fix(parser): a case pattern's ) does not end a command substitution

[bresilla]

$(case ...) doesn't parse:

$ brush -c 'x=$(case a in a) echo Y;; esac); echo "[$x]"'
error: -c: syntax error at end of input

bash, dash and ksh all print [Y].

A case pattern ends in ), and two places read that as the ) that closes the substitution. consume_nested_construct in tokenizer.rs counts parens, so it stops at $(case a in a. The command() rule in word.rs stops at the first ) for the same reason.

The tokenizer now tracks where it is in a case (case / in / ;; / esac), so a pattern's ) closes the pattern. The ( pattern ) form's leading ( doesn't open anything either. In the word grammar I added a rule that consumes a whole case ... esac before the generic one sees it; it recurses, so nested cases work.

Checked against bash: both pattern forms, several arms, nested case, nested $(), subshells, backticks, $(( )), and a quoted ). All 225 brush-parser tests pass.

Fixes #1052

[Copilot]

Pull request overview

Fixes #1052 by distinguishing case-pattern delimiters from command-substitution terminators.

Changes:

• Adds tokenizer state tracking for case constructs.
• Adds recursive case ... esac handling in command substitutions.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File	Description
brush-parser/src/word.rs	Consumes case blocks within substitutions.
brush-parser/src/tokenizer.rs	Tracks case parsing state around parentheses.

Suppressed comments (1)

brush-parser/src/tokenizer.rs:672

• Token spelling does not establish reserved-word context. In the valid $(case x in case) echo Y;; esac) form, the pattern word case pushes AwaitingIn; its following ) is then again mistaken for the substitution terminator. Only a case keyword in command position should push a new state.

                    "case" => case_states.push(CaseState::AwaitingIn),

[github-actions]

Performance Benchmark Report

Benchmark name	Baseline (μs)	Test/PR (μs)	Delta (μs)	Delta %
clone_shell_object	17.71 μs	17.85 μs	0.14 μs	⚪ Unchanged
eval_arithmetic	0.15 μs	0.15 μs	0.00 μs	⚪ Unchanged
expand_one_string	1.72 μs	1.67 μs	-0.05 μs	⚪ Unchanged
for_loop	31.68 μs	32.61 μs	0.93 μs	🟠 +2.94%
full_peg_complex	57.14 μs	56.08 μs	-1.06 μs	🟢 -1.86%
full_peg_for_loop	6.21 μs	6.09 μs	-0.12 μs	🟢 -1.95%
full_peg_nested_expansions	15.93 μs	15.97 μs	0.04 μs	⚪ Unchanged
full_peg_pipeline	4.18 μs	4.20 μs	0.02 μs	⚪ Unchanged
full_peg_simple	1.78 μs	1.81 μs	0.03 μs	🟠 +1.46%
function_call	3.46 μs	3.53 μs	0.06 μs	⚪ Unchanged
instantiate_shell	55.55 μs	55.11 μs	-0.43 μs	⚪ Unchanged
instantiate_shell_with_init_scripts	27202.90 μs	27323.28 μs	120.39 μs	⚪ Unchanged
parse_peg_bash_completion	2097.52 μs	2088.09 μs	-9.43 μs	⚪ Unchanged
parse_peg_complex	20.59 μs	20.38 μs	-0.21 μs	🟢 -1.02%
parse_peg_for_loop	2.12 μs	2.01 μs	-0.11 μs	🟢 -5.24%
parse_peg_pipeline	2.08 μs	2.09 μs	0.00 μs	⚪ Unchanged
parse_peg_simple	1.08 μs	1.13 μs	0.05 μs	🟠 +4.44%
run_echo_builtin_command	16.60 μs	16.15 μs	-0.45 μs	⚪ Unchanged
tokenize_sample_script	3.44 μs	3.34 μs	-0.10 μs	🟢 -2.79%

Code Coverage Report: Only Changed Files listed

Package	Base Coverage	New Coverage	Difference
brush-parser/src/tokenizer.rs	🟢 93.53%	🟢 93.67%	🟢 0.14%
brush-parser/src/word.rs	🟢 93.76%	🟢 93.55%	🔴 -0.21%
Overall Coverage	🟢 76.12%	🟢 76.14%	🟢 0.02%

Minimum allowed coverage is 70%, this run produced 76.14%
Maximum allowed coverage difference is -5%, this run produced 0.02%

Test Summary: bash-completion test suite

Outcome	Count	Percentage
✅ Pass	1581	74.96
❗️ Error	18	0.85
❌ Fail	156	7.40
⏩ Skip	339	16.07
❎ Expected Fail	13	0.62
✔️ Unexpected Pass	2	0.09
📊 Total	2109	100.00