Switch to debruijn indices, genericize sub/lub/glb #18837
Conversation
nikomatsakis
force-pushed
the
hrtb-refactor-4-debruijn
branch
from
6302c59
to
8a7f4a5
Compare
|
@pnkfelix I realized I meant to write some more tests for specialized scenarios. I'll try to do that this morning. But |
nikomatsakis
force-pushed
the
hrtb-refactor-4-debruijn
branch
from
8a7f4a5
to
f903048
Compare
nikomatsakis
force-pushed
the
hrtb-refactor-4-debruijn
branch
from
f903048
to
3d2189b
Compare
|
OK, done writing some more tests and comments. I'm sure I could write more still but good enough for now perhaps. Did find one bug by writing unit tests so huzzah for proper engineering practices. :) |
| * | ||
| * Let's start with the reference type `&'b int` that is the first | ||
| * argument to the inner function. This region `'b` is assigned a De | ||
| * Bruin index of 1, meaning "the innermost binder" (in this case, a |
There was a problem hiding this comment.
Bruijn
Thanks! and I was being so careful to always spell it right :)
|
r=me after addressing the remaining typos pointed out by @tomjakubowski (and potentially addressing my minor feedback). |
|
I'll incorporate the feedback into the main HRTB PR. I've fixed a bunch of random bugs that I found in that branch and it'll be hard to sort out the history if we land this. |
More refactoring in support of HRTB. There should be no change in functionality from these changes.
r? @pnkfelix