[WIP] introduce a region unification table and use it in dropck #30242
Conversation
arielb1
force-pushed
the
region-unification
branch
from
7e271b2
to
80e191f
Compare
| @@ -45,6 +45,41 @@ impl<'a, 'tcx> ty::fold::TypeFolder<'tcx> for OpportunisticTypeResolver<'a, 'tcx | |||
| } | |||
| } | |||
|
|
|||
| /// The opportunistic type and region resolver is similar to the | |||
| /// opportunistic type resolver, but also opportunistly resolves | |||
| /// regions. It is useful for canonicalization. | |||
There was a problem hiding this comment.
Why not just always resolve regions?
|
OK, so, re-reading your comment, I expect that you were probably shooting for a good way to make a kind of minimally invasive (and minimally risky) version of this PR, which makes a lot of sense. |
|
I like this better than the alternative I proposed, and my main complaints about this PR was that it was not going as far as I would expect -- but that's probably good from the POV of being minimally invasive. @bors r+ |
|
📌 Commit 80e191f has been approved by |
|
@bors p=1 |
|
⌛ Testing commit 80e191f with merge d824ab1... |
|
💔 Test failed - auto-mac-64-opt |
|
@bors: retry On Fri, Dec 11, 2015 at 11:38 PM, bors notifications@github.com wrote:
|
Fixes #29844 I would prefer to (a) make some performance measurements (b) use the unification table in a few more places before committing further, but this is probably good enough for beta. r? @nikomatsakis
|
This commit causes a 50% regression in typeck performance (not the WIP). I will write a separate patch to fix it. |
Namely PR #30368 (right?) |
|
Right that was #30368 |
Fixes #29844
I would prefer to
(a) make some performance measurements
(b) use the unification table in a few more places
before committing further, but this is probably good enough for beta.
r? @nikomatsakis