README.md (3482B)
1 # base58 2 3 [](https://hackage.haskell.org/package/ppad-base58) 4  5 [](https://docs.ppad.tech/base58) 6 7 A pure Haskell implementation of base58 and base58check encoding & 8 decoding on strict ByteStrings. 9 10 ## Usage 11 12 A sample GHCi session: 13 14 ``` 15 > :set -XOverloadedStrings 16 > 17 > -- import qualified 18 > import qualified Data.ByteString.Base58 as B58 19 > import qualified Data.ByteString.Base58Check as B58Check 20 > 21 > -- simple base58 encoding and decoding 22 > let b58 = B58.encode "hello world" 23 > b58 24 "StV1DL6CwTryKyV" 25 > 26 > B58.decode b58 27 Just "hello world" 28 > 29 > -- base58check is a versioned, checksummed format 30 > let b58check = B58Check.encode "\NULhello world" -- 0x00 version byte 31 > b58check 32 "13vQB7B6MrGQZaxCqW9KER" 33 > 34 > B58Check.decode b58check 35 Just "\NULhello world" 36 ``` 37 38 ## Documentation 39 40 Haddocks (API documentation, etc.) are hosted at 41 [docs.ppad.tech/base58](https://docs.ppad.tech/base58). 42 43 ## Performance 44 45 The aim is best-in-class performance for pure, highly-auditable Haskell 46 code. 47 48 Current benchmark figures on a M4 Silicon MacBook Air look like (use 49 `cabal bench` to run the benchmark suite): 50 51 ``` 52 benchmarking ppad-base58/base58/encode/hello world 53 time 350.3 ns (349.2 ns .. 352.0 ns) 54 0.999 R² (0.997 R² .. 1.000 R²) 55 mean 353.0 ns (351.9 ns .. 354.9 ns) 56 std dev 4.705 ns (3.411 ns .. 7.449 ns) 57 variance introduced by outliers: 13% (moderately inflated) 58 59 benchmarking ppad-base58/base58/decode/StV1DL6CwTryKyV 60 time 377.7 ns (373.7 ns .. 380.7 ns) 61 0.999 R² (0.999 R² .. 1.000 R²) 62 mean 379.4 ns (376.8 ns .. 381.8 ns) 63 std dev 9.020 ns (7.823 ns .. 10.76 ns) 64 variance introduced by outliers: 32% (moderately inflated) 65 66 benchmarking ppad-base58/base58check/encode/0x00, hello world 67 time 1.197 μs (1.195 μs .. 1.200 μs) 68 1.000 R² (1.000 R² .. 1.000 R²) 69 mean 1.200 μs (1.198 μs .. 1.202 μs) 70 std dev 7.256 ns (6.187 ns .. 8.819 ns) 71 72 benchmarking ppad-base58/base58check/decode/13vQB7B6MrGQZaxCqW9KER 73 time 1.222 μs (1.218 μs .. 1.225 μs) 74 1.000 R² (1.000 R² .. 1.000 R²) 75 mean 1.220 μs (1.217 μs .. 1.223 μs) 76 std dev 9.566 ns (8.250 ns .. 11.28 ns) 77 ``` 78 79 You should build with the 'llvm' flag for maximum performance. 80 81 ## Security 82 83 This library aims at the maximum security achievable in a 84 garbage-collected language under an optimizing compiler such as GHC, in 85 which strict constant-timeness can be challenging to achieve. 86 87 If you discover any vulnerabilities, please disclose them via 88 security@ppad.tech. 89 90 ## Development 91 92 You'll require [Nix][nixos] with [flake][flake] support enabled. Enter a 93 development shell with: 94 95 ``` 96 $ nix develop 97 ``` 98 99 Then do e.g.: 100 101 ``` 102 $ cabal repl ppad-base58 103 ``` 104 105 to get a REPL for the main library. 106 107 ## Attribution 108 109 The vectors used in the test suite for both base58 110 and base58check are verbatim from paulmillr's 111 [scure-base](https://github.com/paulmillr/scure-base) library. 112 113 [nixos]: https://nixos.org/ 114 [flake]: https://nixos.org/manual/nix/unstable/command-ref/new-cli/nix3-flake.html