frog: check that const constructors have potentially const initializers. Also

frog: check that const constructors have potentially const initializers. Also
check that const objects don't have non-final fields (this makes
ConstConstructorNegativeTest fail for the right reasons)

Committed: https://code.google.com/p/dart/source/detail?r=2347

[Siggi Cherem (dart-lang), 2011-12-12 18:04:37]

addressing BUG http://code.google.com/p/dart/issues/detail?id=777

http://codereview.chromium.org/8921009/diff/1/tests/language/language.status
File tests/language/language.status (right):

http://codereview.chromium.org/8921009/diff/1/tests/language/language.status#...
tests/language/language.status:97: ConstInit5NegativeTest: Pass  # Should fail
at compile-time, not at runtime
We should go through negative tests and try to see if we can make them "pass"
when the feature is not implemented. In this example, the code also has a
runtime assertion that is not necessary and that it is hiding the real issue.

[jimhug, 2011-12-12 18:14:30]

LGTM

Nice fix and extension of the test suite all together!

http://codereview.chromium.org/8921009/diff/1/frog/member.dart
File frog/member.dart (right):

http://codereview.chromium.org/8921009/diff/1/frog/member.dart#newcode1148
frog/member.dart:1148: assign.y.span);
Could this be val.span instead?  It doesn't really matter right now, I'm just
thinking a little ahead to richer Values and more analysis with them.

http://codereview.chromium.org/8921009/diff/1/tests/language/language.status
File tests/language/language.status (right):

http://codereview.chromium.org/8921009/diff/1/tests/language/language.status#...
tests/language/language.status:97: ConstInit5NegativeTest: Pass  # Should fail
at compile-time, not at runtime
SGTM - I'd also like to figure out how to ensure that the negative tests are
reporting errors on the correct lines - as an opt-in feature to not break the
VM.

On 2011/12/12 18:04:38, sigmund wrote:
> We should go through negative tests and try to see if we can make them "pass"
> when the feature is not implemented. In this example, the code also has a
> runtime assertion that is not necessary and that it is hiding the real issue.

[Siggi Cherem (dart-lang), 2011-12-12 18:20:55]

http://codereview.chromium.org/8921009/diff/1/frog/member.dart
File frog/member.dart (right):

http://codereview.chromium.org/8921009/diff/1/frog/member.dart#newcode1148
frog/member.dart:1148: assign.y.span);
On 2011/12/12 18:14:30, jimhug wrote:
> Could this be val.span instead?  It doesn't really matter right now, I'm just
> thinking a little ahead to richer Values and more analysis with them.

For some reason in this case they are not the same - we need to look more into
it. In the new test, assing.y.span is the whole expression "1.toString()", but
val.span is just the method call portion ("toString").