Bump the parser submodule: StringLiteral source spans

[revarbat]

Picks up openscad_cpp_parser#3.

A StringLiteral's position spanned only its closing quote, and so did any node wrapping one — a PositionalArgument, a vector element — because a string is matched by several lexer rules and YY_USER_ACTION resets tokenStart on each of them.

a = "txt";        span was '"'   now '"txt"'
f("x,y", b);      argument span '"'   now '"x,y"'
x = ["a,b","c"];  element spans '"'   now '"a,b"' / '"c"'

Why it mattered

parse_ast() consumers slicing source by those offsets could land inside string content. BelfrySCAD's reformatter rewrote a comma inside "h,l,height,length" in BOSL2's isosurface.scad, silently changing the string's value. It needed a guard skipping any argument list with an unbalanced quote span — this bump makes that guard unnecessary.

Verification

• 1399/1399 tests pass
• Spans confirmed correct through the Python parse_ast API, including escapes, empty strings and multi-line strings
• Against BelfrySCAD: the separator-spacing pass now covers 74 of 97 corpus files (up from 73 — the extras are exactly the ones its quote guard had been skipping), with 0 structural changes
• BelfrySCAD's 653 tests and the 145-case differential sweep against real OpenSCAD are unchanged

Minor bump (0.17.0 → 0.18.0): node positions change, which anything using them will see.

🤖 Generated with Claude Code