jouleclaw-backend-apple: 84 clippy warnings -> 0, and -D warnings now covers both configurations #437

Merged

dcharlot merged 2 commits from fix/backend-apple-lints into main 2026-11-01 07:42:04 -05:00

Conversation 0 Commits 2 Files changed 5 +100 -92

2026-10-22

Claude

jouleclaw-backend-apple: 0 warnings, and the gate covers both configurations ...

10 -> 0, and the `-D warnings` step now lints loader-gguf under
`rayon fabric-meter` as well as on default, plus backend-apple on its
own. That closes the per-configuration lint hole #373 opened and #414
halved.

The last ten, all in mps_matmul.rs:

- `struct PipelineCell` with its `unsafe impl Send`/`Sync` was never
  constructed and mentioned nowhere else -- scaffolding for a OnceLock
  wrapper the pipelines ended up doing directly. Removed, and the space
  reused for a `KeepAlive` alias, which is what the two `type_complexity`
  warnings wanted: the inline form is three generic layers deep and
  appeared twice.
- `DeviceKvCache` had a public `len` and no `is_empty`.
- three `let qk_pipe = qk_pipe;` no-op rebindings from a refactor.
- `too_many_arguments` and `needless_range_loop` allowed crate-wide, the
  same two decisions omni and loader-gguf already record and for the same
  reasons: a GPU dispatch signature is its kernel's parameter list, and
  these kernels walk several buffers at one index.

The comment this replaces in standards-build.yml claimed the fabric-meter
configuration could not be gated because backend-apple's warnings "are
unused in THAT configuration and needed in others, so they want cfg work
rather than deletion". Measuring the crate directly refuted both halves:
it has no `[features]` at all, so there is one configuration, and the
count was 84 rather than the 43 that leaked through loader-gguf's lint.

Verified: all ten lint specs in the step pass, 37 backend-apple tests
(32 pass, 5 ignored), 343 loader-gguf tests, 77 omni tests.

Co-Authored-By: Claude Opus 5 <[email protected]>

19904251b7

• View at this point in history

2026-10-22 06:58:23 -04:00

Claude

jouleclaw-backend-apple: 84 clippy warnings -> 10 ...

#414 left this crate as the reason the `-D warnings` gate covers
loader-gguf on default features only: clippy reports it through the path
dependency, so gating the `fabric-meter` configuration needs it clean too.

I described its warnings in #414 as "unused in THAT configuration and
needed in others, so they want cfg work rather than deletion". That was
wrong on two counts, and measuring it directly settled both:

- the crate has NO `[features]` at all, only a macOS-conditional
  dependency block, so there is one configuration and the unused items
  are unconditionally unused;
- there are 84 of them, not 43. The 43 was what leaked through
  loader-gguf's lint, not the crate's own count.

82 of the 84 are in mps_matmul.rs.

41 came from clippy's own suggestions (manual_is_multiple_of 23,
unnecessary_cast 8, needless_option_as_deref 3, excessive_precision,
doc_lazy_continuation, double_parens, chunks_exact_to_as_chunks).

The other 42 -- 27 `unused_unsafe` and 15 `unused_imports` -- carry no
machine-applicable suggestion at all. Neither `clippy --fix` nor
`cargo fix` touches them, with or without `--all-targets`. They are
plain code, not macro expansions, so they were fixed line-guided from
rustc's own line numbers:

- `|| unsafe { .. }` -> `|| { .. }` where objc2 0.6 made the call safe.
  Dropping a block rustc has already proved redundant cannot change
  behaviour.
- `use objc2_metal::{MTLComputePipelineState as _, MTLLibrary as _}` ->
  just the name the function needs. The unused one is a trait brought in
  as `_`; the type itself is referenced fully qualified on the lines
  above, which is why it reads as used at a glance.

Checked before deleting, because this is the shape that broke #372: each
unused import was matched against every other mention of the name in its
file and against the file's six `cfg(not(target_os = "macos"))` branches.
None appears in one, and objc2-metal is a macOS-only dependency, so no
other configuration can want them.

Verified: 0 errors, 37 backend-apple tests (32 pass, 5 ignored), and
loader-gguf still builds and its fabric-meter `process_calibration`
passes.

Co-Authored-By: Claude Opus 5 <[email protected]>

1c10b364a5

• View at this point in history

2026-10-22 06:55:59 -04:00