[Monotone-devel] nvm.experiment.encapsulation review

[Zack Weinberg]

I read through all of the changes on this branch.  Overall, I like it,
and I think we ought to finish it and get it merged.  I brought it up
to date with current mainline and did a few more app_state removals,
as you probably will have noticed if you were watching IRC yesterday.

I do have two big concerns that I think need to get resolved first, though:-

Obviously, we need to do something about the hackish exposure of small
pieces of app_state through the database and keystore (and workspace?)
objects.  I do not have a good general idea, particularly when it
comes to the Lua hooks, which have the problem of being tied to an
intrinsically global object (the Lua interpreter blob).  Wherever
possible, though, I think we ought to be looking at passing down e.g.
information from app.opts in arguments.

Is the goal here to get rid of app_state entirely, or just in
everything below the level of CMD() functions?  I think that getting
rid of app_state entirely might be a good idea but is too ambitious
for a branch that needs to keep up with mainline the way this one
does.  The merge yesterday was No Fun.  You made some changes in CMD()
functions that only make sense if you're trying for a total removal,
and which may interfere with dealing with the hackish app_state
exposure stuff (like, I recall seeing uses of db.get_project()
introduced to CMD()s.

Speaking of get_project(), part of why it's so weird is that it's a
stub; that was where a bunch of the policy branch stuff that never got
implemented was supposed to go.  I believe Tim did the original
implementation there - can you please look over the project-related
changes in .encapsulation, Tim, and comment?

zw