Move decode() into de module

This commit is contained in:
Dirkjan Ochtman 2023-10-25 22:21:34 +02:00
parent 6ea31b721f
commit efdf9334c7
2 changed files with 154 additions and 147 deletions

View File

@ -1,9 +1,10 @@
use std::borrow::Cow;
use std::collections::{BTreeMap, VecDeque};
use std::str::{self, FromStr};
use xmlparser::{ElementEnd, Token, Tokenizer};
use crate::impls::{decode, CowStrAccumulator};
use crate::impls::CowStrAccumulator;
use crate::{Error, Id};
pub struct Deserializer<'cx, 'xml> {
@ -383,6 +384,108 @@ pub fn borrow_cow_slice_u8<'xml>(
Ok(())
}
pub(crate) fn decode(input: &str) -> Result<Cow<'_, str>, Error> {
let mut result = String::with_capacity(input.len());
let (mut state, mut last_end) = (DecodeState::Normal, 0);
for (i, &b) in input.as_bytes().iter().enumerate() {
// use a state machine to find entities
state = match (state, b) {
(DecodeState::Normal, b'&') => DecodeState::Entity([0; 6], 0),
(DecodeState::Normal, _) => DecodeState::Normal,
(DecodeState::Entity(chars, len), b';') => {
let decoded = match &chars[..len] {
[b'a', b'm', b'p'] => '&',
[b'a', b'p', b'o', b's'] => '\'',
[b'g', b't'] => '>',
[b'l', b't'] => '<',
[b'q', b'u', b'o', b't'] => '"',
[b'#', b'x' | b'X', hex @ ..] => {
// Hexadecimal character reference e.g. "&#x007c;" -> '|'
str::from_utf8(hex)
.ok()
.and_then(|hex_str| u32::from_str_radix(hex_str, 16).ok())
.and_then(char::from_u32)
.filter(valid_xml_character)
.ok_or_else(|| {
Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
)
})?
}
[b'#', decimal @ ..] => {
// Decimal character reference e.g. "&#1234;" -> 'Ӓ'
str::from_utf8(decimal)
.ok()
.and_then(|decimal_str| u32::from_str(decimal_str).ok())
.and_then(char::from_u32)
.filter(valid_xml_character)
.ok_or_else(|| {
Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
)
})?
}
_ => {
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
))
}
};
let start = i - (len + 1); // current position - (length of entity characters + 1 for '&')
if last_end < start {
// Unwrap should be safe: `last_end` and `start` must be at character boundaries.
result.push_str(input.get(last_end..start).unwrap());
}
last_end = i + 1;
result.push(decoded);
DecodeState::Normal
}
(DecodeState::Entity(mut chars, len), b) => {
if len >= 6 {
let mut bytes = Vec::with_capacity(7);
bytes.extend(&chars[..len]);
bytes.push(b);
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&bytes).into_owned(),
));
}
chars[len] = b;
DecodeState::Entity(chars, len + 1)
}
};
}
// Unterminated entity (& without ;) at end of input
if let DecodeState::Entity(chars, len) = state {
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
));
}
Ok(match result.is_empty() {
true => Cow::Borrowed(input),
false => {
// Unwrap should be safe: `last_end` and `input.len()` must be at character boundaries.
result.push_str(input.get(last_end..input.len()).unwrap());
Cow::Owned(result)
}
})
}
#[derive(Debug)]
enum DecodeState {
Normal,
Entity([u8; 6], usize),
}
/// Valid character ranges per https://www.w3.org/TR/xml/#NT-Char
fn valid_xml_character(c: &char) -> bool {
matches!(c, '\u{9}' | '\u{A}' | '\u{D}' | '\u{20}'..='\u{D7FF}' | '\u{E000}'..='\u{FFFD}' | '\u{10000}'..='\u{10FFFF}')
}
#[derive(Debug)]
pub enum Node<'xml> {
Attribute(Attribute<'xml>),
@ -418,3 +521,52 @@ pub struct Attribute<'xml> {
pub local: &'xml str,
pub value: &'xml str,
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_decode() {
decode_ok("foo", "foo");
decode_ok("foo &amp; bar", "foo & bar");
decode_ok("foo &lt; bar", "foo < bar");
decode_ok("foo &gt; bar", "foo > bar");
decode_ok("foo &quot; bar", "foo \" bar");
decode_ok("foo &apos; bar", "foo ' bar");
decode_ok("foo &amp;lt; bar", "foo &lt; bar");
decode_ok("&amp; foo", "& foo");
decode_ok("foo &amp;", "foo &");
decode_ok("cbdtéda&amp;sü", "cbdtéda&sü");
// Decimal character references
decode_ok("&#1234;", "Ӓ");
decode_ok("foo &#9; bar", "foo \t bar");
decode_ok("foo &#124; bar", "foo | bar");
decode_ok("foo &#1234; bar", "foo Ӓ bar");
// Hexadecimal character references
decode_ok("&#xc4;", "Ä");
decode_ok("&#x00c4;", "Ä");
decode_ok("foo &#x9; bar", "foo \t bar");
decode_ok("foo &#x007c; bar", "foo | bar");
decode_ok("foo &#xc4; bar", "foo Ä bar");
decode_ok("foo &#x00c4; bar", "foo Ä bar");
decode_ok("foo &#x10de; bar", "foo პ bar");
decode_err("&");
decode_err("&#");
decode_err("&#;");
decode_err("foo&");
decode_err("&bar");
decode_err("&foo;");
decode_err("&foobar;");
decode_err("cbdtéd&ampü");
}
fn decode_ok(input: &str, expected: &'static str) {
assert_eq!(super::decode(input).unwrap(), expected, "{input:?}");
}
fn decode_err(input: &str) {
assert!(super::decode(input).is_err(), "{input:?}");
}
}

View File

@ -8,6 +8,7 @@ use std::{any::type_name, marker::PhantomData};
#[cfg(feature = "chrono")]
use chrono::{DateTime, NaiveDate, Utc};
use crate::de::decode;
use crate::{Accumulate, Deserializer, Error, FromXml, Id, Kind, Serializer, ToXml};
// Deserializer
@ -513,108 +514,6 @@ fn encode(input: &str) -> Result<Cow<'_, str>, Error> {
Ok(Cow::Owned(result))
}
pub(crate) fn decode(input: &str) -> Result<Cow<'_, str>, Error> {
let mut result = String::with_capacity(input.len());
let (mut state, mut last_end) = (DecodeState::Normal, 0);
for (i, &b) in input.as_bytes().iter().enumerate() {
// use a state machine to find entities
state = match (state, b) {
(DecodeState::Normal, b'&') => DecodeState::Entity([0; 6], 0),
(DecodeState::Normal, _) => DecodeState::Normal,
(DecodeState::Entity(chars, len), b';') => {
let decoded = match &chars[..len] {
[b'a', b'm', b'p'] => '&',
[b'a', b'p', b'o', b's'] => '\'',
[b'g', b't'] => '>',
[b'l', b't'] => '<',
[b'q', b'u', b'o', b't'] => '"',
[b'#', b'x' | b'X', hex @ ..] => {
// Hexadecimal character reference e.g. "&#x007c;" -> '|'
str::from_utf8(hex)
.ok()
.and_then(|hex_str| u32::from_str_radix(hex_str, 16).ok())
.and_then(char::from_u32)
.filter(valid_xml_character)
.ok_or_else(|| {
Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
)
})?
}
[b'#', decimal @ ..] => {
// Decimal character reference e.g. "&#1234;" -> 'Ӓ'
str::from_utf8(decimal)
.ok()
.and_then(|decimal_str| u32::from_str(decimal_str).ok())
.and_then(char::from_u32)
.filter(valid_xml_character)
.ok_or_else(|| {
Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
)
})?
}
_ => {
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
))
}
};
let start = i - (len + 1); // current position - (length of entity characters + 1 for '&')
if last_end < start {
// Unwrap should be safe: `last_end` and `start` must be at character boundaries.
result.push_str(input.get(last_end..start).unwrap());
}
last_end = i + 1;
result.push(decoded);
DecodeState::Normal
}
(DecodeState::Entity(mut chars, len), b) => {
if len >= 6 {
let mut bytes = Vec::with_capacity(7);
bytes.extend(&chars[..len]);
bytes.push(b);
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&bytes).into_owned(),
));
}
chars[len] = b;
DecodeState::Entity(chars, len + 1)
}
};
}
// Unterminated entity (& without ;) at end of input
if let DecodeState::Entity(chars, len) = state {
return Err(Error::InvalidEntity(
String::from_utf8_lossy(&chars[..len]).into_owned(),
));
}
Ok(match result.is_empty() {
true => Cow::Borrowed(input),
false => {
// Unwrap should be safe: `last_end` and `input.len()` must be at character boundaries.
result.push_str(input.get(last_end..input.len()).unwrap());
Cow::Owned(result)
}
})
}
#[derive(Debug)]
enum DecodeState {
Normal,
Entity([u8; 6], usize),
}
/// Valid character ranges per https://www.w3.org/TR/xml/#NT-Char
fn valid_xml_character(c: &char) -> bool {
matches!(c, '\u{9}' | '\u{A}' | '\u{D}' | '\u{20}'..='\u{D7FF}' | '\u{E000}'..='\u{FFFD}' | '\u{10000}'..='\u{10FFFF}')
}
impl<'xml, T: FromXml<'xml>> FromXml<'xml> for Vec<T> {
#[inline]
fn matches(id: Id<'_>, field: Option<Id<'_>>) -> bool {
@ -851,50 +750,6 @@ impl<'xml> FromXml<'xml> for IpAddr {
mod tests {
use super::*;
#[test]
fn test_decode() {
decode_ok("foo", "foo");
decode_ok("foo &amp; bar", "foo & bar");
decode_ok("foo &lt; bar", "foo < bar");
decode_ok("foo &gt; bar", "foo > bar");
decode_ok("foo &quot; bar", "foo \" bar");
decode_ok("foo &apos; bar", "foo ' bar");
decode_ok("foo &amp;lt; bar", "foo &lt; bar");
decode_ok("&amp; foo", "& foo");
decode_ok("foo &amp;", "foo &");
decode_ok("cbdtéda&amp;sü", "cbdtéda&sü");
// Decimal character references
decode_ok("&#1234;", "Ӓ");
decode_ok("foo &#9; bar", "foo \t bar");
decode_ok("foo &#124; bar", "foo | bar");
decode_ok("foo &#1234; bar", "foo Ӓ bar");
// Hexadecimal character references
decode_ok("&#xc4;", "Ä");
decode_ok("&#x00c4;", "Ä");
decode_ok("foo &#x9; bar", "foo \t bar");
decode_ok("foo &#x007c; bar", "foo | bar");
decode_ok("foo &#xc4; bar", "foo Ä bar");
decode_ok("foo &#x00c4; bar", "foo Ä bar");
decode_ok("foo &#x10de; bar", "foo პ bar");
decode_err("&");
decode_err("&#");
decode_err("&#;");
decode_err("foo&");
decode_err("&bar");
decode_err("&foo;");
decode_err("&foobar;");
decode_err("cbdtéd&ampü");
}
fn decode_ok(input: &str, expected: &'static str) {
assert_eq!(super::decode(input).unwrap(), expected, "{input:?}");
}
fn decode_err(input: &str) {
assert!(super::decode(input).is_err(), "{input:?}");
}
#[test]
fn encode_unicode() {
let input = "Iñtërnâ&tiônàlizætiøn";