Behn pass errors to client vanes

[belisarius222]

This sets the groundwork for future work for vanes to handle async errors. I just added stubs to the various vanes for now, but at least they now have the error information.

A |reset passed, but I have not done more detailed testing on this. Not sure if there's a convenient way to test. Maybe pH?

[belisarius222]

I resolved a merge conflict in Clay by checking both the error in the %wake move and whether the wire is %tyme.

[belisarius222]

Broke :aqua-behn. Fix incoming.

[belisarius222]

@philipcmonk I think I fixed Aqua; we'll see what the tests say. I added %crud to +unix-task in Zuse, and I modified :aqua-behn to send one as an aqua-event when it receives a Behn error. If you'd prefer different handling for that case, let me know and I can change it.

[belisarius222]

Tests passed; ready for review.

[joemfb]

@belisarius222, since this is going into next, do you want to coordinate the %crud protocol change in vere at the same time? I could do that if you'd like.

[belisarius222]

@joemfb I'm down, although I think we should get this in as-is first. I think it'll be a little cleaner that way, although I could be convinced otherwise.

[philipcmonk]

This looks good. My only question is whether there are apps that will choke when given an erroring wake. This is still strictly an improvement, though.

[joemfb]

It looks to me like all the apps that set timers will fail if the are called with an error %wake. It looks like gall just prints an error in this case, but doesn't crash -- if that's right, %behn will still be unblocked in the event of app timer failure, which is the most important result.