Fixes #163 - small fix in the mangling logic, renames all expected test files

Fixes #163 - small fix in the mangling logic, renames all expected test files
that now don't use mangling.

Committed: https://github.com/dart-lang/dart-web-components/commit/6c41469

[Siggi Cherem (dart-lang), 2012-11-14 22:01:31]

mainly only compiler.dart, info.dart, and run.sh have changes. All the
expectation files are just renamed.

https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.dart
File lib/src/compiler.dart (left):

https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.da...
lib/src/compiler.dart:63: var _mangleFilenames = basePath != outputPath;
I just realized this should be == instead of !=, and should be checked after we
normalize paths (basically in line 82), when I saw that, it made sense to hide
it under the PathInfo constructor.

[justinfagnani, 2012-11-14 22:07:13]

lgtm

https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.dart
File lib/src/compiler.dart (left):

https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.da...
lib/src/compiler.dart:63: var _mangleFilenames = basePath != outputPath;
On 2012/11/14 22:01:31, Siggi Cherem (dart-lang) wrote:
> I just realized this should be == instead of !=, and should be checked after
we
> normalize paths (basically in line 82), when I saw that, it made sense to hide
> it under the PathInfo constructor.

d'oh. the tests don't cover both mangling and no-mangling cases yet, so this
passed. should we cover both?

[Siggi Cherem (dart-lang), 2012-11-14 22:12:46]

On 2012/11/14 22:07:13, justinfagnani wrote:
> lgtm
> 
>
https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.dart
> File lib/src/compiler.dart (left):
> 
>
https://chromiumcodereview.appspot.com/11293285/diff/2001/lib/src/compiler.da...
> lib/src/compiler.dart:63: var _mangleFilenames = basePath != outputPath;
> On 2012/11/14 22:01:31, Siggi Cherem (dart-lang) wrote:
> > I just realized this should be == instead of !=, and should be checked after
> we
> > normalize paths (basically in line 82), when I saw that, it made sense to
hide
> > it under the PathInfo constructor.
> 
> d'oh. the tests don't cover both mangling and no-mangling cases yet, so this
> passed. should we cover both?

good idea - will do. For now I'll submit this fix to get things in a working
state (we might overwrite files in the input dir if --out is omitted in the
current form)