Rewrite the improper_ctypes lint. #26583
Conversation
|
r? @pcwalton (rust_highfive has picked a reviewer for you, use r? to override) |
|
With some help on IRC, I got part of the way handling projection types (54e183f). Still doesn't work completely correctly, though; not sure what's missing. |
|
☔ The latest upstream changes (presumably #26575) made this pull request unmergeable. Please resolve the merge conflicts. |
eefriedman
force-pushed
the
lint-ffi
branch
2 times, most recently
from
060b6eb
to
f951898
Compare
|
Rebased, and updated with some handling for late-bound lifetimes. Ready to be reviewed, I think? |
|
☔ The latest upstream changes (presumably #26694) made this pull request unmergeable. Please resolve the merge conflicts. |
|
Rebased. |
|
Ping. |
|
r? @nrc |
| pub fn is_ffi_safe(&'tcx self, cx: &ctxt<'tcx>) -> bool { | ||
| !self.type_contents(cx).intersects(TC::ReachesFfiUnsafe) | ||
| } | ||
|
|
There was a problem hiding this comment.
Does this mean we no longer use ReachesFfiUnsafe at all? If so, it should be removed, not just this helper function.
eefriedman
force-pushed
the
lint-ffi
branch
3 times, most recently
from
dcc0cbb
to
484d071
Compare
|
Updated to address review comments. Function pointers are now checked. Tuples are now disallowed. Zero-size structs are now disallowed. |
Seems reasonable |
|
@bors r+ |
|
📌 Commit 484d071 has been approved by |
|
⌛ Testing commit 484d071 with merge d7e2f6f... |
|
📌 Commit 8d47c3d has been approved by |
|
⌛ Testing commit 8d47c3d with merge 3825fc1... |
|
💔 Test failed - auto-mac-64-nopt-t |
|
Fixed again... |
|
@bors r+ |
|
📌 Commit 858a1b9 has been approved by |
|
⌛ Testing commit 858a1b9 with merge 48ef346... |
|
💔 Test failed - auto-win-msvc-64-opt |
Makes the lint a bit more accurate, and improves the quality of the diagnostic messages by explicitly returning an error message. The new lint is also a little more aggressive: specifically, it now rejects tuples, and it recurses into function pointers.
|
Fixed again... |
Makes the lint a bit more accurate, and improves the quality of the diagnostic messages by explicitly returning an error message.
ghost
mentioned this pull request
These warnings starting appearing from rust-lang/rust#26583.
These warnings starting appearing from rust-lang/rust#26583.
This warning starting appearing from rust-lang/rust#26583.
…enums This dates back to at least rust-lang#26583. At the time, usize and isize were considered ffi-unsafe to nudge people away from them, but this changed in the aforementioned PR, making it inconsistent to complain about it in enum discriminants. In fact, repr(usize) is probably the best way to interface with `enum Foo : size_t { ... }`.
Makes the lint a bit more accurate, and improves the quality of the diagnostic
messages by explicitly returning an error message.