nx_string_ops.nx

nx_string_ops.nx -- byte-buffer string primitives (Python/JS stdlib analog). Every program ends up needing: substring search, slicing, length, trim, parse-int, equality, hash, split. Without a substrate-level home for these, every caller hand-rolls a while-loop scanning bytes -- the exact kind of duplicated bullshit the user named in the "loops within loops" feedback. PRIMITIVES (16 total): Family 1: scan + length nx_str_len(buf) -- length until first 0 byte nx_str_index_of(buf, n, ch) -- first index of byte ch, -1 if none nx_str_last_index_of(buf, n, ch) -- last index of ch, -1 if none Family 2: equality / prefix / suffix nx_str_equals(a, na, b, nb) nx_str_starts_with(buf, n, prefix, np) nx_str_ends_with(buf, n, suffix, ns) nx_str_find(haystack, nh, needle, nn) -- first occurrence, -1 if none Family 3: slice + copy nx_str_slice_copy(src, lo, hi, dst, cap) -- copy [lo, hi) into dst nx_str_trim_left(buf, n) -- returns new start offset nx_str_trim_right(buf, n) -- returns new end offset (exclusive) Family 4: parse + format nx_str_parse_int(buf, n) -- parse signed decimal nx_str_format_int(v, dst, cap) -- write decimal; returns length Family 5: classify nx_str_is_empty(buf, n) nx_str_is_all_digits(buf, n) nx_str_is_all_whitespace(buf, n) Family 6: hash nx_str_hash_fnv1a(buf, n) -- 64-bit FNV-1a, well-distributed DESIGN PRINCIPLES: - All take explicit length n; no null-terminator assumption (bytes in buffers may contain 0)