v1.2.0

Added

• Add NameMode and NameMode::apply to set how names are displayed in an IDB's pseudocode.
• Add HaruspexError::IdaConfig, returned when an IDA configuration directive can't be applied.
• Add the garble_names test binary and integration tests for names that aren't valid C identifiers.

Changed

• Update idalib to v0.11.1, for IDB::modify_ida_config.
• Make sanitize_filename an allowlist of letters and digits of any script, printable ASCII characters, and spaces
  (except reserved characters), so that Unicode format characters that IDA 9.5 keeps in names (e.g., bidirectional
  isolates and zero-width spaces) no longer reach output file names and printed paths. Other non-ASCII characters that
  aren't letters or digits (e.g., emoji) are now replaced too.
• Mark the linker_messages suppression as permanent.
• Update other dependencies.
• Update documentation.

Fixed

• Garble names that IDA 9.5 wraps in backticks in the pseudocode written by the haruspex CLI (e.g., #memcpy becomes
  _memcpy, as on IDA 9.4), so static analysis tools can parse it again. Library users need to call
  NameMode::Garbled.apply themselves. Some demangled Rust names (e.g., trait implementations such as
  <char as core::fmt::Display>::fmt) and types (e.g., &str) still keep their backticks.