remove superseded lints

[bend-n]

changelog: [transmute_float_to_int, transmute_int_to_char, transmute_int_to_float]: remove lints, now in rustc

these lints are now mostly in rustc, so they dont need to be in clippy anymore

rust-lang/rust#136083 (comment)

pending rust-lang/rust#140431: transmute_int_to_bool

Summary Notes

• "Rust version of new lints should be checked" by @samueltardieu

Generated by triagebot, see help for how to add more

[rustbot]

r? @llogiq

rustbot has assigned @llogiq.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

Failed to set assignee to lcnr: invalid assignee

Note: Only org members with at least the repository "read" role, users with write permissions, or people who have commented on the PR may be assigned.

[Jarcho]

When uplifting lints please run cargo dev rename_lint <lint_name> <uplifted_name> --uplift. This will add all the metadata related to lint renames.

[samueltardieu]

@bend-n Anything I can help with in order to merge this (I noticed it's still in draft)? Could you run cargo dev fmt and squash your commits into one?

r? @samueltardieu

[bend-n]

@samueltardieu
this PR is pending rust-lang/rust#140431 with regards to keeping transmute_int_to_bool. if that PR is closed, ill remove transmute_int_to_bool here.

[samueltardieu]

Any reason to want to put this into the same PR, while there are two PRs on the compiler side? If your second compiler PR is not merged before tomorrow, then it won't be in Rust 1.88 while your first compiler PR would be. It means that, in Clippy, we would have to backport only the deprecation of transmute_float_to_int, transmute_int_to_char, and transmute_int_to_float, while the deprecation of transmute_int_to_bool would stay in nightly as the Rust side would land only in Rust 1.89.

This is why having two PRs, to match the compiler changes, might make things easier to keep Clippy lints deprecation in sync.

[bend-n]

well if this gets merged today then transmute_int_to_bool will be in both clippy and rustc in 1.88, is that what we want?

[samueltardieu]

Yes. But if it is not, please do not add the transmute_int_to_bool deprecation to this PR but open another one.

Marking this PR as blocked in the meantime and adding a note for myself when it comes to reviewing.

@rustbot note "Rust version of new lints should be checked"
@rustbot blocked

[samueltardieu]

It looks like the unnecessary_transmutes compiler builtin lint does not lint for signed integer to float and vice-versa, while this was linted by Clippy. Removing the lints will lose functionality. Maybe the lints should be modified instead of being uplifted to handle only those cases involving signed integers.

The same thing is true for transmuting a i32 to a char: there is no compiler warning there, so the corresponding lint should probably be kept in Clippy for this case.

Also, it looks like this patch removes the code for the transmute_num_to_bytes lint, as this is also linted by the compiler, but does not remove it properly nor does it talk about it in the PR changelog. It looks like this is the only Clippy lint which is completely covered by the unnecessary_transmutes compiler lint, so probably the only one which should be uplifted.