Files
cow-mail/src/header/mod.rs
T
2026-09-13 00:16:49 +02:00

309 lines
9.5 KiB
Rust

mod encoding;
mod headers;
pub use encoding::encode_header_value;
pub use headers::{HeaderMap, parse_headers};
use std::borrow::Cow;
use std::collections::BTreeMap;
use std::fmt::{self, Display};
use charset::decode_latin1;
use mailparse::parse_content_type;
pub trait Header<'a>: Clone + Display {
fn new<K: ToString, V: ToString>(key: K, value: V) -> Self;
fn from(raw_bytes: &'a [u8], key_end: usize, value_start: usize, value_end: usize) -> Self;
fn get_key_raw(&self) -> &[u8];
fn get_value_raw(&self) -> &[u8];
fn set_value_raw(&mut self, value: Vec<u8>);
fn serialize(&self) -> Vec<Cow<'_, [u8]>>;
/// The raw header line, key included. Implementations holding the data
/// should override this to borrow it instead of rebuilding it.
fn get_raw(&self) -> Cow<'_, [u8]> {
let key = self.get_key_raw();
let value = self.get_value_raw();
let mut raw = Vec::with_capacity(key.len() + 2 + value.len());
raw.extend_from_slice(key);
raw.extend_from_slice(b": ");
raw.extend_from_slice(value);
Cow::Owned(raw)
}
fn get_key(&self) -> Cow<'_, str> {
decode_latin1(self.get_key_raw())
}
fn key_eq(&self, key: &str) -> bool {
self.get_key().eq_ignore_ascii_case(key)
}
fn get_value_ref(&self) -> Cow<'_, str> {
// RFC 6532 says that header values can be UTF-8. Let's try that first, and
// fall back to latin1 if that fails
let value = self.get_value_raw();
match std::str::from_utf8(value) {
Ok(s) => Cow::Borrowed(s),
Err(_) => decode_latin1(value),
}
}
/// The header value, unfolded and RFC 2047 decoded.
fn get_value(&self) -> String {
match mailparse::parse_header(&self.get_raw()) {
Ok((header, _)) => header.get_value(),
// Only reachable for a malformed header; hand back the value as it is
// rather than losing it.
Err(_) => self.get_value_ref().into_owned(),
}
}
fn set_value<V: ToString>(&mut self, value: V) {
let value = value.to_string();
assert!(!value.is_empty());
let encoded_value = encode_header_value(&self.get_key(), &value);
self.set_value_raw(encoded_value.into_bytes());
}
fn get_param(&self, key: &str) -> Option<String>
where
Self: Sized,
{
let header = RFC2045HeaderValue::from(self);
header.params.get(key).cloned()
}
fn set_param<K: ToString, V: ToString>(&mut self, key: K, value: V)
where
Self: Sized,
{
let key = key.to_string();
let value = value.to_string();
assert!(!key.is_empty());
assert!(!value.is_empty());
let mut rfc2045_value = RFC2045HeaderValue::from(&*self);
rfc2045_value.params.insert(key, value);
self.set_value(rfc2045_value.to_string());
}
}
struct RFC2045HeaderValue<T> {
pub value: T,
pub params: BTreeMap<T, T>,
}
impl<T: ToString> Display for RFC2045HeaderValue<T> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
match self.params.is_empty() {
true => write!(f, "{}", self.value.to_string()),
false => {
let params = self
.params
.iter()
.map(|(key, value)| {
let key = key.to_string();
let encoded_value = encode_header_value(&key, &value.to_string());
format!("{key}=\"{encoded_value}\"")
})
.collect::<Vec<_>>();
write!(f, "{}; {}", self.value.to_string(), params.join("; "))
}
}
}
}
impl<'a, T: Header<'a>> From<&'_ T> for RFC2045HeaderValue<String> {
fn from(header: &T) -> Self {
let key = header.get_key();
let (value, params) = match key.to_ascii_lowercase().as_str() {
"content-type" | "content-disposition" => {
let parsed = parse_content_type(&header.get_value());
(parsed.mimetype, parsed.params)
}
_ => (header.get_value(), BTreeMap::new()),
};
RFC2045HeaderValue { value, params }
}
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct CowHeader<'a> {
raw_bytes: Cow<'a, [u8]>,
key_end: usize,
value_start: usize,
value_end: usize,
}
impl CowHeader<'_> {
pub fn is_owned(&self) -> bool {
match self.raw_bytes {
Cow::Owned(_) => true,
Cow::Borrowed(_) => false,
}
}
}
impl<'a> Header<'a> for CowHeader<'a> {
fn new<K: ToString, V: ToString>(key: K, value: V) -> Self {
let key = key.to_string();
let value = value.to_string();
assert!(!key.is_empty());
assert!(!value.is_empty());
let encoded_value = encode_header_value(&key, &value);
let mut key_bytes = key.into_bytes();
let mut value_bytes = encoded_value.into_bytes();
let mut raw_bytes: Vec<u8> = Vec::with_capacity(key_bytes.len() + value_bytes.len() + 4);
raw_bytes.append(&mut key_bytes);
let key_end = raw_bytes.len();
raw_bytes.extend(b": ");
let value_start = raw_bytes.len();
raw_bytes.append(&mut value_bytes);
let value_end = raw_bytes.len();
raw_bytes.extend(b"\r\n");
Self {
raw_bytes: Cow::Owned(raw_bytes),
key_end,
value_start,
value_end,
}
}
fn from(raw_bytes: &'a [u8], key_end: usize, value_start: usize, value_end: usize) -> Self {
Self {
raw_bytes: Cow::Borrowed(raw_bytes),
key_end,
value_start,
value_end,
}
}
fn get_raw(&self) -> Cow<'_, [u8]> {
Cow::Borrowed(&self.raw_bytes)
}
fn get_key_raw(&self) -> &[u8] {
&self.raw_bytes[0..self.key_end]
}
fn get_value_raw(&self) -> &[u8] {
&self.raw_bytes[self.value_start..self.value_end]
}
fn set_value_raw(&mut self, mut value: Vec<u8>) {
assert!(!value.is_empty());
let raw_bytes = self.raw_bytes.to_mut();
raw_bytes.truncate(self.value_start);
raw_bytes.append(&mut value);
self.value_end = raw_bytes.len();
raw_bytes.extend(b"\r\n");
}
fn serialize(&self) -> Vec<Cow<'_, [u8]>> {
vec![Cow::Borrowed(&self.raw_bytes)]
}
}
impl Display for CowHeader<'_> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}: {}", self.get_key(), self.get_value_ref())
}
}
enum HeaderParseState {
Initial,
Key,
PreValue,
Value,
ValueNewline,
}
pub fn parse_header<'a, H: Header<'a>>(raw_data: &'a [u8]) -> Result<(H, usize), &'static str> {
let mut it = raw_data.iter();
let mut ix = 0;
let mut c = match it.next() {
None => return Err("Empty string provided"),
Some(v) => *v,
};
let mut ix_key_end = None;
let mut ix_value_start = 0;
let mut ix_value_end = 0;
let mut state = HeaderParseState::Initial;
loop {
match state {
HeaderParseState::Initial => {
if c == b' ' {
return Err(
"Header cannot start with a space; it is likely an overhanging line from a previous header",
);
};
state = HeaderParseState::Key;
continue;
}
HeaderParseState::Key => {
if c == b':' {
ix_key_end = Some(ix);
state = HeaderParseState::PreValue;
} else if c == b'\n' {
// Technically this is invalid. We'll handle it gracefully
// since it does appear to happen in the wild and other
// MTAs deal with it. Our handling is to just treat everything
// encountered so far on this line as the header key, and
// leave the value empty.
ix_key_end = Some(ix);
ix_value_start = ix;
ix_value_end = ix;
ix += 1;
break;
}
}
HeaderParseState::PreValue => {
if c != b' ' {
ix_value_start = ix;
ix_value_end = ix;
state = HeaderParseState::Value;
continue;
}
}
HeaderParseState::Value => {
if c == b'\n' {
state = HeaderParseState::ValueNewline;
} else if c != b'\r' {
ix_value_end = ix + 1;
}
}
HeaderParseState::ValueNewline => {
if c == b' ' || c == b'\t' {
state = HeaderParseState::Value;
continue;
} else {
break;
}
}
}
ix += 1;
c = match it.next() {
None => break,
Some(v) => *v,
};
}
match ix_key_end {
Some(v) => Ok((H::from(&raw_data[0..ix], v, ix_value_start, ix_value_end), ix)),
None => Ok((
// Technically this is invalid. We'll handle it gracefully
// since we handle the analogous situation above. Our handling
// is to just treat everything encountered on this line as
// the header key, and leave the value empty.
H::from(&raw_data[0..ix], ix, ix, ix),
ix,
)),
}
}