Report wrapped pub exceptions with --verbose and --trace.

Report wrapped pub exceptions with --verbose and --trace.

R=rnystrom@google.com
BUG=14677
Committed: https://code.google.com/p/dart/source/detail?r=30006

[Bob Nystrom, 2013-11-04 22:34:38]

Couple of suggestions then LGTM, but I think you probably should not land this
right now.

https://codereview.chromium.org/49293003/diff/1/sdk/lib/_internal/pub/lib/src...
File sdk/lib/_internal/pub/lib/src/solver/version_solver.dart (right):

https://codereview.chromium.org/49293003/diff/1/sdk/lib/_internal/pub/lib/src...
sdk/lib/_internal/pub/lib/src/solver/version_solver.dart:176: final innerTrace =
null;
Doc comments for these.

I'd also prefer them to be getters. Fields whose value can only ever be null and
are intended to be overridden look really perverse to me.

[nweiz, 2013-11-06 22:46:32]

https://codereview.chromium.org/49293003/diff/1/sdk/lib/_internal/pub/lib/src...
File sdk/lib/_internal/pub/lib/src/solver/version_solver.dart (right):

https://codereview.chromium.org/49293003/diff/1/sdk/lib/_internal/pub/lib/src...
sdk/lib/_internal/pub/lib/src/solver/version_solver.dart:176: final innerTrace =
null;
On 2013/11/04 22:34:38, Bob Nystrom wrote:
> Doc comments for these.

I think the usual style is to not write doc comments for inherited members. I've
added docs to the parent class's properties, though.

> I'd also prefer them to be getters. Fields whose value can only ever be null
and
> are intended to be overridden look really perverse to me.

I strongly disagree. Getters that have identical functionality to their
less-verbose field equivalents look really perverse to me.

These are also not intended to be overridden; they're doing the overriding.

[nweiz, 2013-11-06 22:47:28]

Committed patchset #2 manually as r30006 (presubmit successful).