Fixing parsing of factories.

Fixing factories.

Committed: https://code.google.com/p/dart/source/detail?r=2043

[Jennifer Messerly, 2011-12-02 23:27:10]

This is the simple fix that makes it parse.

(I tried doing something better with the type parameters, but backed out of it.
I realized I don't really understand what they're for and anything I wrote would
probably be wrong.)

[Jennifer Messerly, 2011-12-02 23:58:47]

On 2011/12/02 23:27:10, John Messerly wrote:
> This is the simple fix that makes it parse.
> 
> (I tried doing something better with the type parameters, but backed out of
it.
> I realized I don't really understand what they're for and anything I wrote
would
> probably be wrong.)

tbr

[Jennifer Messerly, 2011-12-03 01:39:48]

On 2011/12/02 23:58:47, John Messerly wrote:
> On 2011/12/02 23:27:10, John Messerly wrote:
> > This is the simple fix that makes it parse.
> > 
> > (I tried doing something better with the type parameters, but backed out of
> it.
> > I realized I don't really understand what they're for and anything I wrote
> would
> > probably be wrong.)
> 
> tbr

This change exposed a latent bug where our validation of "extends" happens too
soon, so may or may not work depending on what order the types were returned
from the HashMap (ick). This causes different behavior running Frog on the VM vs
selfhost. I'm working on a proper fix.

[jimhug, 2011-12-08 16:24:30]

lgtm

Your change is fantastic in terms of tests fixed for minimal code changes! 
However, the feature that you're implementing here still hurts my brain...

http://codereview.chromium.org/8788004/diff/6001/frog/parser.dart
File frog/parser.dart (right):

http://codereview.chromium.org/8788004/diff/6001/frog/parser.dart#newcode260
frog/parser.dart:260: // anything as far as I can tell.
Just hold your nose and parse them <smile/frown>

http://codereview.chromium.org/8788004/diff/6001/tests/language/language.status
File tests/language/language.status (right):

http://codereview.chromium.org/8788004/diff/6001/tests/language/language.stat...
tests/language/language.status:343: TypeVariableBoundsTest/none: Pass # I think
the "cancel leg" message is confusing the test harness? Leg is bailing on all
these tests after seeing the "extends" factory syntax.
Please file this as a frog bug to replace this long comment in the status file.