Per Monarch docs, the proper way to make a tokenizer case-insensitive
is `ignoreCase: true` at the language level — flags on individual regex
literals appear to be ignored by Monarch's parser. Switched the algol68
mode and the converter (`-i`) to emit `ignoreCase: true`, dropped the
redundant `/i` suffixes.
Also fixed a CodeQL warning in tm-to-monarch's regex literal escaping:
the previous one-liner `re.replace(/\//g, '\\/')` didn't account for
backslashes, so inputs containing `\/` would emit malformed literals.
Replaced with a char-by-char walk that preserves escape sequences and
only escapes truly unescaped slashes; warns on stray trailing backslash.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
6ac53d1cdd