feat: basic support for trap EXIT

[reubeno]

Adds just enough support for basic trap EXIT handlers to get invoked.

The save/restore of last exit status isn't ideal, but gets the test to pass for now. We should consider coming back and looking more seriously at whether there's a more general pattern there.

[github-actions]

Test Results

    3 files     23 suites   6m 18s ⏱️
1 123 tests 1 123 ✅ 0 💤 0 ❌
3 349 runs  3 349 ✅ 0 💤 0 ❌

Results for commit ea54ae6.

[github-actions]

Public API changes for crate: brush-core

Added items

+pub async fn brush_core::Shell::on_exit(&mut self) -> core::result::Result<(), brush_core::error::Error>

Performance Benchmark Report

Benchmark name	Baseline (μs)	Test/PR (μs)	Delta (μs)	Delta %
clone_shell_object	17.44 μs	17.74 μs	0.30 μs	🟠 +1.74%
eval_arithmetic	0.14 μs	0.17 μs	0.02 μs	🟠 +14.58%
expand_one_string	1.67 μs	1.70 μs	0.03 μs	⚪ Unchanged
for_loop	20.92 μs	20.90 μs	-0.02 μs	⚪ Unchanged
function_call	2.15 μs	2.20 μs	0.05 μs	⚪ Unchanged
instantiate_shell	57.98 μs	57.93 μs	-0.05 μs	⚪ Unchanged
instantiate_shell_with_init_scripts	24385.66 μs	24792.42 μs	406.76 μs	🟠 +1.67%
parse_bash_completion	2024.41 μs	2021.03 μs	-3.38 μs	⚪ Unchanged
parse_sample_script	1.93 μs	1.99 μs	0.06 μs	⚪ Unchanged
run_echo_builtin_command	15.50 μs	15.77 μs	0.27 μs	⚪ Unchanged
run_one_external_command	2346.42 μs	2291.00 μs	-55.42 μs	🟢 -2.36%
tokenize_sample_script	3.48 μs	3.48 μs	0.01 μs	⚪ Unchanged

Code Coverage Report: Only Changed Files listed

Package	Base Coverage	New Coverage	Difference
brush-core/src/shell.rs	🟢 75.67%	🟢 76.25%	🟢 0.58%
brush-interactive/src/interactive_shell.rs	🟠 63.49%	🟠 63.78%	🟢 0.29%
Overall Coverage	🟢 71.83%	🟢 71.86%	🟢 0.03%

Minimum allowed coverage is 70%, this run produced 71.86%

Test Summary: bash-completion test suite

Outcome	Count	Percentage
✅ Pass	1351	64.06
❗️ Error	212	10.05
❌ Fail	177	8.39
⏩ Skip	354	16.79
❎ Expected Fail	14	0.66
✔️ Unexpected Pass	1	0.05
📊 Total	2109	100.00