%clay $flue scry

[tadad]

Resolves #6331

%pyro installs are quite slow because we cannot utilize the host ships's $flue cache. By adding the ability to scry this out, we can build desks on ships in seconds. I have a demo of this working here.

edit: I added the ability to scry out the entire dome. This also required a renaming of $dome:clay to $domo in lull.hoon, since now the entire $dome is in lull.

[belisarius222]

This changes the return type of a scry, which would necessitate burning a Kelvin. We might be able to get away with not burning a Kelvin if we added a new scry instead, although it's somewhat questionable because an agent wouldn't be able to specify that it needs to be run in a kernel that has this scry as a capability.

Alternatively, we could merge this into the next Kelvin release. I think that's better, since the "domes" scry is more useful with this information in it, and duplicating the interface just to add the $flue cache strikes me as unclean. I won't officially approve this PR yet, until we're sure we're burning a Kelvin, but the code looks good. If we're not burning a Kelvin in this next release, then we will hold onto this work to be release when the next Kelvin goes out.

[tadad]

yeah I agree we should burn a kelvin.

I was also thinking maybe we should just scry out the entire dome (we only leave out two pieces from this scry now, mim (mime cache), and wic(commit in waiting). Unless there is something wrong with scrying the whole thing, it will let us get rid of some duplicated code between clay and lull.

[tadad]

just changed it so that you can scry out the entire $dome now. This moves $dome into lull.hoon, which made me have to change the name of $dome:clay to $domo:clay, which I think is better style

[belisarius222]

LGTM. We should hold off on merging until we're ready to burn the next Kelvin.