fix: suppress host-only cuTile lints

[efegokdemir]

Summary

#[cutile::module] emits host-side Rust for device kernels, where lints such as unused variables and unreachable statements do not reflect device execution. Allow these lints within generated modules so strict host builds do not fail on device-only code.

Changes

• Add the host-only lint allowances to both generated module forms.
• Add a strict-lint integration regression case for a print followed by device code.

Testing

• cargo fmt --all --check
• git diff --check
• cargo test -p cutile --test basics_and_inlining --no-run could not reach compilation: cuda-bindings requires a CUDA 13.0+ toolkit, which is not installed here.

Checklist

• All commits signed off (git commit -s)
• SPDX headers on new source files (no new source files added)

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[GiGiKoneti]

Hi @efegokdemir & @elibol! 👋

Just wanted to link my original PR (NVlabs/cutile-rs#321, tracked in #1442). My patch adds the full set of inner-module lint allowances (unused_variables, unreachable_code, unused_mut, nonstandard_style, dead_code) to both macro expansion branches (entry_functions.is_empty() and non-empty) in cutile-macro.

Feel free to pull or cherry-pick those missing allowances into this PR if helpful so all host warning paths are covered!

Signed-off-by: GiGiKoneti <gigikoneti@gmail.com>