On Mon, Mar 31, 2025 at 11:47:42PM +0000, Thomas Mühlbacher wrote:
> Ok, so I have now tried to minimize the patch from the last patch set to
> at least have most of the CI (except for i686 builds) pass.
> 
> I threw out the pedantic automatic clippy fixes, so the `Bpos`
> instantiation change from before is gone. I am pretty sure that wouldn't
> have changed anything about the binary layout of that structure, though.

Err, do we actually have different types for Bpos on the C and Rust
side? Bpos and bpos? I may have to look over that code - I was concerned
about the potential for a bug there

> 
> I know you don't like formatting commits and I wish they wouldn't pick
> up anything except for what I changed. However, since `clippy --fix`
> doesn't format itself, I previously had that in there as part of a
> commit that only includes tool-generated changes.

I see :/

How much of a chore will it be to keep up with manually formatting
clippy fixes?

> 
> I also now just avoid the ugly transmute annotations and just suppress
> the lint instead for that function. This also requires changing the MSRV
> CI check, because that lint did not yet exist for the MSRV Rust version.
> 
> Thomas Mühlbacher (2):
>   fix(nix): allow msrv test to have warnings
>   refactor: make bch_bindgen pass clippy
> 
>  bch_bindgen/src/bcachefs.rs |  5 +++++
>  bch_bindgen/src/bkey.rs     |  3 ++-
>  bch_bindgen/src/btree.rs    | 25 +++++++++++++------------
>  bch_bindgen/src/fs.rs       |  2 +-
>  bch_bindgen/src/lib.rs      | 19 ++++++-------------
>  flake.nix                   |  2 +-
>  6 files changed, 28 insertions(+), 28 deletions(-)
> 
> -- 
> 2.48.1
> 

Reply via email to