@nikomatsakis Submitted a PR for this issue. I've not ran all the tests locally but it fixes the repro in the issue.
@David Wood see my review here
@nikomatsakis Pushed a fix. Didn't take long to do but then spent ages getting confused why the
run-pass test was failing even though it was now compiling correctly, then I noticed that the example had
heh sorry :)
@David Wood left a few comment nits :)
@nikomatsakis I'm not sure I follow your first nit.
Then when we do find an activation we can assert that the field is "not activated".
I took this (from first review) to mean that where we change to
ActivatedAt, we should assert that we were at
NotActivated (as opposed to
that is correct
I'm just saying that the purpose of that check
Are you referring to the comment?
is to make sure that we are "activating" something that we considered 2-phase in the first place
but your comment made it sound like this was checking that we found an activation for each thing
yes, the comment
I see, I thought you meant the
assert! itself, the location or contents of it.
to check that latter thing we'd have to do some kind of iteration at the end
ah, no. the code is fine.
"I think this commit" = "this comment"
I copied the comment from the previous assert that was removed, I'll update it to something more appropriate.
@nikomatsakis Pushed up the improved comments.