Fix .d for ninja on Windows

Fix .ds for ninja on Windows

Currently relies on 'deplist' branch of ninja. Also fix some prefixes/suffixes
for Windows. Tests in test/hello now pass.
Committed: https://code.google.com/p/gyp/source/detail?r=1202

[Evan Martin, 2012-02-15 19:53:04]

LGTM

https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/ninja_syntax.py
File pylib/gyp/ninja_syntax.py (right):

https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/ninja_syntax....
pylib/gyp/ninja_syntax.py:46: self.variable('deplist', deplist, indent=1)
Hm, in theory this file shouldn't diverge from upstream.  I wonder if this code
would be better written as kwargs anyway.
I don't think I care too much though.

[scottmg, 2012-02-15 21:55:46]

On 2012/02/15 19:53:04, Evan Martin wrote:
> LGTM
> 
>
https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/ninja_syntax.py
> File pylib/gyp/ninja_syntax.py (right):
> 
>
https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/ninja_syntax....
> pylib/gyp/ninja_syntax.py:46: self.variable('deplist', deplist, indent=1)
> Hm, in theory this file shouldn't diverge from upstream.  I wonder if this
code
> would be better written as kwargs anyway.
> I don't think I care too much though.

Oops, I didn't read that upstream comment. I guess it's not too bad because at
least won't fail silently if upstream is pulled w/o it, since it's an explicit
parameter. Once we figure out the best way to do deps in ninja-master I'll make
sure to update this.

[Nico, 2012-02-16 19:32:26]

https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/generator/nin...
File pylib/gyp/generator/ninja.py (right):

https://chromiumcodereview.appspot.com/9401016/diff/1/pylib/gyp/generator/nin...
pylib/gyp/generator/ninja.py:834: if self.flavor == 'win':
This should be !=, or you need to swap the branch bodies.

(This is breaking the linux build with nacl enabled)