is still open.  Scott, is there anything you can check on
Rietveld
> to see why the issue is still open?

I'm not sure why, but I just clicked the X to close it.