v3.1

detextive v3.1 (2026-02-17)

Enhancements

• API: Add Behaviors.utf_16_32_requires_byte_order to optionally reject
  BOM-less trial decodes for generic utf-16 and utf-32 while keeping
  permissive behavior as the default.

Repairs

• API: Report UTF charset results from BOM provenance rather than decode codec
  choice so utf-8-sig is returned only when a UTF-8 BOM is present, and
  apply the same provenance normalization to UTF-16/UTF-32 reporting across
  decode, detection, and inference surfaces.

v3.0

detextive v3.0 (2026-02-14)

Enhancements

• API: Add decode_inform to return decoded text together with charset, MIME
  type, and line-separator metadata in a single call.
• API: Honor supplied textual http_content_type metadata consistently across
  decode and inference paths, including header-guided charset trial decode.

Removals

• API: Remove charset_default, mimetype_default, and
  mimetype_supplement parameters from decode so decoding follows
  decode-or-error semantics instead of fallback-return inference semantics.
• API: Replace Behaviors.charset_detect and Behaviors.mimetype_detect
  tristates with booleans; pass True or False instead of
  BehaviorTristate values.

Repairs

• Fix UTF-8 content incorrectly decoded when charset detector misidentifies encoding, causing mojibake with non-ASCII characters and emoji.
• Fix malformed http_content_type parameter parsing so inference no longer
  raises raw ValueError for invalid header parameter syntax.
  Also include the resolved MIME type value in TextualMimetypeInvalidity
  messages.
• Reject binary content with non-textual MIME types instead of attempting to decode, preventing false positives where binary data was incorrectly decoded as text.

v1.0

Detextive v1.0 (2025-08-12)

Enhancements

• Provide LineSeparators enum with detection, normalization, and nativization
  methods.
• Provide detect_charset, detect_mimetype,
  detect_charset_and_mimetype, is_textual_mimetype, and
  is_textual_content.