nx_lex.nx

nx_lex.nx -- canonical lexer primitives (source-code whitespace). Distinct from nx_ascii.nx by design: nx_ascii follows POSIX isspace() / isalpha() semantics (broad, including VT and FF for whitespace). nx_lex provides STRICT source-code-oriented variants used by parsers that need precise control over what counts as whitespace. CONSOLIDATION TRAIL (verified via nx_func_compare 2026-05-13): nx_lex_is_ws (4-char: space, tab, LF, CR) replaces 9 byte-identical adapter-local variants: - nx_coq_is_ws (hash 5493497725180460624) - nx_extract_is_ws (hash 5493497725180460624) - nx_hol_is_ws (hash 5493497725180460624) - nx_isa_is_ws (hash 5493497725180460624) - nx_lean_is_ws (hash 5493497725180460624) - nx_mizar_is_ws (hash 5493497725180460624) - nx_mm_is_ws (hash 5493497725180460624) - nx_json_is_ws (semantically equivalent; hex-vs-decimal literals) Why distinct from nx_ascii_is_space: nx_ascii_is_space also accepts 0x0B (vertical tab) and 0x0C (form feed); adapter is_ws was strict to the 4 source-code chars. Preserving the strict semantics here. nx_lex_is_intraline_ws (2-char: space, tab) replaces: - nx_kv_is_ws (record-separator parsers must NOT treat newline as whitespace) genealogy_id: ascii_1968_standard + posix_isspace_traditions + source_code_lexer_conventions lineage_id: character_class_predicate axioms: NX_AX_LOGIC_EXCLUDED_MIDDLE (each c is or is not ws, no third state)