1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
//! # Configs
//!
//! There isn't just one type of Base64; that would be too simple. You need to choose a character
//! set (standard, URL-safe, etc) and padding suffix (yes/no).
//! The `Config` struct encapsulates this info. There are some common configs included: `STANDARD`,
//! `URL_SAFE`, etc. You can also make your own `Config` if needed.
//!
//! The functions that don't have `config` in the name (e.g. `encode()` and `decode()`) use the
//! `STANDARD` config .
//!
//! The functions that write to a slice (the ones that end in `_slice`) are generally the fastest
//! because they don't need to resize anything. If it fits in your workflow and you care about
//! performance, keep using the same buffer (growing as need be) and use the `_slice` methods for
//! the best performance.
//!
//! # Encoding
//!
//! Several different encoding functions are available to you depending on your desire for
//! convenience vs performance.
//!
//! | Function                | Output                       | Allocates                      |
//! | ----------------------- | ---------------------------- | ------------------------------ |
//! | `encode`                | Returns a new `String`       | Always                         |
//! | `encode_config`         | Returns a new `String`       | Always                         |
//! | `encode_config_buf`     | Appends to provided `String` | Only if `String` needs to grow |
//! | `encode_config_slice`   | Writes to provided `&[u8]`   | Never                          |
//!
//! All of the encoding functions that take a `Config` will pad as per the config.
//!
//! # Decoding
//!
//! Just as for encoding, there are different decoding functions available.
//!
//! | Function                | Output                        | Allocates                      |
//! | ----------------------- | ----------------------------- | ------------------------------ |
//! | `decode`                | Returns a new `Vec<u8>`       | Always                         |
//! | `decode_config`         | Returns a new `Vec<u8>`       | Always                         |
//! | `decode_config_buf`     | Appends to provided `Vec<u8>` | Only if `Vec` needs to grow    |
//! | `decode_config_slice`   | Writes to provided `&[u8]`    | Never                          |
//!
//! Unlike encoding, where all possible input is valid, decoding can fail (see `DecodeError`).
//!
//! Input can be invalid because it has invalid characters or invalid padding. (No padding at all is
//! valid, but excess padding is not.) Whitespace in the input is invalid.
//!
//! # Panics
//!
//! If length calculations result in overflowing `usize`, a panic will result.
//!
//! The `_slice` flavors of encode or decode will panic if the provided output slice is too small,

#![cfg_attr(feature = "cargo-clippy", allow(clippy::cast_lossless))]
#![deny(
    missing_docs,
    trivial_casts,
    trivial_numeric_casts,
    unused_extern_crates,
    unused_import_braces,
    unused_results,
    variant_size_differences,
    warnings
)]
#![forbid(unsafe_code)]
#![cfg_attr(not(any(feature = "std", test)), no_std)]

#[cfg(all(feature = "alloc", not(any(feature = "std", test))))]
extern crate alloc;
#[cfg(any(feature = "std", test))]
extern crate std as alloc;

#[cfg(test)]
#[macro_use]
extern crate doc_comment;

#[cfg(test)]
doctest!("../README.md");

mod chunked_encoder;
pub mod display;
mod tables;
#[cfg(any(feature = "std", test))]
pub mod write;

mod encode;
pub use crate::encode::encode_config_slice;
#[cfg(any(feature = "alloc", feature = "std", test))]
pub use crate::encode::{encode, encode_config, encode_config_buf};

mod decode;
#[cfg(any(feature = "alloc", feature = "std", test))]
pub use crate::decode::{decode, decode_config, decode_config_buf};
pub use crate::decode::{decode_config_slice, DecodeError};

#[cfg(test)]
mod tests;

/// Available encoding character sets
#[derive(Clone, Copy, Debug)]
pub enum CharacterSet {
    /// The standard character set (uses `+` and `/`).
    ///
    /// See [RFC 3548](https://tools.ietf.org/html/rfc3548#section-3).
    Standard,
    /// The URL safe character set (uses `-` and `_`).
    ///
    /// See [RFC 3548](https://tools.ietf.org/html/rfc3548#section-4).
    UrlSafe,
    /// The `crypt(3)` character set (uses `./0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz`).
    ///
    /// Not standardized, but folk wisdom on the net asserts that this alphabet is what crypt uses.
    Crypt,
}

impl CharacterSet {
    fn encode_table(self) -> &'static [u8; 64] {
        match self {
            CharacterSet::Standard => tables::STANDARD_ENCODE,
            CharacterSet::UrlSafe => tables::URL_SAFE_ENCODE,
            CharacterSet::Crypt => tables::CRYPT_ENCODE,
        }
    }

    fn decode_table(self) -> &'static [u8; 256] {
        match self {
            CharacterSet::Standard => tables::STANDARD_DECODE,
            CharacterSet::UrlSafe => tables::URL_SAFE_DECODE,
            CharacterSet::Crypt => tables::CRYPT_DECODE,
        }
    }
}

/// Contains configuration parameters for base64 encoding
#[derive(Clone, Copy, Debug)]
pub struct Config {
    /// Character set to use
    char_set: CharacterSet,
    /// True to pad output with `=` characters
    pad: bool,
    /// True to ignore excess nonzero bits in the last few symbols, otherwise an error is returned.
    decode_allow_trailing_bits: bool,
}

impl Config {
    /// Create a new `Config`.
    pub fn new(char_set: CharacterSet, pad: bool) -> Config {
        Config {
            char_set,
            pad,
            decode_allow_trailing_bits: false,
        }
    }

    /// Sets whether to pad output with `=` characters.
    pub fn pad(self, pad: bool) -> Config {
        Config { pad, ..self }
    }

    /// Sets whether to emit errors for nonzero trailing bits.
    ///
    /// This is useful when implementing
    /// [forgiving-base64 decode](https://infra.spec.whatwg.org/#forgiving-base64-decode).
    pub fn decode_allow_trailing_bits(self, allow: bool) -> Config {
        Config {
            decode_allow_trailing_bits: allow,
            ..self
        }
    }
}

/// Standard character set with padding.
pub const STANDARD: Config = Config {
    char_set: CharacterSet::Standard,
    pad: true,
    decode_allow_trailing_bits: false,
};

/// Standard character set without padding.
pub const STANDARD_NO_PAD: Config = Config {
    char_set: CharacterSet::Standard,
    pad: false,
    decode_allow_trailing_bits: false,
};

/// URL-safe character set with padding
pub const URL_SAFE: Config = Config {
    char_set: CharacterSet::UrlSafe,
    pad: true,
    decode_allow_trailing_bits: false,
};

/// URL-safe character set without padding
pub const URL_SAFE_NO_PAD: Config = Config {
    char_set: CharacterSet::UrlSafe,
    pad: false,
    decode_allow_trailing_bits: false,
};

/// As per `crypt(3)` requirements
pub const CRYPT: Config = Config {
    char_set: CharacterSet::Crypt,
    pad: false,
    decode_allow_trailing_bits: false,
};