use the real matchers for running language tests

use the real matchers for running language tests

R=vsm@google.com

Committed: https://github.com/dart-lang/dev_compiler/commit/78d906d5b8ceb161fc886355ca21c79f93bc0993

[Jennifer Messerly, 2015-08-27 22:04:59]

jmesserly@google.com changed reviewers:
+ vsm@google.com

[vsm, 2015-08-28 00:05:54]

https://codereview.chromium.org/1321683004/diff/20001/karma.conf.js
File karma.conf.js (right):

https://codereview.chromium.org/1321683004/diff/20001/karma.conf.js#newcode28
karma.conf.js:28: 'test/codegen/expect/matcher/src/*.js',
Are these supposed to be checked in with this CL?

[Jennifer Messerly, 2015-08-28 16:09:20]

https://codereview.chromium.org/1321683004/diff/20001/karma.conf.js
File karma.conf.js (right):

https://codereview.chromium.org/1321683004/diff/20001/karma.conf.js#newcode28
karma.conf.js:28: 'test/codegen/expect/matcher/src/*.js',
On 2015/08/28 00:05:54, vsm wrote:
> Are these supposed to be checked in with this CL?

no. I added them to the gitignore.

[vsm, 2015-08-28 16:18:38]

Does matcher need to be added to the pubspec to ensure it's there?  Seems like
we're getting it transitively now?

Otherwise, LGTM.

[Jennifer Messerly, 2015-08-28 16:22:42]

On 2015/08/28 16:18:38, vsm wrote:
> Does matcher need to be added to the pubspec to ensure it's there?  Seems like
> we're getting it transitively now?
> 
> Otherwise, LGTM.

Right, it gets pulled in from `package:test`

Up to us if we want it in the pubspec. I didn't add it because one more version
to juggle, and if suddenly test stopped depending on matcher (very unlikely),
we'd see a test failure anyway to alert us to the problem.

[vsm, 2015-08-28 16:26:44]

On 2015/08/28 16:22:42, John Messerly wrote:
> On 2015/08/28 16:18:38, vsm wrote:
> > Does matcher need to be added to the pubspec to ensure it's there?  Seems
like
> > we're getting it transitively now?
> > 
> > Otherwise, LGTM.
> 
> Right, it gets pulled in from `package:test`
> 
> Up to us if we want it in the pubspec. I didn't add it because one more
version
> to juggle, and if suddenly test stopped depending on matcher (very unlikely),
> we'd see a test failure anyway to alert us to the problem.

lgtm / sgtm

[Jennifer Messerly, 2015-08-28 16:27:40]

Committed patchset #2 (id:20001) manually as
78d906d5b8ceb161fc886355ca21c79f93bc0993 (presubmit successful).