edition 2024 and raise MSRV to 1.88 #731
Conversation
|
down to "only" fixing 32bit Windows unsafety |
workingjubilee
force-pushed
the
edition-2024-cleanup
branch
2 times, most recently
from
96ee90e to
21c02bd
Compare
|
note re: timing, it is possible another MSRV 1.82 release should land first. |
workingjubilee
force-pushed
the
edition-2024-cleanup
branch
2 times, most recently
from
88970de to
eb7f722
Compare
workingjubilee
changed the title
workingjubilee
force-pushed
the
edition-2024-cleanup
branch
from
eb7f722 to
f141372
Compare
|
Could you rebase on master? Github's diff seems to be confused. |
This allows using let-chains in the code, but requires a repo-wide fmt.
workingjubilee
force-pushed
the
edition-2024-cleanup
branch
from
f141372 to
7d52b2a
Compare
|
Done! It probably had trouble identifying the commits as being the same contentfully. |
workingjubilee
force-pushed
the
edition-2024-cleanup
branch
from
f3e305c to
3ae4456
Compare
| // SAFETY: Name must be initialized by SymFromAddrW, but we only interact with it as a pointer anyways. | ||
| let name_ptr = unsafe { (&raw const (*info).Name).cast::<u16>() }; |
There was a problem hiding this comment.
I removed this unsafe block during my post-push editing of #737 because I thought the compiler said it was superfluous, but now it is required for reasons that aren't clear to me.
There was a problem hiding this comment.
Yeah I wondered about that but I trusted the compiler to complain if it's not needed, ha.
There was a problem hiding this comment.
I may need to poke at this a bit more if it's an edition-based unsafety difference: that shouldn't really be the case. It could have just been human confusion on my end though.
There was a problem hiding this comment.
...I identified the diff in question removed deny(unsafe_op_in_unsafe_fn) as well as the unsafe block, so there is no edition-based problem aside from the lint being enabled or not.
With 1.90 being released, the stable - 2 policy now means 1.88 is fair game and we can use let-chains in the code. As we have lots of parsing code, this is a significant improvement. However, first, we have to also update to edition 2024. I have left out let-chain improvements from this PR to make it tractable to review.
Due to the somewhat adventuresome nature of updating the edition, this also includes cleaning up the win32 implementation a fair deal. Mostly this is the obvious unsafe-wrapping. That itself has been mostly split out to #735, modulo one unsafe block that turned out to be required after #737.