MIR sanity check: validate types on assignment #72796
Merged
Conversation
This file contains 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
|
(rust_highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
|
However, something is wrong: this ICEs. Looks like something with consts does not work right here? |
RalfJung
force-pushed
the
mir-assign-sanity
branch
from
621f52a
to
8447da3
Compare
RalfJung
force-pushed
the
mir-assign-sanity
branch
from
8447da3
to
e34ea08
Compare
This comment has been minimized.
This comment has been minimized.
RalfJung
force-pushed
the
mir-assign-sanity
branch
from
3954b29
to
4c14454
Compare
RalfJung
commented
May 31, 2020
This comment has been minimized.
This comment has been minimized.
RalfJung
commented
May 31, 2020
|
Cc @lcnr (as const generics seem to be involved) |
|
Does this happen while building mir, if so, we probably didn't normalize the rhs yet. edit: you can also try running this with --verbose, which prints the actual debug repr instead of |
|
I am seeing this after all optimizations have run, so it's not during MIR building. |
|
@RalfJung You don't see this in miri because the way you do monomorphization also normalizes. The MIR type-checker that runs as part of borrow-checking has to normalize types in a few places in order to work, but what it does involves capturing all of the constraints on lifetimes resulting from normalizing and equating types. I guess you can probably just use |
eddyb
reviewed
May 31, 2020
eddyb
reviewed
May 31, 2020
| // Normalize both of them away. | ||
| // FIXME: Share this code with `interpret/eval_context.rs`. | ||
| let normalize = |ty: Ty<'tcx>| { | ||
| ty.fold_with(&mut BottomUpFolder { |
There was a problem hiding this comment.
Btw you can make this cheaper by implementing TypeFolder instead, which is only a few lines longer, and doing early returns in the type-visiting method based on flags (which are precomputed so they're basically free) that tell you if there are any erasable lifetimes or normalizable consts (for example, you could choose more things to do/look for).
Although you still have that reference "normalization" aspect, hmpf.
There was a problem hiding this comment.
I understood maybe half of that so I think I'll leave this to someone else...
eddyb
reviewed
May 31, 2020
RalfJung
changed the title
RalfJung
force-pushed
the
mir-assign-sanity
branch
from
1de3df7
to
294fb55
Compare
|
Now we get a different failure: The confusing part is that this looks like it's the same type on both sides.^^ |
|
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 |
|
running with |
|
@matthewjasper unfortunately, bootstrap fails to even build with those flags... gives |
|
I passed the flags by modifying Looks like vs |
lcnr
reviewed
Jun 1, 2020
This comment has been minimized.
This comment has been minimized.
|
☀️ Try build successful - checks-azure |
|
Queued 5e7db3c429dc362ffac291de89f7b045b674bfc9 with parent 1a4e2b6, future comparison URL. |
|
Finished benchmarking try commit (5e7db3c429dc362ffac291de89f7b045b674bfc9): comparison url. |
|
Ah lol, the extra check of layouts are really equal if the types are equal actually is measurable for the CTFE stress test. |
|
Okay @matthewjasper I think this is finally ready for review. |
matthewjasper
approved these changes
Jun 27, 2020
| "encountered `Assign` statement with overlapping memory", | ||
| ); | ||
| match &statement.kind { | ||
| StatementKind::Assign(box (dest, rvalue)) => { |
There was a problem hiding this comment.
As a follow up we should also check function calls and DropAndReplace
There was a problem hiding this comment.
SetDiscriminant can also get checked as to whether the type has the mentioned discriminant id and whether the local's type is actually an enum
There was a problem hiding this comment.
Let's collect these ideas in #73832
|
@bors r+ |
|
📌 Commit 35911ee has been approved by |
bors
added
S-waiting-on-bors
Dylan-DPC-zz
pushed a commit
to Dylan-DPC-zz/rust
that referenced
this pull request
Jun 28, 2020
…ewjasper MIR sanity check: validate types on assignment This expands the MIR validation added by @jonas-schievink in rust-lang#72093 to also check that on an assignment, the types of both sides match. Cc @eddyb @oli-obk
Manishearth
added a commit
to Manishearth/rust
that referenced
this pull request
Jun 28, 2020
…ewjasper MIR sanity check: validate types on assignment This expands the MIR validation added by @jonas-schievink in rust-lang#72093 to also check that on an assignment, the types of both sides match. Cc @eddyb @oli-obk
bors
added a commit
to rust-lang-ci/rust
that referenced
this pull request
Jun 28, 2020
…arth Rollup of 10 pull requests Successful merges: - rust-lang#72796 (MIR sanity check: validate types on assignment) - rust-lang#73243 (Add documentation to point to `File::open` or `OpenOptions::open` instead of `is_file` to check read/write possibility) - rust-lang#73525 (Prepare for LLVM 11) - rust-lang#73672 (Adds a clearer message for when the async keyword is missing from a f…) - rust-lang#73708 (Explain move errors that occur due to method calls involving `self` (take two)) - rust-lang#73758 (improper_ctypes: fix remaining `Reveal:All`) - rust-lang#73763 (errors: use `-Z terminal-width` in JSON emitter) - rust-lang#73796 (replace more `DefId`s with `LocalDefId`) - rust-lang#73797 (fix typo in self-profile.md) - rust-lang#73809 (Add links to fs::DirEntry::metadata) Failed merges: r? @ghost
|
This has broken the bootstrap with |
|
@jonas-schievink what exactly is the command for that again? Might be worth adding a comment in the file explaining this. |
|
I think it's |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This expands the MIR validation added by @jonas-schievink in #72093 to also check that on an assignment, the types of both sides match.
Cc @eddyb @oli-obk