perf(runner): collapse detect_runtime_error error scan

[Chemaclass]

Context

bashunit::runner::detect_runtime_error (src/runner.sh:93-114) loops over 23 error strings, running case on each iteration. For test outputs without errors (the common case) it walks the full list.

Proposal

Collapse to a single case with glob alternation, short-circuit early:

function bashunit::runner::detect_runtime_error() {
  local runtime_output=\$1
  case \"\$runtime_output\" in
    *\"command not found\"* | *\"unbound variable\"* | *\"permission denied\"* | \\
    *\"no such file or directory\"* | *\"syntax error\"* | *\"bad substitution\"* | \\
    *\"division by 0\"* | *\"cannot allocate memory\"* | *\"bad file descriptor\"* | \\
    *\"segmentation fault\"* | *\"illegal option\"* | *\"argument list too long\"* | \\
    *\"readonly variable\"* | *\"missing keyword\"* | *killed* | \\
    *\"cannot execute binary file\"* | *\"invalid arithmetic operator\"* | \\
    *\"ambiguous redirect\"* | *\"integer expression expected\"* | \\
    *\"too many arguments\"* | *\"value too great\"* | \\
    *\"not a valid identifier\"* | *\"unexpected EOF\"*)
      local runtime_error=\"\${runtime_output#*: }\"
      printf '%s' \"\${runtime_error//\$'\\n'/}\"
      return 0
      ;;
  esac
  printf ''
}

Single pattern match, no Bash loop, no per-iteration case setup.

Acceptance

• Single case matches all error strings
• Existing tests covering runtime error detection green
• No behavioral change in error string extraction
• Bash 3.0+ compat preserved
• Benchmark on error-free output paths reported in PR