Address lint ambiguous_wide_pointer_comparisons

[jamesmunns]

Noticed this warning:

warning: ambiguous wide pointer comparison, the comparison includes metadata which may not be expected
   --> /Users/james/personal/mycelium/cordyceps/src/list.rs:497:12
    |
497 |         if head == tail {
    |            ^^^^^^^^^^^^
    |
    = note: `#[warn(ambiguous_wide_pointer_comparisons)]` on by default
help: use `std::ptr::addr_eq` or untyped pointers to only compare their addresses
    |
497 |         if std::ptr::addr_eq(head.as_ptr(), tail.as_ptr()) {
    |            ++++++++++++++++++    ~~~~~~~~~~     ++++++++++
help: use explicit `std::ptr::eq` method to compare metadata and addresses
    |
497 |         if std::ptr::eq(head.as_ptr(), tail.as_ptr()) {
    |            +++++++++++++    ~~~~~~~~~~     ++++++++++

warning: ambiguous wide pointer comparison, the comparison includes metadata which may not be expected
   --> /Users/james/personal/mycelium/cordyceps/src/list.rs:966:13
    |
966 |             splice_start == splice_end
    |             ^^^^^^^^^^^^^^^^^^^^^^^^^^
    |
help: use `std::ptr::addr_eq` or untyped pointers to only compare their addresses
    |
966 |             std::ptr::addr_eq(splice_start.as_ptr(), splice_end.as_ptr())
    |             ++++++++++++++++++            ~~~~~~~~~~           ++++++++++
help: use explicit `std::ptr::eq` method to compare metadata and addresses
    |
966 |             std::ptr::eq(splice_start.as_ptr(), splice_end.as_ptr())
    |             +++++++++++++            ~~~~~~~~~~           ++++++++++
    ```

[hawkw]

This is a MSRV bump, since core::ptr::addr_eq was added in Rust 1.76.0, and cordyceps currently claims a minimum Rust version of 1.61.0.

I'm not super fussed about taking the MSRV bump, but if we do, it would be good to make sure to change the Cargo.toml rust-version field as part of this commit, as well.

Alternatively, we could perform the same address-without-metadata comparison by writing

head.as_ptr() as *const () == tail.as_ptr() as *const ()

and

splice_start.as_ptr() as *const () == splice_end.as_ptr() as *const ()

which is basically what ptr::addr_eq does internally, but is permitted on older Rust versions. If we go that route, we can avoid the MSRV bump, but it might be worth adding a comment that the as *const () casts are to ignore the wide pointer metadata, so that a future reader doesn't change them back...

[hawkw]

Ugh, it looks like some of the maitake tests are not feature flagged correctly: https://github.com/hawkw/mycelium/actions/runs/9821097757/job/27116399920?pr=478

I can fix that separately...

[jamesmunns]

Updated MSRV of cordyceps, and all crates that I could see that depend on cordyceps.

[hawkw]

Some smallish nits. I'd like to merge this, though!

[hawkw]

I don't actually think we need to do the MSRV bump for maitake-sync. Since the cordyceps changes will be released in a semver-compatible version, and maitake-sync's minimum cordyceps version isn't changing, users of maitake-sync can continue using the current cordyceps version rather than picking up the one with the new MSRV --- either by checking in a lockfile, or by explicitly pinning the cordyceps version in their Cargo.toml.

IMHO, we would only need to change maitake-sync's MSRV if we were changing the dependency spec for cordyceps to not permit versions that compile on 1.61.0.

[hawkw]

nit, take it or leave it: core::ptr is already imported in this file:

mycelium/cordyceps/src/list.rs

Line 13 in ed9e49b

ptr::{self, NonNull},

so we could drop the core prefix:

Suggested change

	if core::ptr::addr_eq(head.as_ptr(), tail.as_ptr()) {
	if ptr::addr_eq(head.as_ptr(), tail.as_ptr()) {

[hawkw]

similarly, we could write:

Suggested change

	core::ptr::addr_eq(splice_start.as_ptr(), splice_end.as_ptr())
	ptr::addr_eq(splice_start.as_ptr(), splice_end.as_ptr())

[hawkw]

again, i think the MSRV bump isn't strictly necessary unless we are also changing the cordyceps dep

[hawkw]

and again, i don't think this is really necessary...although it's also fine since this isn't a published crate.