4.0.2

Fixes

• Fixes #59 - Base32's Encode(ulong) and DecodeUInt64() works consistently among platforms with different endianness

4.0.1

Fixes

• Fixes #58 - Encode(long) failing -- reported by Pascal Schwarz <@pschwarzpp>

4.0.0

Breaking changes

• This version is built with .NET 6 SDK.
• Benchmark now uses BenchmarkDotNet.
• Changed interface names from Encoder to Coder to signify encoding and
  decoding functionality better.
• Removed obsolete methods.
• Simple (aka allocating) versions of Decode() will now return byte[]'s instead of Span<byte>'s for correct
  ownership semantics. It's even possible that some copying may be avoided in certain scenarios.
• Base16.TryDecode() doesn't throw on invalid input, but returns false instead.
• Base32.Decode() throws separate exceptions for encountered failures.

New features

• Added Bech32 flavor to Base32
• Added RFC 1924 (IPv6) flavor to Base85 along with
  EncodeIpv6 and DecodeIpv6 functions https://tools.ietf.org/html/rfc1924
• Added Base58.Bitcoin.EncodeCheck() and Base58.Bitcoin.TryDecodeCheck() methods.
• Added Base58.Bitcoin.EncodeCb58() and Base58.Bitcoin.TryDecodeCb58() methods.

Improvements

• Added more buffer overflow detection to Base32 coder
• Removed all unsafe code. New Span-based optimizations make the code come close to unsafe perf.
• Removed slow and hard to read optimizations like bit shift operations for multiplication and division
  where compiler almost always does a better job of optimizing.

Fixes

• Fixed output buffer was too small error for certain Base58 cases.
• Avoid redundant memory copy operations

3.1.0

Improvements

• Added FileCoin flavor to Base32 (contributed by @farukterzioglu).
• Upgraded test and benchmark projects to .NET Core 3.1 (LTS).

3.0.2

Fixes

• Fixed output buffer was too small error for certain Base58 cases.
• Avoid redundant memory copy operations.
• Introduce an additional buffer overrun check for Base32.

3.0.1

Fixes

• Fixed padding being ignored by Base32, thanks to @skwasjer!

2.1.0

Improvements

• Produce native .NET 4.7.2 binaries which hopefully removes the scaffolding overhead

3.0.0

Breaking changes

• This version only supports .NET Standard 2.1 (.NET Core 3.0) and up, because I started using nullable annotations and Span/Range constructs in the code. It's time to move to the future, folks!
• Base16 encoding/decoding methods have been moved to Base16.UpperCase and Base16.LowerCase respectively.
• Changed the static version of Base16.Decode() to receive a string as a parameter to avoid signature conflicts. (It's probably more practical this way)
• Removed string/byte[] overloads for Encode functions as they are redundant with .NET Core 3.0. Only Base16.Decode remained as a string due to function signature conflicts.
• Base32 is no longer whitespace tolerant.

New features

• New TryEncode/TryDecode interfaces for non-allocating encoding. Probably would be useful in massive number of encodings.
• Yubico's ModHex support for Base16.

Improvements

• More than 2x faster Base16 decoding
• Faster Base16 encoding
• More test coverage
• No more dependency to System.Runtime.Numerics
• Encoders now conform to common interfaces like IBaseEncoder, IBaseStreamEncoder, INonAllocatingBaseEncoder which make them pluggable, replacable with other implementations.
• Nullable reference annotations

Fixes

• Fix package license expression.
• Fix regression in Base16 optimizations.

2.0.0

Breaking changes

• Adapted to nullable references
• Encoder references are now properties instead of static fields

New stuff

• Async variants of Stream-based encoding/decoding methods
• New Base32 variants: z-base-32 and Geohash
• XML documentation is included in the package now
• SimpleBase is now on GitHub Package Repository

Improvements

• Base58 decoding is 100+% faster and encoding is slightly faster, too
• Reduced startup overhead
• Updated dependencies to newer packages
• Removed redundant null checking code
• Base58's Ripple and Flickr variants are now being tested

1.7.1

• Minor code analysis fixes