Uh oh!
There was an error while loading. Please reload this page.
Handle const-checks for &mut outside of HasMutInterior - #66654
Merged
Conversation
rust-highfive
commented
Nov 23, 2019
Contributor
(rust_highfive has picked a reviewer for you, use r? to override) |
ContributorAuthor
r? @eddyb (although I'd like to work on this a bit more, it's not quite ready) |
&mut outside of HasMutInterior&mut outside of HasMutInteriorecstatic-morseforce-pushed
the
check-consts-ref
branch
from
November 23, 2019 00:52
ba0ce7e to
b40d0bdComparebors
commented
Nov 23, 2019
Collaborator
☔ The latest upstream changes (presumably #66640) made this pull request unmergeable. Please resolve the merge conflicts. |
Centril
reviewed
Nov 23, 2019
Contributor
There was a problem hiding this comment.
(btw, a doc comment on these methods would be great)
Contributor
There was a problem hiding this comment.
Can we delay_span_bug just in case?
Centril
commented
Nov 23, 2019
Contributor
Uh oh!
There was an error while loading. Please reload this page.
eddyb
reviewed
Nov 26, 2019
Contributor
There was a problem hiding this comment.
I thought this was resolved by simply not having any qualifs on &Cell::new(42).
&T is always Freeze + Copy, regardless of T, so no qualif bits need to be set.
The only reason they were before was for promotion, which is why we split promotion into its own check.
ContributorAuthor
There was a problem hiding this comment.
I think you're right. This means I can clean up a lot more!
ecstatic-morseforce-pushed
the
check-consts-ref
branch
2 times, most recently
from
November 26, 2019 20:17
c93bb56 to
639aeecCompareecstatic-morse
marked this pull request as ready for review
November 26, 2019 20:18
rust-highfive
commented
Nov 26, 2019
Contributor
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
bors
commented
Nov 27, 2019
Collaborator
☔ The latest upstream changes (presumably #66677) made this pull request unmergeable. Please resolve the merge conflicts. |
ecstatic-morseforce-pushed
the
check-consts-ref
branch
from
November 27, 2019 21:05
639aeec to
4c1f3b1CompareNow, `derived_from_illegal_borrow` is also applied to reborrows, so we don't show the user a useless error message.
ecstatic-morseforce-pushed
the
check-consts-ref
branch
from
November 27, 2019 21:07
4c1f3b1 to
7f35ab6CompareContributorAuthor
Now that #66640 has been merged, the const-checker treats borrows of a @matthewjasper I'm not quite sure why #66587 didn't cause this test to change. How do I check that |
matthewjasper
commented
Nov 27, 2019
Contributor
I believe it didn't change in that PR because we don't special case reborrows in |
matthewjasper
commented
Nov 27, 2019
Contributor
The |
This error code is never emitted, and the contents of this test are identical to that of `E0017.rs`.
ecstatic-morseforce-pushed
the
check-consts-ref
branch
from
November 27, 2019 22:38
7f35ab6 to
7673400Comparerust-highfive
commented
Nov 27, 2019
Contributor
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
They now look the same in the MIR after rust-lang#66587.
This checks `static mut` as well for E0017, and blesses tests now that we emit an error for a mut deref.
ecstatic-morseforce-pushed
the
check-consts-ref
branch
from
November 28, 2019 00:15
7673400 to
846be82CompareContributorAuthor
@eddyb I had to make a few changes in |
eddyb
reviewed
Nov 28, 2019
| if let LocalInfo::StaticRef { .. } = decl.local_info { | ||
| return None; | ||
| } | ||
| } |
Contributor
There was a problem hiding this comment.
@matthewjasper isn't this supposed to use a helper method?
Contributor
There was a problem hiding this comment.
Yes, there's body.local_decls[local].is_ref_to_static
eddyb
reviewed
Nov 28, 2019
| | Rvalue::Ref(_, BorrowKind::Shared, ref place) | ||
| | Rvalue::Ref(_, BorrowKind::Shallow, ref place) | ||
| if matches!(place.base, PlaceBase::Static(_)) | ||
| => {} |
Contributor
There was a problem hiding this comment.
Does this make sense anymore, with the static changes?
Contributor
There was a problem hiding this comment.
No, this shouldn't be reachable, since the only PlaceBase::Statics are promoted. This maybe should be bug!() for now.
eddyb
commented
Nov 28, 2019
Contributor
r? @matthewjasper for all the |
matthewjasper
commented
Dec 2, 2019
Contributor
@bors r=eddyb,matthewjasper |
bors
commented
Dec 2, 2019
Collaborator
📌 Commit ccb4eed has been approved by |
RalfJung added a commit
to RalfJung/rust
that referenced
this pull request
Dec 2, 2019
…eddyb,matthewjasper Handle const-checks for `&mut` outside of `HasMutInterior` Addresses [this comment](rust-lang#64470 (comment)). Const-checking relied on `HasMutInterior` to forbid `&mut` in a const context. This was strange because all we needed to do was look for an `Rvalue::Ref` with a certain `BorrowKind`, whereas the `Qualif` traits are specifically meant to get the qualifs for a *value*. This PR removes that logic from `HasMutInterior` and moves it into `check_consts::Validator`. As a result, we can now properly handle qualifications for `static`s, which had to be ignored previously since you can e.g. borrow a static `Cell` from another `static`. We also remove the `derived_from_illegal_borrow` logic, since it is no longer necessary; we give good errors for subsequent reborrows/borrows of illegal borrows.
bors added a commit
that referenced
this pull request
Dec 2, 2019
Rollup of 5 pull requests Successful merges: - #66245 (Conditional compilation for sanitizers) - #66654 (Handle const-checks for `&mut` outside of `HasMutInterior`) - #66822 (libunwind_panic: adjust miri panic hack) - #66827 (handle diverging functions forwarding their return place) - #66834 (rustbuild fixes) Failed merges: r? @ghost
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Addresses this comment.
Const-checking relied on
HasMutInteriorto forbid&mutin a const context. This was strange because all we needed to do was look for anRvalue::Refwith a certainBorrowKind, whereas theQualiftraits are specifically meant to get the qualifs for a value. This PR removes that logic fromHasMutInteriorand moves it intocheck_consts::Validator.As a result, we can now properly handle qualifications for
statics, which had to be ignored previously since you can e.g. borrow a staticCellfrom anotherstatic. We also remove thederived_from_illegal_borrowlogic, since it is no longer necessary; we give good errors for subsequent reborrows/borrows of illegal borrows.