`x.py bench --stage 0 src/libcore` fails with deprecation errors

[AaronKutch]

I would like to benchmark a specific subset of all the benchmarks, but adding src/libcore onto the command makes it fail:

$ ./x.py bench --stage 0 src/libcore
Updating only changed submodules
Submodules updated in 0.59 seconds
    Finished dev [unoptimized] target(s) in 0.37s
Building stage0 std artifacts (x86_64-pc-windows-gnu -> x86_64-pc-windows-gnu)
    Finished release [optimized] target(s) in 4.00s
Copying stage0 std from stage0 (x86_64-pc-windows-gnu -> x86_64-pc-windows-gnu / x86_64-pc-windows-gnu)
Building stage0 test artifacts (x86_64-pc-windows-gnu -> x86_64-pc-windows-gnu)
    Finished release [optimized] target(s) in 0.56s
Copying stage0 test from stage0 (x86_64-pc-windows-gnu -> x86_64-pc-windows-gnu / x86_64-pc-windows-gnu)
Benchmarking core stage0 (x86_64-pc-windows-gnu -> x86_64-pc-windows-gnu)
   Compiling core v0.0.0 (C:\Users\Allen\Documents\GitHub\rust\src\libcore)
error: `...` range patterns are deprecated
   --> src\libcore\../libcore/benches\ascii.rs:194:21
    |
194 |                 b'a'...b'z' => true,
    |                     ^^^ help: use `..=` for an inclusive range
    |
    = note: `-D ellipsis-inclusive-range-patterns` implied by `-D rust-2018-idioms`

error: `...` range patterns are deprecated
   --> src\libcore\../libcore/benches\ascii.rs:206:21
    |
206 |                 b'a'...b'z' => true,
    |                     ^^^ help: use `..=` for an inclusive range

error: `...` range patterns are deprecated
   --> src\libcore\../libcore/benches\ascii.rs:218:21
    |
218 |                 b'a'...b'z' => true,
    |                     ^^^ help: use `..=` for an inclusive range

error: `...` range patterns are deprecated
   --> src\libcore\../libcore/benches\ascii.rs:230:21
    |
230 |                 b'a'...b'z' => true,
    |                     ^^^ help: use `..=` for an inclusive range

error: `...` range patterns are deprecated
   --> src\libcore\../libcore/benches\ascii.rs:194:21
    |
194 |                 b'a'...b'z' => true,
    |                     ^^^ help: use `..=` for an inclusive range

error: aborting due to 5 previous errors

error: Could not compile `core`.

To learn more, run the command again with --verbose.


command did not execute successfully: "C:\\Users\\Allen\\Documents\\GitHub\\rust\\build\\x86_64-pc-windows-gnu\\stage0\\bin\\cargo.exe" "bench" "--target" "x86_64-pc-windows-gnu" "-j" "4" "--features" "panic-unwind backtrace compiler-builtins-c" "--manifest-path" "C:\\Users\\Allen\\Documents\\GitHub\\rust\\src/libstd/Cargo.toml" "-p" "core" "--" "--quiet"
expected success, got: exit code: 101


        finished in 8.642

1 command(s) did not execute successfully:

  - "C:\\Users\\Allen\\Documents\\GitHub\\rust\\build\\x86_64-pc-windows-gnu\\stage0\\bin\\cargo.exe" "bench" "--target" "x86_64-pc-windows-gnu" "-j" "4" "--features" "panic-unwind backtrace compiler-builtins-c" "--manifest-path" "C:\\Users\\Allen\\Documents\\GitHub\\rust\\src/libstd/Cargo.toml" "-p" "core" "--" "--quiet"

[AaronKutch]

Also, I should mention that this occurs even after running x.py clean

[Centril]

@AaronKutch Could you file a PR with ..= instead of ...?

[ehuss]

I was curious why this hadn't been noticed earlier. Apparently benchmarks aren't tested on CI to verify that they still build. This libcore benchmark has been broken since #60133. I'm quite surprised nobody has run x.py bench src/libcore since then.

Another wrinkle that confused me is that ./x.py test --stage=0 --no-doc src/libcore also fails. Why does this command build benchmarks? Well this little snippet makes --no-doc also enable benchmarks. I suspect that was a mistake, or maybe the mistake is that benchmarks aren't always tested?

The libs team may want to consider possibly testing benchmarks in CI to ensure that they stay functional. libcore at least only takes a few seconds to build and run its benchmarks in test mode (one iteration).