Since dartc has no backend anymore, a runtime error in a multitest does *not* make it a negative te…

Sicne dartc has no backend anymore, a runtime error in a multitest does *not* make it a negative test.
Committed: https://code.google.com/p/dart/source/detail?r=4531

[zundel, 2012-02-23 13:00:24]

lgtm

in light of recent discussion, testing the exception type and using a 'string'
as an exception value might be change, but this isn't really about the test
itself.

I was just trying  hard to think of a way to recast this as a positive test, but
I see that if we don't test this fundamental assumption, we could otherwise get
lots of false positives from running the tests.

[ahe, 2012-02-23 13:30:30]

On 2012/02/23 13:00:24, zundel wrote:
> lgtm
> 
> in light of recent discussion, testing the exception type and using a 'string'
> as an exception value might be change, but this isn't really about the test
> itself.
> 
> I was just trying  hard to think of a way to recast this as a positive test,
but
> I see that if we don't test this fundamental assumption, we could otherwise
get
> lots of false positives from running the tests.

LGTM.

Eric, you're absolutely right. This test was motivated by a bug that broke stuff
like Expect.throws.

Also, recasting this as a multi-test means that we ensure that the test doesn't
grow stale or contains an unintended syntax error.

[ngeoffray, 2012-02-23 13:56:12]

Thanks both of you for having taken a look.