refactor: improve benchmark compile output

[secretfader]

Currently, several of our benchmarks output warnings concerning unused imports in addition to the benchmark data we requested. This is unaffected by the --quiet flag, and in at least one case, we may be compiling more dependency code than necessary.

This PR resolves the above issues through use of Rust compiler flags.

[codecov]

Codecov Report

Merging #445 (f4d5b1e) into master (522a04f) will not change coverage.
The diff coverage is n/a.

❗ Current head f4d5b1e differs from pull request most recent head 68f0e79. Consider uploading reports for the commit 68f0e79 to get more accurate results

@@           Coverage Diff           @@
##           master     #445   +/-   ##
=======================================
  Coverage   75.15%   75.15%           
=======================================
  Files          83       83           
  Lines        9800     9800           
=======================================
  Hits         7365     7365           
  Misses       2435     2435           

Impacted Files	Coverage Δ
pallets/allocations/src/benchmarking.rs	100.00% <ø> (ø)
pallets/allocations/src/lib.rs	82.97% <ø> (ø)
pallets/amendments/src/benchmarking.rs	100.00% <ø> (ø)
pallets/emergency-shutdown/src/benchmarking.rs	100.00% <ø> (ø)
pallets/grants/src/benchmarking.rs	100.00% <ø> (ø)
pallets/reserve/src/benchmarking.rs	100.00% <ø> (ø)
pallets/root-of-trust/src/benchmarking.rs	100.00% <ø> (ø)
pallets/staking/src/benchmarking.rs	95.31% <ø> (ø)
pallets/tcr/src/benchmarking.rs	100.00% <ø> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 522a04f...68f0e79. Read the comment docs.

[secretfader]

This branch should be executed with the following command and return no errors in a clean environment.

cargo run --release --features runtime-benchmarks --manifest-path=node/Cargo.toml -- benchmark --pallet="frame_system" --extrinsic="*"

Note the inclusion of --release; this ensures we're not compiling debug or test code (which should make the runs quicker), and enable production optimizations (to place our results closer to real-world conditions).

[secretfader]

I resolved the errors when running cargo test --all --all-features by allowing unused imports in the benchmarking module specifically.

I'll merge after the test run passes.