build(xtask): use resolved cargo for clippy,doc invocation

[fogti]

It felt quite odd to use two completely different resolution mechanisms for cargo.

[fogti]

After rust-osdev/ovmf-prebuilt#239 (or such) is merged, it would be probably possible to get rid of the flate2 dependency of ureq entirely.

[mkroening]

I'm torn on this. The reason the sanitized cargo exists is that building the kernel and building example applications need to make sure to make Rustup's Cargo proxy.

The other uses of Cargo don't need this. Xshell was chosen for its scripting-like ease of use in the other cases.

The ureq change can be merged separately, right? Please keep it in alphabetical order and don't use tabular format.

[fogti]

Okay, I split the part which messes around with dependencies into #2132.

[github-actions]

Benchmark Results

Details

Benchmark	Current: 1c99f60	Previous: 52aa3f4	Performance Ratio
startup_benchmark Build Time	114.25 s	115.13 s	0.99 ❗
startup_benchmark File Size	0.87 MB	0.87 MB	1.00 ❗
Startup Time - 1 core	1.00 s (±0.02 s)	0.96 s (±0.02 s)	1.05 ❗
Startup Time - 2 cores	1.00 s (±0.03 s)	0.97 s (±0.02 s)	1.03
Startup Time - 4 cores	1.01 s (±0.02 s)	0.99 s (±0.04 s)	1.02
multithreaded_benchmark Build Time	117.50 s	117.46 s	1.00 ❗
multithreaded_benchmark File Size	0.97 MB	0.97 MB	1.00 ❗
Multithreaded Pi Efficiency - 2 Threads	92.45 % (±8.74 %)	89.90 % (±8.53 %)	1.03
Multithreaded Pi Efficiency - 4 Threads	45.24 % (±3.19 %)	44.03 % (±3.56 %)	1.03
Multithreaded Pi Efficiency - 8 Threads	25.75 % (±1.73 %)	25.43 % (±1.13 %)	1.01
micro_benchmarks Build Time	204.41 s	205.72 s	0.99 ❗
micro_benchmarks File Size	0.98 MB	0.98 MB	1.00 ❗
Scheduling time - 1 thread	73.88 ticks (±3.54 ticks)	75.99 ticks (±3.44 ticks)	0.97
Scheduling time - 2 threads	61.12 ticks (±17.78 ticks)	57.42 ticks (±19.04 ticks)	1.06
Micro - Time for syscall (getpid)	4.74 ticks (±0.31 ticks)	4.79 ticks (±0.32 ticks)	0.99
Memcpy speed - (built_in) block size 4096	56230.93 MByte/s (±39835.05 MByte/s)	55141.43 MByte/s (±39045.94 MByte/s)	1.02
Memcpy speed - (built_in) block size 1048576	29026.75 MByte/s (±24326.60 MByte/s)	28433.61 MByte/s (±23758.34 MByte/s)	1.02
Memcpy speed - (built_in) block size 16777216	22365.36 MByte/s (±18879.72 MByte/s)	26629.71 MByte/s (±22366.20 MByte/s)	0.84
Memset speed - (built_in) block size 4096	56996.95 MByte/s (±40393.72 MByte/s)	55434.09 MByte/s (±39238.17 MByte/s)	1.03
Memset speed - (built_in) block size 1048576	29461.13 MByte/s (±24488.91 MByte/s)	29247.17 MByte/s (±24233.73 MByte/s)	1.01
Memset speed - (built_in) block size 16777216	22654.79 MByte/s (±18988.25 MByte/s)	27017.06 MByte/s (±22507.35 MByte/s)	0.84
Memcpy speed - (rust) block size 4096	49801.08 MByte/s (±36415.58 MByte/s)	51653.42 MByte/s (±37465.88 MByte/s)	0.96
Memcpy speed - (rust) block size 1048576	28965.38 MByte/s (±24258.23 MByte/s)	28526.07 MByte/s (±23780.07 MByte/s)	1.02
Memcpy speed - (rust) block size 16777216	23319.72 MByte/s (±19751.48 MByte/s)	26406.45 MByte/s (±22136.69 MByte/s)	0.88
Memset speed - (rust) block size 4096	50467.51 MByte/s (±36869.50 MByte/s)	52383.36 MByte/s (±37932.30 MByte/s)	0.96
Memset speed - (rust) block size 1048576	29807.47 MByte/s (±24750.97 MByte/s)	29339.64 MByte/s (±24250.24 MByte/s)	1.02
Memset speed - (rust) block size 16777216	23600.27 MByte/s (±19851.94 MByte/s)	26877.78 MByte/s (±22333.36 MByte/s)	0.88
alloc_benchmarks Build Time	198.31 s	188.73 s	1.05 ❗
alloc_benchmarks File Size	0.94 MB	0.94 MB	1.00 ❗
Allocations - Allocation success	100.00 %	100.00 %	1
Allocations - Deallocation success	100.00 %	100.00 %	1
Allocations - Pre-fail Allocations	100.00 %	100.00 %	1
Allocations - Average Allocation time	9008.09 Ticks (±99.90 Ticks)	8899.94 Ticks (±109.88 Ticks)	1.01
Allocations - Average Allocation time (no fail)	9008.09 Ticks (±99.90 Ticks)	8899.94 Ticks (±109.88 Ticks)	1.01
Allocations - Average Deallocation time	1460.13 Ticks (±451.10 Ticks)	1937.45 Ticks (±495.96 Ticks)	0.75
mutex_benchmark Build Time	201.86 s	190.92 s	1.06 ❗
mutex_benchmark File Size	0.98 MB	0.98 MB	1.00 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	16.66 ns (±0.79 ns)	16.00 ns (±0.80 ns)	1.04
Mutex Stress Test Average Time per Iteration - 2 Threads	19.20 ns (±2.10 ns)	19.16 ns (±2.29 ns)	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[mkroening]

I have refactored the code a bit.

When testing, I noticed and remembered the reason I would rather not use the proxy explicitly for these development tasks:

cargo clippy --target=x86_64-unknown-none --all-targets

cd "/Users/mkroening/devel/hermit-rs/kernel" && env -u CARGO -u CARGO_MANIFEST_DIR -u CARGO_MANIFEST_PATH -u CARGO_PKG_AUTHORS -u CARGO_PKG_DESCRIPTION -u CARGO_PKG_HOMEPAGE -u CARGO_PKG_LICENSE -u CARGO_PKG_LICENSE_FILE -u CARGO_PKG_NAME -u CARGO_PKG_README -u CARGO_PKG_REPOSITORY -u CARGO_PKG_RUST_VERSION -u CARGO_PKG_VERSION -u CARGO_PKG_VERSION_MAJOR -u CARGO_PKG_VERSION_MINOR -u CARGO_PKG_VERSION_PATCH -u CARGO_PKG_VERSION_PRE -u LD_LIBRARY_PATH -u RUSTUP_TOOLCHAIN -u RUST_RECURSION_COUNT "/Users/mkroening/.cargo/bin/cargo" "clippy" "--all-targets" "--target=x86_64-unknown-none"

I strongly prefer the top for development purposes. I'll cherry-pick a few of these changes, but overall I think we can close this PR. Does this work for you? :)

[fogti]

Yes, this works for me.