Add wrong_any_coerce lint
#3376
Conversation
HMPerson1
force-pushed
the
wrong_any_coerce
branch
from
22ad055
to
ca7408b
Compare
| src_ty: Ty<'tcx>, | ||
| tgt_ty: Ty<'tcx>, | ||
| ) -> Option<LintData<'tcx>> { | ||
| // redo the typechecking for this coercion to see if it required unsizing something to `dyn Any` |
There was a problem hiding this comment.
What's the reason for this? Would it be possible to use https://doc.rust-lang.org/nightly/nightly-rustc/rustc/ty/struct.TyCtxt.html#method.struct_lockstep_tails to figure out whether the src_ty already is Any at the same level that tgt_ty is?
There was a problem hiding this comment.
struct_lockstep_tails doesn't go through references and Box<T> ends up as PhantomData<T>, so we'd have to do some additional work to use that. That also wouldn't work if the type is generic (i.e. T: Deref<Target=dyn Any>).
There was a problem hiding this comment.
ugh, makes sense :/
Do you think we can share this function with rustc by refactoring the original rustc code to factor out this function and make it public?
There was a problem hiding this comment.
Something like resolve_obligations_for_unsize_coercion(.., obligation_callback: impl FnMut(&PredicateObligation) -> ()) -> ..? Then typechecking could use the callback to check for unsizing to a tuple (which apparently is feature gated) and we could check for unsizing to dyn Any?
HMPerson1
force-pushed
the
wrong_any_coerce
branch
from
babfbde
to
0fcd88d
Compare
phansch
added
the
S-waiting-on-review
|
Ping @HMPerson1. I'm going over old PRs, that were abandoned by us reviewers (sorry for that!) or by the authors. Are you still interested in completing this? |
HMPerson1
force-pushed
the
wrong_any_coerce
branch
from
151e966
to
8e9c8f6
Compare
| return true; | ||
| } | ||
| } | ||
| // TODO: check for `RefCell<dyn Any>`? |
There was a problem hiding this comment.
Do you want to do this in this PR? Otherwise you can put this in the Known Problems section.
flip1995
added
S-waiting-on-author
|
☔ The latest upstream changes (presumably #4259) made this pull request unmergeable. Please resolve the merge conflicts. |
|
Hi @HMPerson1, since this PR has been inactive for a while, I'm going to go ahead and close it. Please don't hesitate to re-open it/ask for help. (As far as I can tell, the last status was that there were some remaining |
This lints any coercion to
dyn Anywhere the source type derefs todyn Any(e.g.Box<dyn Any>,Rc<dyn Any>, etc.). We might also want to lintRefCell<dyn Any>orMutex<dyn Any>, but I don't know of a good way to check for things like that without just hardcoding a list of "container-ish things".Fixes #2988