Enable report_in_external_macro in unaligned_references #82587
Conversation
rust-highfive
added
the
S-waiting-on-review
| = note: fields of packed structs are not properly aligned, and creating a misaligned reference is undefined behavior (even if that reference is never dereferenced) | ||
| = note: this error originates in a macro (in Nightly builds, run with -Z macro-backtrace for more info) | ||
|
|
||
| warning: borrow of packed field is unsafe and requires unsafe function or block (error E0133) |
There was a problem hiding this comment.
Can you allow this lint, just to make sure we focus on the one we really care about here?
There was a problem hiding this comment.
That makes sense. Updated to allow safe_packed_borrows.
|
The change makes sense at first sight, but I don't feel comfortable approving this since I have no idea if there are any side-effects to using this flag. Cc @rust-lang/compiler -- does anyone know what |
taiki-e
force-pushed
the
unaligned_references
branch
from
8f16211 to
62b4b8d
Compare
|
Technically all lint changes need to be approved by the lang team. I'm going ahead and accept this however, since the lint is allow-by-default so if it is denied, the user must be expecting it to occur. This can theoretically be a breaking change in a bin crate that uses a macro that (uselessly) denies this right now. @bors r+ rollup=never (for bisection) |
|
📌 Commit 62b4b8d has been approved by |
bors
added
S-waiting-on-bors
|
Furthermore, when/if this becomes a hard error, it will certainly also be a hard error when the code originates in an external macro.
FWIW, I am hoping to change this soon with #82525. |
|
☀️ Test successful - checks-actions |
Fixes an issue where
unaligned_referencesis not triggered in external macros, unlikesafe_packed_borrows.Also, given that this lint is planned to eventually change to hard error (#82525), it would make sense for this lint to be triggered for external macros as well.
See taiki-e/pin-project-lite#55 (comment) and taiki-e/pin-project-lite#55 (comment) for more.
r? @RalfJung