A Go implementation of base100 with a permissive software license.

Base💯 can represent any byte with a unique emoji symbol, therefore it can represent binary data with zero printable overhead.

The API is nearly identical to other modules from the Go encoding/* standard library. See the Go Docs for more information.

A CLI tool is also provided for convenience, and ease of cross-compilation for multiple operating systems.

base💯 (Go)
Encodes things into emoji

USAGE:
    base100 [FLAGS]

FLAGS:
    -d, --decode     Decodes input
    -i, --input      Input file (default use STDIN)
    -o, --output     Output file (default use STDOUT)
    -h, --help       Prints help information

base100 will read from stdin unless a file is specified, will write UTF-8 to stdout, and has a similar API to GNU's base64. Data is encoded by default, unless --decode is specified

The implementation is fairly performant, and appears to perform roughly equivalent to the scalar Rust version on my machine. A future optimization utilizing SIMD/AVX could be possible with Go assembly code, however the throughput already far exceeds any known use case I can see for this, so I'll leave that out unless I get incredibly bored some day.

Library single-cpu benchmarks from my laptop (the throughput values are the relevant ones):

$ go test -bench=. -cpu=1
goos: darwin
goarch: arm64
pkg: github.com/mroth/base100-go
cpu: Apple M2 Pro
BenchmarkEncode                 24076875                49.24 ns/op      913.81 MB/s
BenchmarkEncodeToString         12909928                90.94 ns/op      494.83 MB/s
BenchmarkDecode                 31709400                37.85 ns/op     1189.02 MB/s
BenchmarkDecodeString           25889036                45.45 ns/op      990.11 MB/s
BenchmarkEncoder                   18339             65083 ns/op        1006.96 MB/s
BenchmarkDecoder                   21528             56465 ns/op        1160.65 MB/s