feat: reduce allocations

[Elsie19]

Some of the things I did:

• Initialize with with_capacity where applicable
• Delay or remove vector initialization
• Replace individual collection pushes to one extend
• Replace some function arguments with IntoIterator bounds for looser constraints
• Use Cow when appropriate
• Replace Vec in function arguments with &[]
• Replace map_or_else with map_or when default value is static

I also replaced these functions entirely:

- uppercase_first_char
- lowercase_first_char
+ pattern_to_first_char

and

- uppercase_pattern
- lowercase_pattern
+ pattern_to_string

Where the new functions take a closure over their respective argument and can transform it at the call site. You may want to pull out that specific commit and merge it outside of this PR, but I added it here anyways.

Feel free to pick and choose which ones to keep and leave! If you need I can merge likewise commits into one to make the merge cleaner.

[github-actions]

Public API changes for crate: brush-core

Changed items

-pub const fn brush_core::commands::SimpleCommand<'a, SE>::new(shell: brush_core::commands::ShellForCommand<'a, SE>, params: brush_core::ExecutionParameters, command_name: alloc::string::String, args: alloc::vec::Vec<brush_core::commands::CommandArg>) -> Self
+pub fn brush_core::commands::SimpleCommand<'a, SE>::new<I>(shell: brush_core::commands::ShellForCommand<'a, SE>, params: brush_core::ExecutionParameters, command_name: alloc::string::String, args: I) -> Self where I: core::iter::traits::collect::IntoIterator<Item = brush_core::commands::CommandArg>
-pub async fn brush_core::Shell<SE>::invoke_function<N: core::convert::AsRef<str>, I: core::iter::traits::collect::IntoIterator<Item = A>, A: core::convert::AsRef<str>>(&mut self, name: N, args: I, params: &brush_core::ExecutionParameters) -> core::result::Result<u8, brush_core::error::Error>
+pub async fn brush_core::Shell<SE>::invoke_function<N: core::convert::AsRef<str>, I: core::iter::traits::collect::IntoIterator<Item = A>, A: core::convert::AsRef<str>>(&mut self, name: N, args: I, params: brush_core::ExecutionParameters) -> core::result::Result<u8, brush_core::error::Error>

Performance Benchmark Report

Benchmark name	Baseline (μs)	Test/PR (μs)	Delta (μs)	Delta %
clone_shell_object	13.54 μs	13.68 μs	0.15 μs	🟠 +1.09%
eval_arithmetic	0.12 μs	0.12 μs	0.00 μs	⚪ Unchanged
expand_one_string	1.22 μs	1.20 μs	-0.02 μs	⚪ Unchanged
for_loop	24.57 μs	23.95 μs	-0.62 μs	🟢 -2.52%
full_peg_complex	43.41 μs	43.16 μs	-0.25 μs	⚪ Unchanged
full_peg_for_loop	4.83 μs	4.79 μs	-0.04 μs	🟢 -0.91%
full_peg_nested_expansions	12.92 μs	12.70 μs	-0.22 μs	⚪ Unchanged
full_peg_pipeline	3.23 μs	3.24 μs	0.00 μs	⚪ Unchanged
full_peg_simple	1.32 μs	1.35 μs	0.03 μs	🟠 +2.05%
function_call	2.63 μs	2.68 μs	0.05 μs	⚪ Unchanged
instantiate_shell	43.27 μs	43.09 μs	-0.18 μs	⚪ Unchanged
instantiate_shell_with_init_scripts	21119.40 μs	21290.80 μs	171.40 μs	⚪ Unchanged
parse_peg_bash_completion	1614.43 μs	1619.21 μs	4.78 μs	⚪ Unchanged
parse_peg_complex	14.70 μs	14.71 μs	0.00 μs	⚪ Unchanged
parse_peg_for_loop	1.51 μs	1.45 μs	-0.06 μs	🟢 -4.04%
parse_peg_pipeline	1.56 μs	1.47 μs	-0.09 μs	🟢 -5.52%
parse_peg_simple	0.80 μs	0.78 μs	-0.02 μs	🟢 -2.62%
run_echo_builtin_command	13.27 μs	13.36 μs	0.09 μs	⚪ Unchanged
tokenize_sample_script	2.85 μs	2.75 μs	-0.10 μs	🟢 -3.48%

Code Coverage Report: Only Changed Files listed

Package	Base Coverage	New Coverage	Difference
brush-builtins\src\bind.rs	🔴 25.26%	🔴 25.17%	🔴 -0.09%
brush-builtins\src\history.rs	🔴 7.81%	🔴 7.87%	🟢 0.06%
brush-builtins\src\read.rs	🔴 38.24%	🔴 37.93%	🔴 -0.31%
brush-core\src\completion.rs	🔴 14.54%	🔴 14.58%	🟢 0.04%
brush-core\src\expansion.rs	🔴 28.89%	🔴 29.4%	🟢 0.51%
brush-core\src\variables.rs	🔴 10.31%	🔴 10.34%	🟢 0.03%
brush-shell\src\config.rs	🟢 87.85%	🟢 88.26%	🟢 0.41%
Overall Coverage	🟢 25.8%	🟢 25.84%	🟢 0.04%

Minimum allowed coverage is 20%, this run produced 25.84%

Test Summary: bash-completion test suite

Outcome	Count	Percentage
✅ Pass	1581	74.96
❗️ Error	19	0.90
❌ Fail	155	7.35
⏩ Skip	339	16.07
❎ Expected Fail	13	0.62
✔️ Unexpected Pass	2	0.09
📊 Total	2109	100.00

[lu-zero]

I read it twice and it looks very nice.

[lu-zero]

Probably we could cache them in the future

[reubeno]

Thanks for putting this all togther, @Elsie19!

I've read through everything other than expansion.rs so far -- and it's all looking pretty good. Your changes have left the code cleaner, more idiomatic, and definitely avoided some unneeded allocations 😄.

There are 2 API surface changes flagged by the PR -- I think both of them are warranted, and they're relatively low impact, so I'm supportive of making them.

I'll spend some more time looking through expansion.rs, particularly given that the expand_one_string microbenchmark noted a possible perf change.

[lu-zero]

It does crash with #1069 so once it lands I'd re-roll that branch and if we have time would be nice to land it as follow up.

[reubeno]

It does crash with #1069 so once it lands I'd re-roll that branch and if we have time would be nice to land it as follow up.

@lu-zero Any more info on what the crash is? Is there anything you'd recommend get adjusted in this PR before we merge? I just finished reviewing the latest changes (for this PR) and it's looking pretty good to me.

[lu-zero]

Just I managed to typoing clash with crash... It works but on the portage-repo it seems to not speed up anything :/