0.2.0 — Mojo 1.0

• Builds on released Mojo 1.0. The pixi channel moved from
  max-nightly to max and the dependency pin from >=1.0.0b3.dev… to
  >=1.0,<2; CI installs mojo==1.0.0 from PyPI instead of the nightly
  index with --prerelease allow. The weekly drift check still runs
  against the latest nightly, so the repo still reports when the language
  moves under it. No source changes were needed.

• README figures re-measured on 1.0: 70 → 102 tests (the errors module
  added a fifth test file), 5,400 fuzz mutations with zero crashes and
  zero hangs, and a live corpus run where all 137 feeds that fetched
  parsed without a raise, crash, or hang. The throughput range is
  restated for the machine it was re-measured on — b3 and 1.0 benchmark
  the same there, so the move is the machine, not the compiler.

• New feed.errors module (exported from the package): line_col(source, offset) maps a byte offset to a 1-based (line, column) pair — the column
  is the 1-based BYTE offset within the line, no UTF-8 decoding — and
  parse_error(msg, source, offset) builds an Error reading
  <msg> at line <L>, column <C>: '<snippet>', where the snippet is up to
  ~30 bytes of the offending line centered on the column,
  whitespace-trimmed, with ... where truncated, and never multi-line.
  This is the error-reporting pattern shared with mojo-xml.

• Parse errors now carry that position + snippet wherever a byte offset
  exists at the raise site: every strict-mode XML error (previously a bare
  (line L, column C) suffix with no snippet), the structural XML errors
  both modes raise (unterminated constructs / start tags / attributes /
  attribute values, unquoted attribute values, malformed start/end tags,
  empty element names), every JSON Feed syntax error, and the positioned
  date-parsing errors (offsets relative to the date string).

• No mechanism change: parsers still raise Error(...), no new error
  types, and existing contains=-style message checks keep matching.