-
Notifications
You must be signed in to change notification settings - Fork 12.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Ban registering obligations during InferCtxt snapshots. #41325
Conversation
// | ||
// HOWEVER, in some cases the flag is wrong. In particular, we |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe it's better to have a within_snapshot
boolean on FulfillmentCx
that disables the check?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah it's not super clear what the best option is. It's probably a good idea to tie fulfillment contexts to inference contexts (through lifetimes?) and maybe have a snapshot depth... or maybe an "epoch"?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Longer term, I have been hoping to just get rid of the notion of snapshots except for very short-term (e.g., to make a unification operation atomic) and then create nested fulfillment contexts for "probing sessions".
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Well, maybe that's not a complete answer. Still, I guess I hope that we can move things into nested contexts and generally remove the need for snapshots.
Crater run has found one root regression. EDIT: reproduced it (looks like obligations need to be selected eagerly for binops - Travis failure is similar): fn main() {
let mut s = String::new();
s += {&String::new()};
// ^^^^^^^^^^^^^^ expected str, found struct `std::string::String`
// note: expected type `&str`
// found type `&std::string::String`
} EDIT2: pushed a fix. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me, do we want to do a crater run or anything like that? I could imagine some subtle effects (as evidenced by the last commit).
Ah, I see @eddyb is ahead of me here. Very good. |
@bors r+ |
📌 Commit cd64ff9 has been approved by |
@bors r- |
Actually, @arielb1, were you satisfied? (If so, feel free to r=me, or you :) |
@bors r+ |
📌 Commit cd64ff9 has been approved by |
…=arielb1 Ban registering obligations during InferCtxt snapshots. Back in rust-lang#33852, a flag was added to `InferCtxt` to prevent rolling back a snapshot if obligations were added to some `FulfillmentContext` during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse. But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations *do* get caught nowadays, those errors prevent, e.g. the speculative coercions from rust-lang#37658, which *have to* be rolled back *even* if they succeed. We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse. Instead, this PR modifies `lookup_method_in_trait_adjusted` to return `InferOk` containing the obligations that `Autoderef::finalize_as_infer_ok` can propagate to deref coercions. As there shouldn't be *anything* left that registers obligations during snapshots, it's completely banned. r? @nikomatsakis @arielb1
…=arielb1 Ban registering obligations during InferCtxt snapshots. Back in rust-lang#33852, a flag was added to `InferCtxt` to prevent rolling back a snapshot if obligations were added to some `FulfillmentContext` during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse. But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations *do* get caught nowadays, those errors prevent, e.g. the speculative coercions from rust-lang#37658, which *have to* be rolled back *even* if they succeed. We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse. Instead, this PR modifies `lookup_method_in_trait_adjusted` to return `InferOk` containing the obligations that `Autoderef::finalize_as_infer_ok` can propagate to deref coercions. As there shouldn't be *anything* left that registers obligations during snapshots, it's completely banned. r? @nikomatsakis @arielb1
…=arielb1 Ban registering obligations during InferCtxt snapshots. Back in rust-lang#33852, a flag was added to `InferCtxt` to prevent rolling back a snapshot if obligations were added to some `FulfillmentContext` during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse. But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations *do* get caught nowadays, those errors prevent, e.g. the speculative coercions from rust-lang#37658, which *have to* be rolled back *even* if they succeed. We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse. Instead, this PR modifies `lookup_method_in_trait_adjusted` to return `InferOk` containing the obligations that `Autoderef::finalize_as_infer_ok` can propagate to deref coercions. As there shouldn't be *anything* left that registers obligations during snapshots, it's completely banned. r? @nikomatsakis @arielb1
…=arielb1 Ban registering obligations during InferCtxt snapshots. Back in rust-lang#33852, a flag was added to `InferCtxt` to prevent rolling back a snapshot if obligations were added to some `FulfillmentContext` during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse. But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations *do* get caught nowadays, those errors prevent, e.g. the speculative coercions from rust-lang#37658, which *have to* be rolled back *even* if they succeed. We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse. Instead, this PR modifies `lookup_method_in_trait_adjusted` to return `InferOk` containing the obligations that `Autoderef::finalize_as_infer_ok` can propagate to deref coercions. As there shouldn't be *anything* left that registers obligations during snapshots, it's completely banned. r? @nikomatsakis @arielb1
Ban registering obligations during InferCtxt snapshots. Back in #33852, a flag was added to `InferCtxt` to prevent rolling back a snapshot if obligations were added to some `FulfillmentContext` during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse. But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations *do* get caught nowadays, those errors prevent, e.g. the speculative coercions from #37658, which *have to* be rolled back *even* if they succeed. We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse. Instead, this PR modifies `lookup_method_in_trait_adjusted` to return `InferOk` containing the obligations that `Autoderef::finalize_as_infer_ok` can propagate to deref coercions. As there shouldn't be *anything* left that registers obligations during snapshots, it's completely banned. r? @nikomatsakis @arielb1
☀️ Test successful - status-appveyor, status-travis |
Back in #33852, a flag was added to
InferCtxt
to prevent rolling back a snapshot if obligations were added to someFulfillmentContext
during the snapshot, to prevent leaking fresh inference variables (created during that snapshot, so their indices would get reused) in obligations, which could ICE or worse.But that isn't enough in the long run, as type-checking ends up relying on success implying that eager side-effects are fine, and while stray obligations do get caught nowadays, those errors prevent, e.g. the speculative coercions from #37658, which have to be rolled back even if they succeed.
We can't just allow those obligations to stay around though, because we end up, again, in ICEs or worse.
Instead, this PR modifies
lookup_method_in_trait_adjusted
to returnInferOk
containing the obligations thatAutoderef::finalize_as_infer_ok
can propagate to deref coercions.As there shouldn't be anything left that registers obligations during snapshots, it's completely banned.
r? @nikomatsakis @arielb1