Feature/tag/urlencode #11

Merged
bazaah merged 8 commits from feature/tag/urlencode into master 2021-06-26 08:24:08 +00:00
2 changed files with 189 additions and 14 deletions

View File

@ -70,7 +70,7 @@ macro_rules! check {
(@priv $buffer:expr, $offset:expr => $( $match:tt )|+) => { (@priv $buffer:expr, $offset:expr => $( $match:tt )|+) => {
match $buffer.get($offset..) { match $buffer.get($offset..) {
Some(buffer) => check!(@priv buffer => $( $match )|+), Some(buffer) => check!(@priv buffer => $( $match )|+),
None => check!(@eofck $( $match )|+ ) None => check!(@eofck $( $match )|+ ),
} }
}; };
(@priv $buffer:expr => $( $match:tt )|+) => { (@priv $buffer:expr => $( $match:tt )|+) => {
@ -81,16 +81,16 @@ macro_rules! check {
}; };
(@priv $buffer:expr, $offset:expr => $( $match:tt )|+, else $error:expr) => { (@priv $buffer:expr, $offset:expr => $( $match:tt )|+, else $error:expr) => {
match $buffer.get($offset..) { match $buffer.get($offset..) {
Some(buffer) => check!(@priv buffer => $( $match )|+ else $error), Some(buffer) => check!(@priv buffer => $( $match )|+, else $error),
None if check!(@eofck $( $match )|+ ) => Ok(()) None if check!(@eofck $( $match )|+ ) => Ok(()),
_ => Err($crate::scanner::error::ScanError::UnexpectedEOF) _ => Err($crate::scanner::error::ScanError::UnexpectedEOF),
} }
}; };
(@priv $buffer:expr => $( $match:tt )|+, else $error:expr) => { (@priv $buffer:expr => $( $match:tt )|+, else $error:expr) => {
match $buffer { match $buffer {
$( check!(@ptn $match) )|+ => Ok(()), $( check!(@ptn $match) )|+ => Ok(()),
[] => Err($crate::scanner::error::ScanError::UnexpectedEOF), [] => Err($crate::scanner::error::ScanError::UnexpectedEOF),
_ => Err($error) _ => Err($error),
} }
}; };
@ -186,6 +186,29 @@ macro_rules! isBlankZ {
}; };
} }
/// Checks if byte (@ .offset) in .buffer is hexadecimal
///
/// Modifiers:
/// ~ .buffer := .buffer.as_bytes()
///
/// Variants:
/// /1 .buffer := /2 .buffer, 0
/// /2 .buffer, .offset
/// /3 .buffer, else .error
/// := /4 .buffer, 0, else .error
/// /4 .buffer, .offset, else .error
macro_rules! isHex {
(~ $buffer:expr $(, $offset:expr )? $(, else $error:expr )? ) => {
isHex!($buffer.as_bytes() $(, $offset)? $(, else $error)? )
};
($buffer:expr $(, $offset:expr )? $(, else $error:expr)? ) => {
check!($buffer $(, $offset)? =>
[b'0'..=b'9', ..] | [b'A'..=b'F', ..] | [b'a'..=b'f', ..]
$(, else $error )?
)
};
}
#[cfg(test)] #[cfg(test)]
mod tests mod tests
{ {

View File

@ -9,8 +9,8 @@ use crate::scanner::error::{ScanError, ScanResult as Result};
/// 32 bit unicode points. /// 32 bit unicode points.
/// ///
/// It writes the unescaped character to .scratch, returning /// It writes the unescaped character to .scratch, returning
/// the length of .buffer advanced, or an error if the /// the length of .base advanced, or an error if the
/// escape sequence is invalid. It expects .buffer->0 is a /// escape sequence is invalid. It expects .base->0 is a
/// backslash (\\), as this is the only valid start of an /// backslash (\\), as this is the only valid start of an
/// escape sequence. /// escape sequence.
/// ///
@ -67,6 +67,67 @@ pub(super) fn flow_unescape(base: &str, scratch: &mut Vec<u8>) -> Result<usize>
Ok(base.len() - buffer.len()) Ok(base.len() - buffer.len())
} }
/// Unescape a percent encoded UTF8 tag escape sequence as
/// defined in [Section 5.6][Link], writing the code point
/// to the scratch, returning the length of .base consumed.
///
/// [Link]: https://yaml.org/spec/1.2/spec.html#ns-uri-char
pub(super) fn tag_uri_unescape(base: &str, scratch: &mut Vec<u8>, _directive: bool)
-> Result<usize>
{
let mut buffer = base;
let mut codepoint_len: i8 = 0;
while {
if buffer.len() < 3
{
return Err(ScanError::UnexpectedEOF);
}
if !(check!(~buffer => b'%') && isHex!(~buffer, 1) && isHex!(~buffer, 2))
{
return Err(ScanError::UnknownEscape);
}
// Safety: we just checked that there are at least three
// bytes in the buffer
let octet: u8 = (as_hex(buffer.as_bytes()[1]) << 4) + as_hex(buffer.as_bytes()[2]);
match codepoint_len
{
// First time through, determine how many octets this codepoint has
0 =>
{
codepoint_len = match octet
{
o if (o & 0x80) == 0x00 => 1,
o if (o & 0xE0) == 0xC0 => 2,
o if (o & 0xF0) == 0xE0 => 3,
o if (o & 0xF8) == 0xF0 => 4,
_ => return Err(ScanError::UnknownEscape),
}
},
// Else ensure that the trailing octet is valid
_ =>
{
if (octet & 0xC0) != 0x80
{
return Err(ScanError::UnknownEscape);
}
},
}
scratch.push(octet);
codepoint_len -= 1;
advance!(buffer, 3);
codepoint_len > 0
}
{}
Ok(base.len() - buffer.len())
}
/// Writes a UTF8 codepoint to the scratch space /// Writes a UTF8 codepoint to the scratch space
fn write_unicode_point(base: &str, scratch: &mut Vec<u8>, codepoint_len: u8) -> Result<usize> fn write_unicode_point(base: &str, scratch: &mut Vec<u8>, codepoint_len: u8) -> Result<usize>
{ {
@ -86,7 +147,7 @@ fn write_unicode_point(base: &str, scratch: &mut Vec<u8>, codepoint_len: u8) ->
None => return Err(ScanError::UnexpectedEOF), None => return Err(ScanError::UnexpectedEOF),
Some(c) if !c.is_ascii_hexdigit() => return Err(ScanError::UnknownEscape), Some(c) if !c.is_ascii_hexdigit() => return Err(ScanError::UnknownEscape),
Some(b) => value = (value << 4) + as_hex(*b), Some(b) => value = (value << 4) + as_hex(*b) as u32,
} }
advance!(buffer, 1, i); advance!(buffer, 1, i);
} }
@ -128,9 +189,9 @@ fn write_unicode_point(base: &str, scratch: &mut Vec<u8>, codepoint_len: u8) ->
*/ */
#[allow(clippy::manual_range_contains)] #[allow(clippy::manual_range_contains)]
#[inline] #[inline]
fn as_hex(b: u8) -> u32 fn as_hex(b: u8) -> u8
{ {
let ret = if b >= b'A' && b <= b'F' if b >= b'A' && b <= b'F'
{ {
b - b'A' + 10 b - b'A' + 10
} }
@ -141,9 +202,7 @@ fn as_hex(b: u8) -> u32
else else
{ {
b - b'0' b - b'0'
}; }
ret as u32
} }
/// <Next Line> (U+0085) /// <Next Line> (U+0085)
@ -158,7 +217,7 @@ const PS: [u8; 3] = [b'\xE2', b'\x80', b'\xA9'];
#[cfg(test)] #[cfg(test)]
mod tests mod tests
{ {
use anyhow::anyhow; use anyhow::{anyhow, bail};
use pretty_assertions::assert_eq; use pretty_assertions::assert_eq;
use super::*; use super::*;
@ -323,4 +382,97 @@ mod tests
Ok(()) Ok(())
} }
#[test]
fn tag_uri_unescape_codepoint() -> TestResult
{
let data = &[
r#"%C2%85"#,
r#"%c5%b4"#,
r#"%E2%B1%bf"#,
r#"%E2%B8%BF"#,
r#"%f0%90%8f%95"#,
r#"%F0%90%AD%81"#,
];
let expected: &[&[u8]] = &[
&[0xC2, 0x85],
&[0xC5, 0xB4],
&[0xE2, 0xB1, 0xBF],
&[0xE2, 0xB8, 0xBF],
&[0xF0, 0x90, 0x8F, 0x95],
&[0xF0, 0x90, 0xAD, 0x81],
];
let scratch = &mut Vec::new();
assert_eq!(
data.len(),
expected.len(),
"test data and expected data are not the same length"
);
for (i, (&t, &e)) in data.into_iter().zip(expected).enumerate()
{
scratch.clear();
let consumed = tag_uri_unescape(t, scratch, true)
.map_err(|e| anyhow!("on iteration {}, test errored with {}", i, e))?;
assert_eq!(
&*scratch, e,
"on iteration {}, expected byte sequence {:?}, got {:?}",
i, e, &*scratch
);
assert_eq!(
consumed,
t.len(),
"on iteration {}, expected to consume {}, got {}",
i,
t.len(),
consumed
)
}
Ok(())
}
#[test]
fn tag_uri_unescape_eof() -> TestResult
{
let data = r#"%C2%8"#;
let scratch = &mut Vec::new();
let expected = ScanError::UnexpectedEOF;
match tag_uri_unescape(data, scratch, true)
{
Err(e) if e == expected => Ok(()),
Err(e) => bail!("expected error: {}, got different error: {}", expected, e),
Ok(amt) => bail!(
"expected error: {}, got unexpected value: {}",
expected,
amt
),
}
}
#[test]
fn tag_uri_unescape_invalid() -> TestResult
{
let data = r#"\xC285"#;
let scratch = &mut Vec::new();
let expected = ScanError::UnknownEscape;
match tag_uri_unescape(data, scratch, true)
{
Err(e) if e == expected => Ok(()),
Err(e) => bail!("expected error: {}, got different error: {}", expected, e),
Ok(amt) => bail!(
"expected error: {}, got unexpected value: {}",
expected,
amt
),
}
}
} }