Add validation logic for non-nullable types.

Add validation logic for non-nullable types.

This CL only turns on the non-nullable check for validation tests.

There will be separate CLs for:
- add DCHECK at the sending side.
- make the existing APIs pass the non-nullable check and turn on the check everywhere.

BUG=324170
TEST=New and revised validation tests.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=290002

[yzshen1, 2014-08-13 21:00:41]

Hi, Darin and Tom.

Would you please take a look? Thanks!

Darin: everything except the tests.
Tom: everything.

Also CC-ing Hans who is working on JS validation implementation.

[darin (slow to review), 2014-08-13 21:39:09]

LGTM

[Tom Sepez, 2014-08-15 17:12:00]

Q. Are we going to be able to extend this mechanism you've introduced to
validate out-of-range enums as well?  Otherwise LGTM.

[yzshen1, 2014-08-15 17:29:08]

On 2014/08/15 17:12:00, Tom Sepez wrote:
> Q. Are we going to be able to extend this mechanism you've introduced to
> validate out-of-range enums as well?  Otherwise LGTM.

It seems good to validate out-of-range enums automagically.
The tricky thing is that we need to consider what "extensibility" means in this
case:
If we allow appending new values to an enum, then what should we do when a
client (running new version) sends an unrecognizable value to a service (running
old version). Does the service reject the message or somehow tolerate it?
Trung has a suggestion in this thread "Mojom: enums, expressions, and default
values", which I think is good.

Yeah, I would like to work on that at some point, not sure whether I have enough
time recently, though. :)

Thanks for the suggestion!

[yzshen1, 2014-08-15 17:29:21]

The CQ bit was checked by yzshen@chromium.org

[commit-bot: I haz the power, 2014-08-15 17:30:47]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/yzshen@chromium.org/466613002/80001

[commit-bot: I haz the power, 2014-08-15 20:34:39]

Committed patchset #5 (80001) as 290002