Version 3.5.1

• Widen fromUtf8 to accept any List<int> input instead of requiring a
  Uint8List. Non-Uint8List inputs are copied to a Uint8List before
  decoding.

Full Changelog: v3.5.0...v3.5.1

Version 3.5.0

• [Breaking Changes]
  • Renames internal abstract class HashlibConverter to CipherlibConverter.
  • Removes encodeString, decodeString from IterableCodec.
• Fix CryptDataBuilder.param to throw an ArgumentError when the value is
  null, instead of a raw TypeError.
• Speed up Base-64 and Base-32 encoding with specialized single-pass encoders.
  Base-64 encoding now outperforms dart:convert.
• Speed up Base-64 and Base-32 decoding with specialized single-pass decoders.
  Base-64 decoding now outperforms dart:convert.
• Speed up toUtf8 using Uint8List buffer instead of plain List<int>.
  UTF-8 encoding now outperforms dart:convert.
• Speed up fromUtf8 with a dedicated byte-to-string decoder.
  UTF-8 decoding now matches closely with dart:convert.
• Refactor existing benchmarks and add missing ones. Dropping benchmark_harness
  dependency in favor of custom benchmarking class with better flexibility.

Full Changelog: v3.4.0...v3.5.0

Version 3.4.0

• Rename the package from hashlib_codecs to convertlib. The public API is
  unchanged: update the dependency to convertlib and imports to
  package:convertlib/convertlib.dart.
• The hashlib_codecs package continues to be published as a thin re-export of
  convertlib for backward compatibility, and will be discontinued after a
  migration period.

Full Changelog: v3.3.1...v3.4.0

Version 3.3.1

• Fix CryptData.validate rejecting valid Modular Crypt Format hashes such as
  bcrypt, whose base64 alphabet uses .. The hash field now accepts the
  same characters as the salt ([a-zA-Z0-9/+.-]), reverting the overly
  strict B64 restriction introduced in 3.3.0.
• Document every public API element (100% dartdoc coverage) and correct several
  inaccurate doc comments, including the Base8Codec alphabet (01234567) and
  the octal/binary notes on toOctal.

Full Changelog: v3.3.0...v3.3.1

Version 3.3.0

• [Breaking Changes]
  • ByteCollector.isEqual requires an Iterable argument to match the full
    length; previously a strict prefix of the bytes compared equal.
  • CryptData.validate now follows the PHC string format specification
    strictly: the hash must be a B64 string ([a-zA-Z0-9/+], no . or -),
    and the version must not have leading zeros.
  • Reject invalid values after padding in AlphabetDecoder. Affected codecs are Base32 and Base64.
• ByteCollector.isEqual now compares content in constant time, making it safe
  for comparing MACs and message digests
• Performance improvements (measured on Apple Silicon, 10KB inputs):
  • Base-32 decoding: ~6x faster (425 Mbps -> 2.52 Gbps)
  • Base-64 decoding: ~6.4x faster (453 Mbps -> 2.9 Gbps)
  • UTF-8 encoding: ~6.5x faster (403 Mbps -> 2.61 Gbps)
• Update the PHC string format specification link to its new home at C2SP

Full Changelog: v3.2.0...v3.3.0

Version 3.2.0

• [Breaking Changes]
  • Fix Base32Codec.wordSafe to use the documented word-safe alphabet
    (23456789CFGHJMPQRVWXcfghjmpqrvwx); it previously used the z-base-32 alphabet.
  • ByteCollector equality is now value-based: == and hashCode compare the
    collected bytes instead of the identity of the underlying list.
• Fix UTF-8 decoder producing wrong code points for some 2-byte sequences (e.g. U+0100)
• Fix ByteCollector.isEqual comparing against its own buffer when given a ByteBuffer
• Fix ByteCollector.isEqual ignoring the offset and length of a partial TypedData view
• Fix incorrect alphabets of crockford and geohash in the README
• Fix incorrect parameter descriptions of toBase32, fromBase32, fromBase64, and fromUtf8
• Document that toUtf8 rejects unpaired surrogates

Full Changelog: v3.1.2...v3.2.0

Version 3.1.2

• Add const constructor to ByteCollector class.

Full Changelog: v3.1.1...v3.1.2

Version 3.1.1

• Add const constructor to ByteCollector class.

Full Changelog: v3.1.0...v3.1.1

Version 3.1.0

• Introduce ByteCollector abstract class with methods for various encoding formats (hex, binary, octal, base32, base64, BigInt).
  and utility methods for equality check.

Full Changelog: v3.0.1...v3.1.0

Version 3.0.1

• Test release with workflow

Full Changelog: v3.0.0...v3.0.1