Fix paths with different seperators being compared in the analyzer on Windows.

Fix paths with different seperators being compared in the analyzer on Windows.

Passes analyzer tests (were previously failing on Windows).

BUG=527935
R=dpranke@chromium.org

Committed: https://chromium.googlesource.com/external/gyp/+/2b17e0b26a93e8c9758c23aec6c554da4ca8f0a9

[Geoff Lang, 2015-09-03 19:25:05]

geofflang@chromium.org changed reviewers:
+ dpranke@google.com

[Geoff Lang, 2015-09-03 19:25:05]

PTAL

[Geoff Lang, 2015-09-03 19:25:19]

geofflang@chromium.org changed reviewers:
+ dpranke@chromium.org
- dpranke@google.com

[Ken Russell (switch to Gerrit), 2015-09-03 19:56:36]

kbr@chromium.org changed reviewers:
+ kbr@chromium.org

[Ken Russell (switch to Gerrit), 2015-09-03 19:56:36]

Are the analyzer tests run on Chromium's commit queue? They should be -- bugs in
this area will allow serious regressions to slip through.

[sky, 2015-09-03 20:01:46]

On 2015/09/03 19:56:36, Ken Russell wrote:
> Are the analyzer tests run on Chromium's commit queue?

Not too my knowledge, just as gyp tests in general aren't run on the cq.

> They should be -- bugs in
> this area will allow serious regressions to slip through.

Generally if there is a bug in gyp/analyzer you'll know about it quickly!

[Dirk Pranke, 2015-09-03 20:17:51]

On 2015/09/03 20:01:46, sky wrote:
> On 2015/09/03 19:56:36, Ken Russell wrote:
> > Are the analyzer tests run on Chromium's commit queue?
> 
> Not too my knowledge, just as gyp tests in general aren't run on the cq.
>

We do now have a functioning GYP waterfall and GYP trybots, and I've just sent
this patch
to them. Unfortunately, triggering try jobs through the rietveld UI doesn't
quite work yet,
so you have to do this manually using `git cl try` in a GYP checkout.

[Dirk Pranke, 2015-09-03 20:18:03]

lgtm, though.

[Dirk Pranke, 2015-09-03 20:44:13]

Committed patchset #1 (id:1) manually as
2b17e0b26a93e8c9758c23aec6c554da4ca8f0a9 (presubmit successful).

[Ken Russell (switch to Gerrit), 2015-09-03 20:48:02]

On 2015/09/03 20:01:46, sky wrote:
> On 2015/09/03 19:56:36, Ken Russell wrote:
> > Are the analyzer tests run on Chromium's commit queue?
> 
> Not too my knowledge, just as gyp tests in general aren't run on the cq.
> 
> > They should be -- bugs in
> > this area will allow serious regressions to slip through.
> 
> Generally if there is a bug in gyp/analyzer you'll know about it quickly!

FYI -- Geoff did some analysis and it seems this bug slipped into the Chromium
tree Tuesday. Nobody noticed in the intervening day or two that Windows tryjobs
were basically running no tests. Increased automated test coverage in this area
is needed.

[Dirk Pranke, 2015-09-03 21:00:32]

On 2015/09/03 20:48:02, Ken Russell wrote:
> On 2015/09/03 20:01:46, sky wrote:
> > On 2015/09/03 19:56:36, Ken Russell wrote:
> > > Are the analyzer tests run on Chromium's commit queue?
> > 
> > Not too my knowledge, just as gyp tests in general aren't run on the cq.
> > 
> > > They should be -- bugs in
> > > this area will allow serious regressions to slip through.
> > 
> > Generally if there is a bug in gyp/analyzer you'll know about it quickly!
> 
> FYI -- Geoff did some analysis and it seems this bug slipped into the Chromium
> tree Tuesday. Nobody noticed in the intervening day or two that Windows
tryjobs
> were basically running no tests. Increased automated test coverage in this
area
> is needed.

For the record, we actually have automated test coverage for this on the gyp
waterfall,
we just ignored it :(.