b32_rs/src/decode.rs

132 lines
4.3 KiB
Rust

use {
crate::{error::Error, B32Alphabet},
std::io::{ErrorKind, Read, Write},
};
pub struct Decoder<R: Read, W: Write> {
reader: R,
writer: W,
alphabet: B32Alphabet,
ignore_whitespace: bool
}
impl<R: Read, W: Write> Decoder<R, W> {
pub fn new(reader: R, writer: W, alphabet: Option<B32Alphabet>, ignore_whitespace: bool) -> Self {
Self {
reader,
writer,
alphabet: alphabet.unwrap_or_default(),
ignore_whitespace,
}
}
/// Decodes the bytes provided by `self.reader` and writes the resulting
/// bytes into `self.writer`.
/// # Important
/// Both reader and writer should be buffered for good performance, as the
/// process of decoding will perform a series of very short read and write
/// operations.
/// # Errors
/// - io failure will return `Error::Io`
/// - a character not in the alphabet, or encountering any character after
/// the first occurance of the padding character, will return `Error::IllegalChar`
/// - if the reader comes up short of the final four byte block will return
/// `Error::MissingPadding`
/// - `intError` should never be returned. If this error is recieved the code
/// is somehow broken. Please file a bug.
pub fn decode(mut self) -> Result<W, Error> {
let mut byte_reader = self.reader.bytes();
'outer: loop {
let mut in_buf = [0_u8; 8];
let mut out_buf = [0_u8; 5];
let mut num: u64 = 0;
let mut n_bytes = 0;
while n_bytes < 8 {
match byte_reader.next() {
Some(Ok(b)) => {
if self.ignore_whitespace && b.is_ascii_whitespace() {
continue;
} else if b == b'\n' || b == b'\r' {
continue;
} else {
in_buf[n_bytes] = b;
n_bytes += 1;
}
}
Some(Err(e)) if e.kind() == ErrorKind::UnexpectedEof => break,
Some(Err(e)) if e.kind() == ErrorKind::Interrupted => continue,
Some(Err(e)) => return Err(e.into()),
None => break,
}
}
for c in &in_buf {
num <<= 5;
if !matches!(self.alphabet.pad(), Some(ch) if ch == *c) {
let idx = self.alphabet.idx(*c).ok_or(Error::IllegalChar((*c).into()))?;
num |= idx as u64;
}
}
for i in (0..5).rev() {
let b = (num & 0xff) as u8;
out_buf[i] = b;
num >>= 8;
}
for c in &out_buf {
if *c == b'\0' {
break 'outer;
} else {
self.writer.write_all(&[*c])?;
}
}
}
self.writer.flush()?;
Ok(self.writer)
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::B32_RFC4648_ALPHABET;
use std::{
fs::File,
io::{BufReader, Read},
};
static HELLO: &'static str = "Hello, World!";
static ENCODED: &'static str = "JBSWY3DPFQQFO33SNRSCC===";
#[test]
fn get_idx() {
let idx = B32_RFC4648_ALPHABET.idx(b'S').unwrap();
assert_eq!(idx, 18);
}
#[test]
fn hello() {
let reader = ENCODED.as_bytes();
let writer = Vec::<u8>::new();
let decoder = Decoder::new(reader, writer, None, false);
assert_eq!(decoder.decode().unwrap(), HELLO.as_bytes());
}
#[test]
fn ignore_whitespace() {
let decoder = Decoder::new(" JBSWY3DP\tFQQFO33SNRSCC===".as_bytes(), vec![], None, true);
assert_eq!(decoder.decode().unwrap(), HELLO.as_bytes());
}
#[test]
fn decode_wrapped() {
let infile = File::open("src/testdata/lorem_b32.txt").unwrap();
let reader = BufReader::new(infile);
let mut outfile = vec![];
File::open("src/testdata/lorem.txt")
.unwrap()
.read_to_end(&mut outfile)
.unwrap();
let decoder = Decoder::new(reader, vec![], None, false);
assert_eq!(decoder.decode().unwrap(), outfile);
}
}