Remove support for the old Span classes from source_maps.

Remove support for the old Span classes from source_maps.

This releases source_maps 0.10.0, code_transformers 0.2.0+2, observe 0.11.0+3,
and polymer 0.12.0+4.

BUG=19930
R=sigmund@google.com

Committed: https://code.google.com/p/dart/source/detail?r=38803

[Siggi Cherem (dart-lang), 2014-07-31 21:01:38]

lgtm, modulo fixing the type inconsistency  (probably just changing Url back to
String is all you need)

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/lib/parser.dart
File pkg/source_maps/lib/parser.dart (right):

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/lib/parser.d...
pkg/source_maps/lib/parser.dart:189: sourceUrl == null ? '' :
sourceUrl.toString(), () => urls.length);
This seems inconsistent with the change on the type. If the key is a Url, we
could revert the change in these last 3 lines, but we'd also need to change
below and convert the keys to strings before passing them to the constructor.

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/test/refacto...
File pkg/source_maps/test/refactor_test.dart (right):

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/test/refacto...
pkg/source_maps/test/refactor_test.dart:148: return span == null ? null :
span.message('').trim();
mmm... seeing this test made me realize that 'message' is always including the ^
char. I wonder if we had usages/places where we don't expect it to be there.

[nweiz, 2014-07-31 22:01:19]

Code review changes

[nweiz, 2014-07-31 22:01:25]

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/lib/parser.dart
File pkg/source_maps/lib/parser.dart (right):

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/lib/parser.d...
pkg/source_maps/lib/parser.dart:189: sourceUrl == null ? '' :
sourceUrl.toString(), () => urls.length);
On 2014/07/31 21:01:38, Siggi Cherem (dart-lang) wrote:
> This seems inconsistent with the change on the type. If the key is a Url, we
> could revert the change in these last 3 lines, but we'd also need to change
> below and convert the keys to strings before passing them to the constructor.

Oops, meant to get rid of the type change above. Done.

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/test/refacto...
File pkg/source_maps/test/refactor_test.dart (right):

https://codereview.chromium.org/421723004/diff/1/pkg/source_maps/test/refacto...
pkg/source_maps/test/refactor_test.dart:148: return span == null ? null :
span.message('').trim();
On 2014/07/31 21:01:38, Siggi Cherem (dart-lang) wrote:
> mmm... seeing this test made me realize that 'message' is always including the
^
> char. I wonder if we had usages/places where we don't expect it to be there.

What's the point of formatting a message if it's not displaying information
about where the span is? I think it's a feature that [message] is now consistent
about what format it emits, especially since that consistency uniformly provides
the end user with more information.

[Siggi Cherem (dart-lang), 2014-07-31 22:02:26]

lgtm

[nweiz, 2014-07-31 22:04:43]

Committed patchset #2 manually as r38803 (presubmit successful).