Remove cygwin disease from grit (python) invocations :)

Remove cygwin disease from grit (python) invocations :)

R=scottmg,joi
BUG=123026
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=149892

[iannucci, 2012-08-03 00:40:35]

Small change (gyp only)

[scottmg, 2012-08-03 03:10:47]

Seems a bit odd that we process strings as file names in actions in one shell
but not in the other.

Did you try a ninja build too?

lgtm

[iannucci, 2012-08-03 03:57:22]

On 2012/08/03 03:10:47, scottmg wrote:
> Seems a bit odd that we process strings as file names in actions in one shell
> but not in the other.
> 
> Did you try a ninja build too?
> 
> lgtm

On 2012/08/03 03:10:47, scottmg wrote:
> Seems a bit odd that we process strings as file names in actions in one shell
> but not in the other.
> 
> Did you try a ninja build too?
> 
> lgtm

I'll double check wininja.. I thought it would be covered by the trybots, but I
guess not in reality :)

I think the issue is that we process the strings poorly for the combination of
(msvs+native shell). Theoretically, I should dive in and make all that quoting
better, but secretly I want to move to ninja on windows as quickly as possible
(getting rid of cygwin along the way).

I think, really, there needs to be some way to explicitly call out something as
a path... For example, in this instance, even the very best heuristic would
label that string as a path :/. Something like:
  * _("some/path")  # Yes, we can have _ :)
  * p("some/path")  # If there are objections to 'magic' variable names

They could behave like regular strings, too, so things like:
  '--option=['+p('a/path/thats/awesome')+', '+p('another/path')+']'

would behave in all ways like a string, except that there would be a way to
declare (before generation) if paths should be unix or windows style. Explicitly
wrapping paths would also be able to have gyp warn/error about embedded paths
which are in windows-style (For Best Consistency). It would also let gyp do more
smart things about path manipulation.

I guess telling people that they have to add an additional 3 characters per path
could be annoying, but I think the benefits are numerous/plentiful.

What do you think? If this seems like a good thing to pursue, I'll repost onto
gyp-developers.

[Jói, 2012-08-03 08:50:09]

LGTM for this change, thanks!

A "cleaner" plan like you suggested for a larger-scale fix to the path issue
sounds good to pursue if you have time.

Cheers,
Jói

[commit-bot: I haz the power, 2012-08-03 18:20:08]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/iannucci@chromium.org/10832121/1

[commit-bot: I haz the power, 2012-08-03 19:41:52]

Change committed as 149892

[scottmg, 2012-08-06 18:29:51]

On 2012/08/03 19:41:52, I haz the power (commit-bot) wrote:
> Change committed as 149892

trybots (win and win_rel) are currently failing in strange ways.
http://build.chromium.org/p/tryserver.chromium/builders/win_rel/builds/50016/...

I don't actually think it's due to this change (because its tries were clean)
but I'm going to try reverting it anyway because I don't see any other CLs in
the last 3 days that look relevant.

One other plausible candidate is the grit roll, perhaps we should try that one
too.

[scottmg, 2012-08-06 19:40:07]

On 2012/08/06 18:29:51, scottmg wrote:
> On 2012/08/03 19:41:52, I haz the power (commit-bot) wrote:
> > Change committed as 149892
> 
> trybots (win and win_rel) are currently failing in strange ways.
>
http://build.chromium.org/p/tryserver.chromium/builders/win_rel/builds/50016/...
> 
> I don't actually think it's due to this change (because its tries were clean)
> but I'm going to try reverting it anyway because I don't see any other CLs in
> the last 3 days that look relevant.
> 
> One other plausible candidate is the grit roll, perhaps we should try that one
> too.

Reverting seems to have fixed the trybots. I'm not sure if there's necessarily
any causality though.

Perhaps it's because the path handling at http://crrev.com/149861 interacted
poorly with this? Perhaps just that dependencies are messed up?

I'll try a try on the revert of the revert. (! wat)

[scottmg, 2012-08-06 19:42:36]

On 2012/08/06 19:40:07, scottmg wrote:
> On 2012/08/06 18:29:51, scottmg wrote:
> > On 2012/08/03 19:41:52, I haz the power (commit-bot) wrote:
> > > Change committed as 149892
> > 
> > trybots (win and win_rel) are currently failing in strange ways.
> >
>
http://build.chromium.org/p/tryserver.chromium/builders/win_rel/builds/50016/...
> > 
> > I don't actually think it's due to this change (because its tries were
clean)
> > but I'm going to try reverting it anyway because I don't see any other CLs
in
> > the last 3 days that look relevant.
> > 
> > One other plausible candidate is the grit roll, perhaps we should try that
one
> > too.
> 
> Reverting seems to have fixed the trybots. I'm not sure if there's necessarily
> any causality though.
> 
> Perhaps it's because the path handling at http://crrev.com/149861 interacted
> poorly with this? Perhaps just that dependencies are messed up?
> 
> I'll try a try on the revert of the revert. (! wat)

http://chromiumcodereview.appspot.com/10823188