summaryrefslogtreecommitdiffstats
path: root/third_party/rust/url/src
diff options
context:
space:
mode:
Diffstat (limited to 'third_party/rust/url/src')
-rw-r--r--third_party/rust/url/src/encoding.rs135
-rw-r--r--third_party/rust/url/src/form_urlencoded.rs364
-rw-r--r--third_party/rust/url/src/host.rs418
-rw-r--r--third_party/rust/url/src/lib.rs1456
-rw-r--r--third_party/rust/url/src/origin.rs114
-rw-r--r--third_party/rust/url/src/parser.rs1179
-rw-r--r--third_party/rust/url/src/path_segments.rs187
-rw-r--r--third_party/rust/url/src/percent_encoding.rs344
-rw-r--r--third_party/rust/url/src/quirks.rs217
-rw-r--r--third_party/rust/url/src/slicing.rs182
10 files changed, 4596 insertions, 0 deletions
diff --git a/third_party/rust/url/src/encoding.rs b/third_party/rust/url/src/encoding.rs
new file mode 100644
index 000000000..0703c788f
--- /dev/null
+++ b/third_party/rust/url/src/encoding.rs
@@ -0,0 +1,135 @@
+// Copyright 2013-2014 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+
+//! Abstraction that conditionally compiles either to rust-encoding,
+//! or to only support UTF-8.
+
+#[cfg(feature = "query_encoding")] extern crate encoding;
+
+use std::borrow::Cow;
+
+#[cfg(feature = "query_encoding")] use self::encoding::types::{DecoderTrap, EncoderTrap};
+#[cfg(feature = "query_encoding")] use self::encoding::label::encoding_from_whatwg_label;
+#[cfg(feature = "query_encoding")] pub use self::encoding::types::EncodingRef;
+
+#[cfg(feature = "query_encoding")]
+#[derive(Copy, Clone)]
+pub struct EncodingOverride {
+ /// `None` means UTF-8.
+ encoding: Option<EncodingRef>
+}
+
+#[cfg(feature = "query_encoding")]
+impl EncodingOverride {
+ pub fn from_opt_encoding(encoding: Option<EncodingRef>) -> Self {
+ encoding.map(Self::from_encoding).unwrap_or_else(Self::utf8)
+ }
+
+ pub fn from_encoding(encoding: EncodingRef) -> Self {
+ EncodingOverride {
+ encoding: if encoding.name() == "utf-8" { None } else { Some(encoding) }
+ }
+ }
+
+ #[inline]
+ pub fn utf8() -> Self {
+ EncodingOverride { encoding: None }
+ }
+
+ pub fn lookup(label: &[u8]) -> Option<Self> {
+ // Don't use String::from_utf8_lossy since no encoding label contains U+FFFD
+ // https://encoding.spec.whatwg.org/#names-and-labels
+ ::std::str::from_utf8(label)
+ .ok()
+ .and_then(encoding_from_whatwg_label)
+ .map(Self::from_encoding)
+ }
+
+ /// https://encoding.spec.whatwg.org/#get-an-output-encoding
+ pub fn to_output_encoding(self) -> Self {
+ if let Some(encoding) = self.encoding {
+ if matches!(encoding.name(), "utf-16le" | "utf-16be") {
+ return Self::utf8()
+ }
+ }
+ self
+ }
+
+ pub fn is_utf8(&self) -> bool {
+ self.encoding.is_none()
+ }
+
+ pub fn name(&self) -> &'static str {
+ match self.encoding {
+ Some(encoding) => encoding.name(),
+ None => "utf-8",
+ }
+ }
+
+ pub fn decode<'a>(&self, input: Cow<'a, [u8]>) -> Cow<'a, str> {
+ match self.encoding {
+ // `encoding.decode` never returns `Err` when called with `DecoderTrap::Replace`
+ Some(encoding) => encoding.decode(&input, DecoderTrap::Replace).unwrap().into(),
+ None => decode_utf8_lossy(input),
+ }
+ }
+
+ pub fn encode<'a>(&self, input: Cow<'a, str>) -> Cow<'a, [u8]> {
+ match self.encoding {
+ // `encoding.encode` never returns `Err` when called with `EncoderTrap::NcrEscape`
+ Some(encoding) => Cow::Owned(encoding.encode(&input, EncoderTrap::NcrEscape).unwrap()),
+ None => encode_utf8(input)
+ }
+ }
+}
+
+
+#[cfg(not(feature = "query_encoding"))]
+#[derive(Copy, Clone)]
+pub struct EncodingOverride;
+
+#[cfg(not(feature = "query_encoding"))]
+impl EncodingOverride {
+ #[inline]
+ pub fn utf8() -> Self {
+ EncodingOverride
+ }
+
+ pub fn decode<'a>(&self, input: Cow<'a, [u8]>) -> Cow<'a, str> {
+ decode_utf8_lossy(input)
+ }
+
+ pub fn encode<'a>(&self, input: Cow<'a, str>) -> Cow<'a, [u8]> {
+ encode_utf8(input)
+ }
+}
+
+pub fn decode_utf8_lossy(input: Cow<[u8]>) -> Cow<str> {
+ match input {
+ Cow::Borrowed(bytes) => String::from_utf8_lossy(bytes),
+ Cow::Owned(bytes) => {
+ let raw_utf8: *const [u8];
+ match String::from_utf8_lossy(&bytes) {
+ Cow::Borrowed(utf8) => raw_utf8 = utf8.as_bytes(),
+ Cow::Owned(s) => return s.into(),
+ }
+ // from_utf8_lossy returned a borrow of `bytes` unchanged.
+ debug_assert!(raw_utf8 == &*bytes as *const [u8]);
+ // Reuse the existing `Vec` allocation.
+ unsafe { String::from_utf8_unchecked(bytes) }.into()
+ }
+ }
+}
+
+pub fn encode_utf8(input: Cow<str>) -> Cow<[u8]> {
+ match input {
+ Cow::Borrowed(s) => Cow::Borrowed(s.as_bytes()),
+ Cow::Owned(s) => Cow::Owned(s.into_bytes())
+ }
+}
diff --git a/third_party/rust/url/src/form_urlencoded.rs b/third_party/rust/url/src/form_urlencoded.rs
new file mode 100644
index 000000000..f4a655507
--- /dev/null
+++ b/third_party/rust/url/src/form_urlencoded.rs
@@ -0,0 +1,364 @@
+// Copyright 2013-2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+//! Parser and serializer for the [`application/x-www-form-urlencoded` syntax](
+//! http://url.spec.whatwg.org/#application/x-www-form-urlencoded),
+//! as used by HTML forms.
+//!
+//! Converts between a string (such as an URL’s query string)
+//! and a sequence of (name, value) pairs.
+
+use encoding::EncodingOverride;
+use percent_encoding::{percent_encode_byte, percent_decode};
+use std::borrow::{Borrow, Cow};
+use std::str;
+
+
+/// Convert a byte string in the `application/x-www-form-urlencoded` syntax
+/// into a iterator of (name, value) pairs.
+///
+/// Use `parse(input.as_bytes())` to parse a `&str` string.
+///
+/// The names and values are percent-decoded. For instance, `%23first=%25try%25` will be
+/// converted to `[("#first", "%try%")]`.
+#[inline]
+pub fn parse(input: &[u8]) -> Parse {
+ Parse {
+ input: input,
+ encoding: EncodingOverride::utf8(),
+ }
+}
+
+
+/// Convert a byte string in the `application/x-www-form-urlencoded` syntax
+/// into a iterator of (name, value) pairs.
+///
+/// Use `parse(input.as_bytes())` to parse a `&str` string.
+///
+/// This function is only available if the `query_encoding` Cargo feature is enabled.
+///
+/// Arguments:
+///
+/// * `encoding_override`: The character encoding each name and values is decoded as
+/// after percent-decoding. Defaults to UTF-8.
+/// * `use_charset`: The *use _charset_ flag*. If in doubt, set to `false`.
+#[cfg(feature = "query_encoding")]
+pub fn parse_with_encoding<'a>(input: &'a [u8],
+ encoding_override: Option<::encoding::EncodingRef>,
+ use_charset: bool)
+ -> Result<Parse<'a>, ()> {
+ use std::ascii::AsciiExt;
+
+ let mut encoding = EncodingOverride::from_opt_encoding(encoding_override);
+ if !(encoding.is_utf8() || input.is_ascii()) {
+ return Err(())
+ }
+ if use_charset {
+ for sequence in input.split(|&b| b == b'&') {
+ // No '+' in "_charset_" to replace with ' '.
+ if sequence.starts_with(b"_charset_=") {
+ let value = &sequence[b"_charset_=".len()..];
+ // Skip replacing '+' with ' ' in value since no encoding label contains either:
+ // https://encoding.spec.whatwg.org/#names-and-labels
+ if let Some(e) = EncodingOverride::lookup(value) {
+ encoding = e;
+ break
+ }
+ }
+ }
+ }
+ Ok(Parse {
+ input: input,
+ encoding: encoding,
+ })
+}
+
+/// The return type of `parse()`.
+#[derive(Copy, Clone)]
+pub struct Parse<'a> {
+ input: &'a [u8],
+ encoding: EncodingOverride,
+}
+
+impl<'a> Iterator for Parse<'a> {
+ type Item = (Cow<'a, str>, Cow<'a, str>);
+
+ fn next(&mut self) -> Option<Self::Item> {
+ loop {
+ if self.input.is_empty() {
+ return None
+ }
+ let mut split2 = self.input.splitn(2, |&b| b == b'&');
+ let sequence = split2.next().unwrap();
+ self.input = split2.next().unwrap_or(&[][..]);
+ if sequence.is_empty() {
+ continue
+ }
+ let mut split2 = sequence.splitn(2, |&b| b == b'=');
+ let name = split2.next().unwrap();
+ let value = split2.next().unwrap_or(&[][..]);
+ return Some((
+ decode(name, self.encoding),
+ decode(value, self.encoding),
+ ))
+ }
+ }
+}
+
+fn decode(input: &[u8], encoding: EncodingOverride) -> Cow<str> {
+ let replaced = replace_plus(input);
+ encoding.decode(match percent_decode(&replaced).if_any() {
+ Some(vec) => Cow::Owned(vec),
+ None => replaced,
+ })
+}
+
+/// Replace b'+' with b' '
+fn replace_plus<'a>(input: &'a [u8]) -> Cow<'a, [u8]> {
+ match input.iter().position(|&b| b == b'+') {
+ None => Cow::Borrowed(input),
+ Some(first_position) => {
+ let mut replaced = input.to_owned();
+ replaced[first_position] = b' ';
+ for byte in &mut replaced[first_position + 1..] {
+ if *byte == b'+' {
+ *byte = b' ';
+ }
+ }
+ Cow::Owned(replaced)
+ }
+ }
+}
+
+impl<'a> Parse<'a> {
+ /// Return a new iterator that yields pairs of `String` instead of pairs of `Cow<str>`.
+ pub fn into_owned(self) -> ParseIntoOwned<'a> {
+ ParseIntoOwned { inner: self }
+ }
+}
+
+/// Like `Parse`, but yields pairs of `String` instead of pairs of `Cow<str>`.
+pub struct ParseIntoOwned<'a> {
+ inner: Parse<'a>
+}
+
+impl<'a> Iterator for ParseIntoOwned<'a> {
+ type Item = (String, String);
+
+ fn next(&mut self) -> Option<Self::Item> {
+ self.inner.next().map(|(k, v)| (k.into_owned(), v.into_owned()))
+ }
+}
+
+/// The [`application/x-www-form-urlencoded` byte serializer](
+/// https://url.spec.whatwg.org/#concept-urlencoded-byte-serializer).
+///
+/// Return an iterator of `&str` slices.
+pub fn byte_serialize(input: &[u8]) -> ByteSerialize {
+ ByteSerialize {
+ bytes: input,
+ }
+}
+
+/// Return value of `byte_serialize()`.
+pub struct ByteSerialize<'a> {
+ bytes: &'a [u8],
+}
+
+fn byte_serialized_unchanged(byte: u8) -> bool {
+ matches!(byte, b'*' | b'-' | b'.' | b'0' ... b'9' | b'A' ... b'Z' | b'_' | b'a' ... b'z')
+}
+
+impl<'a> Iterator for ByteSerialize<'a> {
+ type Item = &'a str;
+
+ fn next(&mut self) -> Option<&'a str> {
+ if let Some((&first, tail)) = self.bytes.split_first() {
+ if !byte_serialized_unchanged(first) {
+ self.bytes = tail;
+ return Some(if first == b' ' { "+" } else { percent_encode_byte(first) })
+ }
+ let position = tail.iter().position(|&b| !byte_serialized_unchanged(b));
+ let (unchanged_slice, remaining) = match position {
+ // 1 for first_byte + i unchanged in tail
+ Some(i) => self.bytes.split_at(1 + i),
+ None => (self.bytes, &[][..]),
+ };
+ self.bytes = remaining;
+ Some(unsafe { str::from_utf8_unchecked(unchanged_slice) })
+ } else {
+ None
+ }
+ }
+
+ fn size_hint(&self) -> (usize, Option<usize>) {
+ if self.bytes.is_empty() {
+ (0, Some(0))
+ } else {
+ (1, Some(self.bytes.len()))
+ }
+ }
+}
+
+/// The [`application/x-www-form-urlencoded` serializer](
+/// https://url.spec.whatwg.org/#concept-urlencoded-serializer).
+pub struct Serializer<T: Target> {
+ target: Option<T>,
+ start_position: usize,
+ encoding: EncodingOverride,
+}
+
+pub trait Target {
+ fn as_mut_string(&mut self) -> &mut String;
+ fn finish(self) -> Self::Finished;
+ type Finished;
+}
+
+impl Target for String {
+ fn as_mut_string(&mut self) -> &mut String { self }
+ fn finish(self) -> Self { self }
+ type Finished = Self;
+}
+
+impl<'a> Target for &'a mut String {
+ fn as_mut_string(&mut self) -> &mut String { &mut **self }
+ fn finish(self) -> Self { self }
+ type Finished = Self;
+}
+
+// `as_mut_string` string here exposes the internal serialization of an `Url`,
+// which should not be exposed to users.
+// We achieve that by not giving users direct access to `UrlQuery`:
+// * Its fields are private
+// (and so can not be constructed with struct literal syntax outside of this crate),
+// * It has no constructor
+// * It is only visible (on the type level) to users in the return type of
+// `Url::query_pairs_mut` which is `Serializer<UrlQuery>`
+// * `Serializer` keeps its target in a private field
+// * Unlike in other `Target` impls, `UrlQuery::finished` does not return `Self`.
+impl<'a> Target for ::UrlQuery<'a> {
+ fn as_mut_string(&mut self) -> &mut String { &mut self.url.serialization }
+ fn finish(self) -> &'a mut ::Url { self.url }
+ type Finished = &'a mut ::Url;
+}
+
+impl<T: Target> Serializer<T> {
+ /// Create a new `application/x-www-form-urlencoded` serializer for the given target.
+ ///
+ /// If the target is non-empty,
+ /// its content is assumed to already be in `application/x-www-form-urlencoded` syntax.
+ pub fn new(target: T) -> Self {
+ Self::for_suffix(target, 0)
+ }
+
+ /// Create a new `application/x-www-form-urlencoded` serializer
+ /// for a suffix of the given target.
+ ///
+ /// If that suffix is non-empty,
+ /// its content is assumed to already be in `application/x-www-form-urlencoded` syntax.
+ pub fn for_suffix(mut target: T, start_position: usize) -> Self {
+ &target.as_mut_string()[start_position..]; // Panic if out of bounds
+ Serializer {
+ target: Some(target),
+ start_position: start_position,
+ encoding: EncodingOverride::utf8(),
+ }
+ }
+
+ /// Remove any existing name/value pair.
+ ///
+ /// Panics if called after `.finish()`.
+ pub fn clear(&mut self) -> &mut Self {
+ string(&mut self.target).truncate(self.start_position);
+ self
+ }
+
+ /// Set the character encoding to be used for names and values before percent-encoding.
+ #[cfg(feature = "query_encoding")]
+ pub fn encoding_override(&mut self, new: Option<::encoding::EncodingRef>) -> &mut Self {
+ self.encoding = EncodingOverride::from_opt_encoding(new).to_output_encoding();
+ self
+ }
+
+ /// Serialize and append a name/value pair.
+ ///
+ /// Panics if called after `.finish()`.
+ pub fn append_pair(&mut self, name: &str, value: &str) -> &mut Self {
+ append_pair(string(&mut self.target), self.start_position, self.encoding, name, value);
+ self
+ }
+
+ /// Serialize and append a number of name/value pairs.
+ ///
+ /// This simply calls `append_pair` repeatedly.
+ /// This can be more convenient, so the user doesn’t need to introduce a block
+ /// to limit the scope of `Serializer`’s borrow of its string.
+ ///
+ /// Panics if called after `.finish()`.
+ pub fn extend_pairs<I, K, V>(&mut self, iter: I) -> &mut Self
+ where I: IntoIterator, I::Item: Borrow<(K, V)>, K: AsRef<str>, V: AsRef<str> {
+ {
+ let string = string(&mut self.target);
+ for pair in iter {
+ let &(ref k, ref v) = pair.borrow();
+ append_pair(string, self.start_position, self.encoding, k.as_ref(), v.as_ref());
+ }
+ }
+ self
+ }
+
+ /// Add a name/value pair whose name is `_charset_`
+ /// and whose value is the character encoding’s name.
+ /// (See the `encoding_override()` method.)
+ ///
+ /// Panics if called after `.finish()`.
+ #[cfg(feature = "query_encoding")]
+ pub fn append_charset(&mut self) -> &mut Self {
+ {
+ let string = string(&mut self.target);
+ append_separator_if_needed(string, self.start_position);
+ string.push_str("_charset_=");
+ string.push_str(self.encoding.name());
+ }
+ self
+ }
+
+ /// If this serializer was constructed with a string, take and return that string.
+ ///
+ /// ```rust
+ /// use url::form_urlencoded;
+ /// let encoded: String = form_urlencoded::Serializer::new(String::new())
+ /// .append_pair("foo", "bar & baz")
+ /// .append_pair("saison", "Été+hiver")
+ /// .finish();
+ /// assert_eq!(encoded, "foo=bar+%26+baz&saison=%C3%89t%C3%A9%2Bhiver");
+ /// ```
+ ///
+ /// Panics if called more than once.
+ pub fn finish(&mut self) -> T::Finished {
+ self.target.take().expect("url::form_urlencoded::Serializer double finish").finish()
+ }
+}
+
+fn append_separator_if_needed(string: &mut String, start_position: usize) {
+ if string.len() > start_position {
+ string.push('&')
+ }
+}
+
+fn string<T: Target>(target: &mut Option<T>) -> &mut String {
+ target.as_mut().expect("url::form_urlencoded::Serializer finished").as_mut_string()
+}
+
+fn append_pair(string: &mut String, start_position: usize, encoding: EncodingOverride,
+ name: &str, value: &str) {
+ append_separator_if_needed(string, start_position);
+ string.extend(byte_serialize(&encoding.encode(name.into())));
+ string.push('=');
+ string.extend(byte_serialize(&encoding.encode(value.into())));
+}
diff --git a/third_party/rust/url/src/host.rs b/third_party/rust/url/src/host.rs
new file mode 100644
index 000000000..47b049a27
--- /dev/null
+++ b/third_party/rust/url/src/host.rs
@@ -0,0 +1,418 @@
+// Copyright 2013-2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+#[cfg(feature = "heapsize")] use heapsize::HeapSizeOf;
+use std::cmp;
+use std::fmt::{self, Formatter};
+use std::io;
+use std::net::{Ipv4Addr, Ipv6Addr, SocketAddr, SocketAddrV4, SocketAddrV6, ToSocketAddrs};
+use std::vec;
+use parser::{ParseResult, ParseError};
+use percent_encoding::percent_decode;
+use idna;
+
+#[derive(Copy, Clone, Debug, Eq, PartialEq)]
+pub enum HostInternal {
+ None,
+ Domain,
+ Ipv4(Ipv4Addr),
+ Ipv6(Ipv6Addr),
+}
+
+#[cfg(feature = "heapsize")]
+known_heap_size!(0, HostInternal);
+
+impl<S> From<Host<S>> for HostInternal {
+ fn from(host: Host<S>) -> HostInternal {
+ match host {
+ Host::Domain(_) => HostInternal::Domain,
+ Host::Ipv4(address) => HostInternal::Ipv4(address),
+ Host::Ipv6(address) => HostInternal::Ipv6(address),
+ }
+ }
+}
+
+/// The host name of an URL.
+#[derive(Clone, Debug, Eq, PartialEq, Ord, PartialOrd, Hash)]
+pub enum Host<S=String> {
+ /// A DNS domain name, as '.' dot-separated labels.
+ /// Non-ASCII labels are encoded in punycode per IDNA.
+ Domain(S),
+
+ /// An IPv4 address.
+ /// `Url::host_str` returns the serialization of this address,
+ /// as four decimal integers separated by `.` dots.
+ Ipv4(Ipv4Addr),
+
+ /// An IPv6 address.
+ /// `Url::host_str` returns the serialization of that address between `[` and `]` brackets,
+ /// in the format per [RFC 5952 *A Recommendation
+ /// for IPv6 Address Text Representation*](https://tools.ietf.org/html/rfc5952):
+ /// lowercase hexadecimal with maximal `::` compression.
+ Ipv6(Ipv6Addr),
+}
+
+#[cfg(feature = "heapsize")]
+impl<S: HeapSizeOf> HeapSizeOf for Host<S> {
+ fn heap_size_of_children(&self) -> usize {
+ match *self {
+ Host::Domain(ref s) => s.heap_size_of_children(),
+ _ => 0,
+ }
+ }
+}
+
+impl<'a> Host<&'a str> {
+ /// Return a copy of `self` that owns an allocated `String` but does not borrow an `&Url`.
+ pub fn to_owned(&self) -> Host<String> {
+ match *self {
+ Host::Domain(domain) => Host::Domain(domain.to_owned()),
+ Host::Ipv4(address) => Host::Ipv4(address),
+ Host::Ipv6(address) => Host::Ipv6(address),
+ }
+ }
+}
+
+impl Host<String> {
+ /// Parse a host: either an IPv6 address in [] square brackets, or a domain.
+ ///
+ /// https://url.spec.whatwg.org/#host-parsing
+ pub fn parse(input: &str) -> Result<Self, ParseError> {
+ if input.starts_with("[") {
+ if !input.ends_with("]") {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ return parse_ipv6addr(&input[1..input.len() - 1]).map(Host::Ipv6)
+ }
+ let domain = percent_decode(input.as_bytes()).decode_utf8_lossy();
+ let domain = try!(idna::domain_to_ascii(&domain));
+ if domain.find(|c| matches!(c,
+ '\0' | '\t' | '\n' | '\r' | ' ' | '#' | '%' | '/' | ':' | '?' | '@' | '[' | '\\' | ']'
+ )).is_some() {
+ return Err(ParseError::InvalidDomainCharacter)
+ }
+ if let Some(address) = try!(parse_ipv4addr(&domain)) {
+ Ok(Host::Ipv4(address))
+ } else {
+ Ok(Host::Domain(domain.into()))
+ }
+ }
+}
+
+impl<S: AsRef<str>> fmt::Display for Host<S> {
+ fn fmt(&self, f: &mut Formatter) -> fmt::Result {
+ match *self {
+ Host::Domain(ref domain) => domain.as_ref().fmt(f),
+ Host::Ipv4(ref addr) => addr.fmt(f),
+ Host::Ipv6(ref addr) => {
+ try!(f.write_str("["));
+ try!(write_ipv6(addr, f));
+ f.write_str("]")
+ }
+ }
+ }
+}
+
+/// This mostly exists because coherence rules don’t allow us to implement
+/// `ToSocketAddrs for (Host<S>, u16)`.
+pub struct HostAndPort<S=String> {
+ pub host: Host<S>,
+ pub port: u16,
+}
+
+impl<'a> HostAndPort<&'a str> {
+ /// Return a copy of `self` that owns an allocated `String` but does not borrow an `&Url`.
+ pub fn to_owned(&self) -> HostAndPort<String> {
+ HostAndPort {
+ host: self.host.to_owned(),
+ port: self.port
+ }
+ }
+}
+
+impl<S: AsRef<str>> ToSocketAddrs for HostAndPort<S> {
+ type Iter = SocketAddrs;
+
+ fn to_socket_addrs(&self) -> io::Result<Self::Iter> {
+ let port = self.port;
+ match self.host {
+ Host::Domain(ref domain) => Ok(SocketAddrs {
+ // FIXME: use std::net::lookup_host when it’s stable.
+ state: SocketAddrsState::Domain(try!((domain.as_ref(), port).to_socket_addrs()))
+ }),
+ Host::Ipv4(address) => Ok(SocketAddrs {
+ state: SocketAddrsState::One(SocketAddr::V4(SocketAddrV4::new(address, port)))
+ }),
+ Host::Ipv6(address) => Ok(SocketAddrs {
+ state: SocketAddrsState::One(SocketAddr::V6(SocketAddrV6::new(address, port, 0, 0)))
+ }),
+ }
+ }
+}
+
+/// Socket addresses for an URL.
+pub struct SocketAddrs {
+ state: SocketAddrsState
+}
+
+enum SocketAddrsState {
+ Domain(vec::IntoIter<SocketAddr>),
+ One(SocketAddr),
+ Done,
+}
+
+impl Iterator for SocketAddrs {
+ type Item = SocketAddr;
+ fn next(&mut self) -> Option<SocketAddr> {
+ match self.state {
+ SocketAddrsState::Domain(ref mut iter) => iter.next(),
+ SocketAddrsState::One(s) => {
+ self.state = SocketAddrsState::Done;
+ Some(s)
+ }
+ SocketAddrsState::Done => None
+ }
+ }
+}
+
+fn write_ipv6(addr: &Ipv6Addr, f: &mut Formatter) -> fmt::Result {
+ let segments = addr.segments();
+ let (compress_start, compress_end) = longest_zero_sequence(&segments);
+ let mut i = 0;
+ while i < 8 {
+ if i == compress_start {
+ try!(f.write_str(":"));
+ if i == 0 {
+ try!(f.write_str(":"));
+ }
+ if compress_end < 8 {
+ i = compress_end;
+ } else {
+ break;
+ }
+ }
+ try!(write!(f, "{:x}", segments[i as usize]));
+ if i < 7 {
+ try!(f.write_str(":"));
+ }
+ i += 1;
+ }
+ Ok(())
+}
+
+fn longest_zero_sequence(pieces: &[u16; 8]) -> (isize, isize) {
+ let mut longest = -1;
+ let mut longest_length = -1;
+ let mut start = -1;
+ macro_rules! finish_sequence(
+ ($end: expr) => {
+ if start >= 0 {
+ let length = $end - start;
+ if length > longest_length {
+ longest = start;
+ longest_length = length;
+ }
+ }
+ };
+ );
+ for i in 0..8 {
+ if pieces[i as usize] == 0 {
+ if start < 0 {
+ start = i;
+ }
+ } else {
+ finish_sequence!(i);
+ start = -1;
+ }
+ }
+ finish_sequence!(8);
+ (longest, longest + longest_length)
+}
+
+
+fn parse_ipv4number(mut input: &str) -> Result<u32, ()> {
+ let mut r = 10;
+ if input.starts_with("0x") || input.starts_with("0X") {
+ input = &input[2..];
+ r = 16;
+ } else if input.len() >= 2 && input.starts_with("0") {
+ input = &input[1..];
+ r = 8;
+ }
+ if input.is_empty() {
+ return Ok(0);
+ }
+ if input.starts_with("+") {
+ return Err(())
+ }
+ match u32::from_str_radix(&input, r) {
+ Ok(number) => Ok(number),
+ Err(_) => Err(()),
+ }
+}
+
+fn parse_ipv4addr(input: &str) -> ParseResult<Option<Ipv4Addr>> {
+ if input.is_empty() {
+ return Ok(None)
+ }
+ let mut parts: Vec<&str> = input.split('.').collect();
+ if parts.last() == Some(&"") {
+ parts.pop();
+ }
+ if parts.len() > 4 {
+ return Ok(None);
+ }
+ let mut numbers: Vec<u32> = Vec::new();
+ for part in parts {
+ if part == "" {
+ return Ok(None);
+ }
+ if let Ok(n) = parse_ipv4number(part) {
+ numbers.push(n);
+ } else {
+ return Ok(None);
+ }
+ }
+ let mut ipv4 = numbers.pop().expect("a non-empty list of numbers");
+ // Equivalent to: ipv4 >= 256 ** (4 − numbers.len())
+ if ipv4 > u32::max_value() >> (8 * numbers.len() as u32) {
+ return Err(ParseError::InvalidIpv4Address);
+ }
+ if numbers.iter().any(|x| *x > 255) {
+ return Err(ParseError::InvalidIpv4Address);
+ }
+ for (counter, n) in numbers.iter().enumerate() {
+ ipv4 += n << (8 * (3 - counter as u32))
+ }
+ Ok(Some(Ipv4Addr::from(ipv4)))
+}
+
+
+fn parse_ipv6addr(input: &str) -> ParseResult<Ipv6Addr> {
+ let input = input.as_bytes();
+ let len = input.len();
+ let mut is_ip_v4 = false;
+ let mut pieces = [0, 0, 0, 0, 0, 0, 0, 0];
+ let mut piece_pointer = 0;
+ let mut compress_pointer = None;
+ let mut i = 0;
+
+ if len < 2 {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+
+ if input[0] == b':' {
+ if input[1] != b':' {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ i = 2;
+ piece_pointer = 1;
+ compress_pointer = Some(1);
+ }
+
+ while i < len {
+ if piece_pointer == 8 {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ if input[i] == b':' {
+ if compress_pointer.is_some() {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ i += 1;
+ piece_pointer += 1;
+ compress_pointer = Some(piece_pointer);
+ continue
+ }
+ let start = i;
+ let end = cmp::min(len, start + 4);
+ let mut value = 0u16;
+ while i < end {
+ match (input[i] as char).to_digit(16) {
+ Some(digit) => {
+ value = value * 0x10 + digit as u16;
+ i += 1;
+ },
+ None => break
+ }
+ }
+ if i < len {
+ match input[i] {
+ b'.' => {
+ if i == start {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ i = start;
+ is_ip_v4 = true;
+ },
+ b':' => {
+ i += 1;
+ if i == len {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ },
+ _ => return Err(ParseError::InvalidIpv6Address)
+ }
+ }
+ if is_ip_v4 {
+ break
+ }
+ pieces[piece_pointer] = value;
+ piece_pointer += 1;
+ }
+
+ if is_ip_v4 {
+ if piece_pointer > 6 {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ let mut dots_seen = 0;
+ while i < len {
+ // FIXME: https://github.com/whatwg/url/commit/1c22aa119c354e0020117e02571cec53f7c01064
+ let mut value = 0u16;
+ while i < len {
+ let digit = match input[i] {
+ c @ b'0' ... b'9' => c - b'0',
+ _ => break
+ };
+ value = value * 10 + digit as u16;
+ if value == 0 || value > 255 {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ }
+ if dots_seen < 3 && !(i < len && input[i] == b'.') {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ pieces[piece_pointer] = pieces[piece_pointer] * 0x100 + value;
+ if dots_seen == 0 || dots_seen == 2 {
+ piece_pointer += 1;
+ }
+ i += 1;
+ if dots_seen == 3 && i < len {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ dots_seen += 1;
+ }
+ }
+
+ match compress_pointer {
+ Some(compress_pointer) => {
+ let mut swaps = piece_pointer - compress_pointer;
+ piece_pointer = 7;
+ while swaps > 0 {
+ pieces[piece_pointer] = pieces[compress_pointer + swaps - 1];
+ pieces[compress_pointer + swaps - 1] = 0;
+ swaps -= 1;
+ piece_pointer -= 1;
+ }
+ }
+ _ => if piece_pointer != 8 {
+ return Err(ParseError::InvalidIpv6Address)
+ }
+ }
+ Ok(Ipv6Addr::new(pieces[0], pieces[1], pieces[2], pieces[3],
+ pieces[4], pieces[5], pieces[6], pieces[7]))
+}
diff --git a/third_party/rust/url/src/lib.rs b/third_party/rust/url/src/lib.rs
new file mode 100644
index 000000000..9378318b4
--- /dev/null
+++ b/third_party/rust/url/src/lib.rs
@@ -0,0 +1,1456 @@
+// Copyright 2013-2015 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+/*!
+
+rust-url is an implementation of the [URL Standard](http://url.spec.whatwg.org/)
+for the [Rust](http://rust-lang.org/) programming language.
+
+It builds with [Cargo](http://crates.io/).
+To use it in your project, add this to your `Cargo.toml` file:
+
+```Cargo
+[dependencies.url]
+git = "https://github.com/servo/rust-url"
+```
+
+Supporting encodings other than UTF-8 in query strings is an optional feature
+that requires [rust-encoding](https://github.com/lifthrasiir/rust-encoding)
+and is off by default.
+You can enable it with
+[Cargo’s *features* mechanism](http://doc.crates.io/manifest.html#the-[features]-section):
+
+```Cargo
+[dependencies.url]
+git = "https://github.com/servo/rust-url"
+features = ["query_encoding"]
+```
+
+… or by passing `--cfg 'feature="query_encoding"'` to rustc.
+
+
+# URL parsing and data structures
+
+First, URL parsing may fail for various reasons and therefore returns a `Result`.
+
+```
+use url::{Url, ParseError};
+
+assert!(Url::parse("http://[:::1]") == Err(ParseError::InvalidIpv6Address))
+```
+
+Let’s parse a valid URL and look at its components.
+
+```
+use url::{Url, Host};
+
+let issue_list_url = Url::parse(
+ "https://github.com/rust-lang/rust/issues?labels=E-easy&state=open"
+).unwrap();
+
+
+assert!(issue_list_url.scheme() == "https");
+assert!(issue_list_url.username() == "");
+assert!(issue_list_url.password() == None);
+assert!(issue_list_url.host_str() == Some("github.com"));
+assert!(issue_list_url.host() == Some(Host::Domain("github.com")));
+assert!(issue_list_url.port() == None);
+assert!(issue_list_url.path() == "/rust-lang/rust/issues");
+assert!(issue_list_url.path_segments().map(|c| c.collect::<Vec<_>>()) ==
+ Some(vec!["rust-lang", "rust", "issues"]));
+assert!(issue_list_url.query() == Some("labels=E-easy&state=open"));
+assert!(issue_list_url.fragment() == None);
+assert!(!issue_list_url.cannot_be_a_base());
+```
+
+Some URLs are said to be *cannot-be-a-base*:
+they don’t have a username, password, host, or port,
+and their "path" is an arbitrary string rather than slash-separated segments:
+
+```
+use url::Url;
+
+let data_url = Url::parse("data:text/plain,Hello?World#").unwrap();
+
+assert!(data_url.cannot_be_a_base());
+assert!(data_url.scheme() == "data");
+assert!(data_url.path() == "text/plain,Hello");
+assert!(data_url.path_segments().is_none());
+assert!(data_url.query() == Some("World"));
+assert!(data_url.fragment() == Some(""));
+```
+
+
+# Base URL
+
+Many contexts allow URL *references* that can be relative to a *base URL*:
+
+```html
+<link rel="stylesheet" href="../main.css">
+```
+
+Since parsed URL are absolute, giving a base is required for parsing relative URLs:
+
+```
+use url::{Url, ParseError};
+
+assert!(Url::parse("../main.css") == Err(ParseError::RelativeUrlWithoutBase))
+```
+
+Use the `join` method on an `Url` to use it as a base URL:
+
+```
+use url::Url;
+
+let this_document = Url::parse("http://servo.github.io/rust-url/url/index.html").unwrap();
+let css_url = this_document.join("../main.css").unwrap();
+assert_eq!(css_url.as_str(), "http://servo.github.io/rust-url/main.css")
+*/
+
+#[cfg(feature="rustc-serialize")] extern crate rustc_serialize;
+#[macro_use] extern crate matches;
+#[cfg(feature="serde")] extern crate serde;
+#[cfg(feature="heapsize")] #[macro_use] extern crate heapsize;
+
+pub extern crate idna;
+
+use encoding::EncodingOverride;
+#[cfg(feature = "heapsize")] use heapsize::HeapSizeOf;
+use host::HostInternal;
+use parser::{Parser, Context, SchemeType, to_u32};
+use percent_encoding::{PATH_SEGMENT_ENCODE_SET, USERINFO_ENCODE_SET,
+ percent_encode, percent_decode, utf8_percent_encode};
+use std::cmp;
+use std::fmt::{self, Write};
+use std::hash;
+use std::io;
+use std::mem;
+use std::net::{ToSocketAddrs, IpAddr};
+use std::ops::{Range, RangeFrom, RangeTo};
+use std::path::{Path, PathBuf};
+use std::str;
+
+pub use origin::{Origin, OpaqueOrigin};
+pub use host::{Host, HostAndPort, SocketAddrs};
+pub use path_segments::PathSegmentsMut;
+pub use parser::ParseError;
+pub use slicing::Position;
+
+mod encoding;
+mod host;
+mod origin;
+mod path_segments;
+mod parser;
+mod slicing;
+
+pub mod form_urlencoded;
+pub mod percent_encoding;
+pub mod quirks;
+
+/// A parsed URL record.
+#[derive(Clone)]
+pub struct Url {
+ /// Syntax in pseudo-BNF:
+ ///
+ /// url = scheme ":" [ hierarchical | non-hierarchical ] [ "?" query ]? [ "#" fragment ]?
+ /// non-hierarchical = non-hierarchical-path
+ /// non-hierarchical-path = /* Does not start with "/" */
+ /// hierarchical = authority? hierarchical-path
+ /// authority = "//" userinfo? host [ ":" port ]?
+ /// userinfo = username [ ":" password ]? "@"
+ /// hierarchical-path = [ "/" path-segment ]+
+ serialization: String,
+
+ // Components
+ scheme_end: u32, // Before ':'
+ username_end: u32, // Before ':' (if a password is given) or '@' (if not)
+ host_start: u32,
+ host_end: u32,
+ host: HostInternal,
+ port: Option<u16>,
+ path_start: u32, // Before initial '/', if any
+ query_start: Option<u32>, // Before '?', unlike Position::QueryStart
+ fragment_start: Option<u32>, // Before '#', unlike Position::FragmentStart
+}
+
+#[cfg(feature = "heapsize")]
+impl HeapSizeOf for Url {
+ fn heap_size_of_children(&self) -> usize {
+ self.serialization.heap_size_of_children()
+ }
+}
+
+/// Full configuration for the URL parser.
+#[derive(Copy, Clone)]
+pub struct ParseOptions<'a> {
+ base_url: Option<&'a Url>,
+ encoding_override: encoding::EncodingOverride,
+ log_syntax_violation: Option<&'a Fn(&'static str)>,
+}
+
+impl<'a> ParseOptions<'a> {
+ /// Change the base URL
+ pub fn base_url(mut self, new: Option<&'a Url>) -> Self {
+ self.base_url = new;
+ self
+ }
+
+ /// Override the character encoding of query strings.
+ /// This is a legacy concept only relevant for HTML.
+ #[cfg(feature = "query_encoding")]
+ pub fn encoding_override(mut self, new: Option<encoding::EncodingRef>) -> Self {
+ self.encoding_override = EncodingOverride::from_opt_encoding(new).to_output_encoding();
+ self
+ }
+
+ /// Call the provided function or closure on non-fatal parse errors.
+ pub fn log_syntax_violation(mut self, new: Option<&'a Fn(&'static str)>) -> Self {
+ self.log_syntax_violation = new;
+ self
+ }
+
+ /// Parse an URL string with the configuration so far.
+ pub fn parse(self, input: &str) -> Result<Url, ::ParseError> {
+ Parser {
+ serialization: String::with_capacity(input.len()),
+ base_url: self.base_url,
+ query_encoding_override: self.encoding_override,
+ log_syntax_violation: self.log_syntax_violation,
+ context: Context::UrlParser,
+ }.parse_url(input)
+ }
+}
+
+impl Url {
+ /// Parse an absolute URL from a string.
+ #[inline]
+ pub fn parse(input: &str) -> Result<Url, ::ParseError> {
+ Url::options().parse(input)
+ }
+
+ /// Parse a string as an URL, with this URL as the base URL.
+ #[inline]
+ pub fn join(&self, input: &str) -> Result<Url, ::ParseError> {
+ Url::options().base_url(Some(self)).parse(input)
+ }
+
+ /// Return a default `ParseOptions` that can fully configure the URL parser.
+ pub fn options<'a>() -> ParseOptions<'a> {
+ ParseOptions {
+ base_url: None,
+ encoding_override: EncodingOverride::utf8(),
+ log_syntax_violation: None,
+ }
+ }
+
+ /// Return the serialization of this URL.
+ ///
+ /// This is fast since that serialization is already stored in the `Url` struct.
+ #[inline]
+ pub fn as_str(&self) -> &str {
+ &self.serialization
+ }
+
+ /// Return the serialization of this URL.
+ ///
+ /// This consumes the `Url` and takes ownership of the `String` stored in it.
+ #[inline]
+ pub fn into_string(self) -> String {
+ self.serialization
+ }
+
+ /// For internal testing, not part of the public API.
+ ///
+ /// Methods of the `Url` struct assume a number of invariants.
+ /// This checks each of these invariants and panic if one is not met.
+ /// This is for testing rust-url itself.
+ #[doc(hidden)]
+ pub fn assert_invariants(&self) {
+ macro_rules! assert {
+ ($x: expr) => {
+ if !$x {
+ panic!("!( {} ) for URL {:?}", stringify!($x), self.serialization)
+ }
+ }
+ }
+
+ macro_rules! assert_eq {
+ ($a: expr, $b: expr) => {
+ {
+ let a = $a;
+ let b = $b;
+ if a != b {
+ panic!("{:?} != {:?} ({} != {}) for URL {:?}",
+ a, b, stringify!($a), stringify!($b), self.serialization)
+ }
+ }
+ }
+ }
+
+ assert!(self.scheme_end >= 1);
+ assert!(matches!(self.byte_at(0), b'a'...b'z' | b'A'...b'Z'));
+ assert!(self.slice(1..self.scheme_end).chars()
+ .all(|c| matches!(c, 'a'...'z' | 'A'...'Z' | '0'...'9' | '+' | '-' | '.')));
+ assert_eq!(self.byte_at(self.scheme_end), b':');
+
+ if self.slice(self.scheme_end + 1 ..).starts_with("//") {
+ // URL with authority
+ match self.byte_at(self.username_end) {
+ b':' => {
+ assert!(self.host_start >= self.username_end + 2);
+ assert_eq!(self.byte_at(self.host_start - 1), b'@');
+ }
+ b'@' => assert!(self.host_start == self.username_end + 1),
+ _ => assert_eq!(self.username_end, self.scheme_end + 3),
+ }
+ assert!(self.host_start >= self.username_end);
+ assert!(self.host_end >= self.host_start);
+ let host_str = self.slice(self.host_start..self.host_end);
+ match self.host {
+ HostInternal::None => assert_eq!(host_str, ""),
+ HostInternal::Ipv4(address) => assert_eq!(host_str, address.to_string()),
+ HostInternal::Ipv6(address) => assert_eq!(host_str, format!("[{}]", address)),
+ HostInternal::Domain => {
+ if SchemeType::from(self.scheme()).is_special() {
+ assert!(!host_str.is_empty())
+ }
+ }
+ }
+ if self.path_start == self.host_end {
+ assert_eq!(self.port, None);
+ } else {
+ assert_eq!(self.byte_at(self.host_end), b':');
+ let port_str = self.slice(self.host_end + 1..self.path_start);
+ assert_eq!(self.port, Some(port_str.parse::<u16>().expect("Couldn't parse port?")));
+ }
+ assert_eq!(self.byte_at(self.path_start), b'/');
+ } else {
+ // Anarchist URL (no authority)
+ assert_eq!(self.username_end, self.scheme_end + 1);
+ assert_eq!(self.host_start, self.scheme_end + 1);
+ assert_eq!(self.host_end, self.scheme_end + 1);
+ assert_eq!(self.host, HostInternal::None);
+ assert_eq!(self.port, None);
+ assert_eq!(self.path_start, self.scheme_end + 1);
+ }
+ if let Some(start) = self.query_start {
+ assert!(start > self.path_start);
+ assert_eq!(self.byte_at(start), b'?');
+ }
+ if let Some(start) = self.fragment_start {
+ assert!(start > self.path_start);
+ assert_eq!(self.byte_at(start), b'#');
+ }
+ if let (Some(query_start), Some(fragment_start)) = (self.query_start, self.fragment_start) {
+ assert!(fragment_start > query_start);
+ }
+
+ let other = Url::parse(self.as_str()).expect("Failed to parse myself?");
+ assert_eq!(&self.serialization, &other.serialization);
+ assert_eq!(self.scheme_end, other.scheme_end);
+ assert_eq!(self.username_end, other.username_end);
+ assert_eq!(self.host_start, other.host_start);
+ assert_eq!(self.host_end, other.host_end);
+ assert!(self.host == other.host ||
+ // XXX No host round-trips to empty host.
+ // See https://github.com/whatwg/url/issues/79
+ (self.host_str(), other.host_str()) == (None, Some("")));
+ assert_eq!(self.port, other.port);
+ assert_eq!(self.path_start, other.path_start);
+ assert_eq!(self.query_start, other.query_start);
+ assert_eq!(self.fragment_start, other.fragment_start);
+ }
+
+ /// Return the origin of this URL (https://url.spec.whatwg.org/#origin)
+ ///
+ /// Note: this returns an opaque origin for `file:` URLs, which causes
+ /// `url.origin() != url.origin()`.
+ ///
+ /// # Examples
+ ///
+ /// URL with `ftp` scheme:
+ ///
+ /// ```rust
+ /// use url::{Host, Origin, Url};
+ ///
+ /// let url = Url::parse("ftp://example.com/foo").unwrap();
+ /// assert_eq!(url.origin(),
+ /// Origin::Tuple("ftp".into(),
+ /// Host::Domain("example.com".into()),
+ /// 21));
+ /// ```
+ ///
+ /// URL with `blob` scheme:
+ ///
+ /// ```rust
+ /// use url::{Host, Origin, Url};
+ ///
+ /// let url = Url::parse("blob:https://example.com/foo").unwrap();
+ /// assert_eq!(url.origin(),
+ /// Origin::Tuple("https".into(),
+ /// Host::Domain("example.com".into()),
+ /// 443));
+ /// ```
+ ///
+ /// URL with `file` scheme:
+ ///
+ /// ```rust
+ /// use url::{Host, Origin, Url};
+ ///
+ /// let url = Url::parse("file:///tmp/foo").unwrap();
+ /// assert!(!url.origin().is_tuple());
+ ///
+ /// let other_url = Url::parse("file:///tmp/foo").unwrap();
+ /// assert!(url.origin() != other_url.origin());
+ /// ```
+ ///
+ /// URL with other scheme:
+ ///
+ /// ```rust
+ /// use url::{Host, Origin, Url};
+ ///
+ /// let url = Url::parse("foo:bar").unwrap();
+ /// assert!(!url.origin().is_tuple());
+ /// ```
+ #[inline]
+ pub fn origin(&self) -> Origin {
+ origin::url_origin(self)
+ }
+
+ /// Return the scheme of this URL, lower-cased, as an ASCII string without the ':' delimiter.
+ ///
+ /// # Examples
+ ///
+ /// ```
+ /// use url::Url;
+ ///
+ /// let url = Url::parse("file:///tmp/foo").unwrap();
+ /// assert_eq!(url.scheme(), "file");
+ /// ```
+ #[inline]
+ pub fn scheme(&self) -> &str {
+ self.slice(..self.scheme_end)
+ }
+
+ /// Return whether the URL has an 'authority',
+ /// which can contain a username, password, host, and port number.
+ ///
+ /// URLs that do *not* are either path-only like `unix:/run/foo.socket`
+ /// or cannot-be-a-base like `data:text/plain,Stuff`.
+ #[inline]
+ pub fn has_authority(&self) -> bool {
+ debug_assert!(self.byte_at(self.scheme_end) == b':');
+ self.slice(self.scheme_end..).starts_with("://")
+ }
+
+ /// Return whether this URL is a cannot-be-a-base URL,
+ /// meaning that parsing a relative URL string with this URL as the base will return an error.
+ ///
+ /// This is the case if the scheme and `:` delimiter are not followed by a `/` slash,
+ /// as is typically the case of `data:` and `mailto:` URLs.
+ #[inline]
+ pub fn cannot_be_a_base(&self) -> bool {
+ self.byte_at(self.path_start) != b'/'
+ }
+
+ /// Return the username for this URL (typically the empty string)
+ /// as a percent-encoded ASCII string.
+ ///
+ /// # Examples
+ ///
+ /// ```
+ /// use url::Url;
+ ///
+ /// let url = Url::parse("ftp://rms@example.com").unwrap();
+ /// assert_eq!(url.username(), "rms");
+ ///
+ /// let url = Url::parse("ftp://:secret123@example.com").unwrap();
+ /// assert_eq!(url.username(), "");
+ ///
+ /// let url = Url::parse("https://example.com").unwrap();
+ /// assert_eq!(url.username(), "");
+ /// ```
+ pub fn username(&self) -> &str {
+ if self.has_authority() {
+ self.slice(self.scheme_end + ("://".len() as u32)..self.username_end)
+ } else {
+ ""
+ }
+ }
+
+ /// Return the password for this URL, if any, as a percent-encoded ASCII string.
+ ///
+ /// # Examples
+ ///
+ /// ```
+ /// use url::Url;
+ ///
+ /// let url = Url::parse("ftp://rms:secret123@example.com").unwrap();
+ /// assert_eq!(url.password(), Some("secret123"));
+ ///
+ /// let url = Url::parse("ftp://:secret123@example.com").unwrap();
+ /// assert_eq!(url.password(), Some("secret123"));
+ ///
+ /// let url = Url::parse("ftp://rms@example.com").unwrap();
+ /// assert_eq!(url.password(), None);
+ ///
+ /// let url = Url::parse("https://example.com").unwrap();
+ /// assert_eq!(url.password(), None);
+ /// ```
+ pub fn password(&self) -> Option<&str> {
+ // This ':' is not the one marking a port number since a host can not be empty.
+ // (Except for file: URLs, which do not have port numbers.)
+ if self.has_authority() && self.byte_at(self.username_end) == b':' {
+ debug_assert!(self.byte_at(self.host_start - 1) == b'@');
+ Some(self.slice(self.username_end + 1..self.host_start - 1))
+ } else {
+ None
+ }
+ }
+
+ /// Equivalent to `url.host().is_some()`.
+ pub fn has_host(&self) -> bool {
+ !matches!(self.host, HostInternal::None)
+ }
+
+ /// Return the string representation of the host (domain or IP address) for this URL, if any.
+ ///
+ /// Non-ASCII domains are punycode-encoded per IDNA.
+ /// IPv6 addresses are given between `[` and `]` brackets.
+ ///
+ /// Cannot-be-a-base URLs (typical of `data:` and `mailto:`) and some `file:` URLs
+ /// don’t have a host.
+ ///
+ /// See also the `host` method.
+ pub fn host_str(&self) -> Option<&str> {
+ if self.has_host() {
+ Some(self.slice(self.host_start..self.host_end))
+ } else {
+ None
+ }
+ }
+
+ /// Return the parsed representation of the host for this URL.
+ /// Non-ASCII domain labels are punycode-encoded per IDNA.
+ ///
+ /// Cannot-be-a-base URLs (typical of `data:` and `mailto:`) and some `file:` URLs
+ /// don’t have a host.
+ ///
+ /// See also the `host_str` method.
+ pub fn host(&self) -> Option<Host<&str>> {
+ match self.host {
+ HostInternal::None => None,
+ HostInternal::Domain => Some(Host::Domain(self.slice(self.host_start..self.host_end))),
+ HostInternal::Ipv4(address) => Some(Host::Ipv4(address)),
+ HostInternal::Ipv6(address) => Some(Host::Ipv6(address)),
+ }
+ }
+
+ /// If this URL has a host and it is a domain name (not an IP address), return it.
+ pub fn domain(&self) -> Option<&str> {
+ match self.host {
+ HostInternal::Domain => Some(self.slice(self.host_start..self.host_end)),
+ _ => None,
+ }
+ }
+
+ /// Return the port number for this URL, if any.
+ #[inline]
+ pub fn port(&self) -> Option<u16> {
+ self.port
+ }
+
+ /// Return the port number for this URL, or the default port number if it is known.
+ ///
+ /// This method only knows the default port number
+ /// of the `http`, `https`, `ws`, `wss`, `ftp`, and `gopher` schemes.
+ ///
+ /// For URLs in these schemes, this method always returns `Some(_)`.
+ /// For other schemes, it is the same as `Url::port()`.
+ #[inline]
+ pub fn port_or_known_default(&self) -> Option<u16> {
+ self.port.or_else(|| parser::default_port(self.scheme()))
+ }
+
+ /// If the URL has a host, return something that implements `ToSocketAddrs`.
+ ///
+ /// If the URL has no port number and the scheme’s default port number is not known
+ /// (see `Url::port_or_known_default`),
+ /// the closure is called to obtain a port number.
+ /// Typically, this closure can match on the result `Url::scheme`
+ /// to have per-scheme default port numbers,
+ /// and panic for schemes it’s not prepared to handle.
+ /// For example:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// # use std::net::TcpStream;
+ /// # use std::io;
+ ///
+ /// fn connect(url: &Url) -> io::Result<TcpStream> {
+ /// TcpStream::connect(try!(url.with_default_port(default_port)))
+ /// }
+ ///
+ /// fn default_port(url: &Url) -> Result<u16, ()> {
+ /// match url.scheme() {
+ /// "git" => Ok(9418),
+ /// "git+ssh" => Ok(22),
+ /// "git+https" => Ok(443),
+ /// "git+http" => Ok(80),
+ /// _ => Err(()),
+ /// }
+ /// }
+ /// ```
+ pub fn with_default_port<F>(&self, f: F) -> io::Result<HostAndPort<&str>>
+ where F: FnOnce(&Url) -> Result<u16, ()> {
+ Ok(HostAndPort {
+ host: try!(self.host()
+ .ok_or(())
+ .or_else(|()| io_error("URL has no host"))),
+ port: try!(self.port_or_known_default()
+ .ok_or(())
+ .or_else(|()| f(self))
+ .or_else(|()| io_error("URL has no port number")))
+ })
+ }
+
+ /// Return the path for this URL, as a percent-encoded ASCII string.
+ /// For cannot-be-a-base URLs, this is an arbitrary string that doesn’t start with '/'.
+ /// For other URLs, this starts with a '/' slash
+ /// and continues with slash-separated path segments.
+ pub fn path(&self) -> &str {
+ match (self.query_start, self.fragment_start) {
+ (None, None) => self.slice(self.path_start..),
+ (Some(next_component_start), _) |
+ (None, Some(next_component_start)) => {
+ self.slice(self.path_start..next_component_start)
+ }
+ }
+ }
+
+ /// Unless this URL is cannot-be-a-base,
+ /// return an iterator of '/' slash-separated path segments,
+ /// each as a percent-encoded ASCII string.
+ ///
+ /// Return `None` for cannot-be-a-base URLs.
+ ///
+ /// When `Some` is returned, the iterator always contains at least one string
+ /// (which may be empty).
+ pub fn path_segments(&self) -> Option<str::Split<char>> {
+ let path = self.path();
+ if path.starts_with('/') {
+ Some(path[1..].split('/'))
+ } else {
+ None
+ }
+ }
+
+ /// Return this URL’s query string, if any, as a percent-encoded ASCII string.
+ pub fn query(&self) -> Option<&str> {
+ match (self.query_start, self.fragment_start) {
+ (None, _) => None,
+ (Some(query_start), None) => {
+ debug_assert!(self.byte_at(query_start) == b'?');
+ Some(self.slice(query_start + 1..))
+ }
+ (Some(query_start), Some(fragment_start)) => {
+ debug_assert!(self.byte_at(query_start) == b'?');
+ Some(self.slice(query_start + 1..fragment_start))
+ }
+ }
+ }
+
+ /// Parse the URL’s query string, if any, as `application/x-www-form-urlencoded`
+ /// and return an iterator of (key, value) pairs.
+ #[inline]
+ pub fn query_pairs(&self) -> form_urlencoded::Parse {
+ form_urlencoded::parse(self.query().unwrap_or("").as_bytes())
+ }
+
+ /// Return this URL’s fragment identifier, if any.
+ ///
+ /// **Note:** the parser did *not* percent-encode this component,
+ /// but the input may have been percent-encoded already.
+ pub fn fragment(&self) -> Option<&str> {
+ self.fragment_start.map(|start| {
+ debug_assert!(self.byte_at(start) == b'#');
+ self.slice(start + 1..)
+ })
+ }
+
+ fn mutate<F: FnOnce(&mut Parser) -> R, R>(&mut self, f: F) -> R {
+ let mut parser = Parser::for_setter(mem::replace(&mut self.serialization, String::new()));
+ let result = f(&mut parser);
+ self.serialization = parser.serialization;
+ result
+ }
+
+ /// Change this URL’s fragment identifier.
+ pub fn set_fragment(&mut self, fragment: Option<&str>) {
+ // Remove any previous fragment
+ if let Some(start) = self.fragment_start {
+ debug_assert!(self.byte_at(start) == b'#');
+ self.serialization.truncate(start as usize);
+ }
+ // Write the new one
+ if let Some(input) = fragment {
+ self.fragment_start = Some(to_u32(self.serialization.len()).unwrap());
+ self.serialization.push('#');
+ self.mutate(|parser| parser.parse_fragment(parser::Input::new(input)))
+ } else {
+ self.fragment_start = None
+ }
+ }
+
+ fn take_fragment(&mut self) -> Option<String> {
+ self.fragment_start.take().map(|start| {
+ debug_assert!(self.byte_at(start) == b'#');
+ let fragment = self.slice(start + 1..).to_owned();
+ self.serialization.truncate(start as usize);
+ fragment
+ })
+ }
+
+ fn restore_already_parsed_fragment(&mut self, fragment: Option<String>) {
+ if let Some(ref fragment) = fragment {
+ assert!(self.fragment_start.is_none());
+ self.fragment_start = Some(to_u32(self.serialization.len()).unwrap());
+ self.serialization.push('#');
+ self.serialization.push_str(fragment);
+ }
+ }
+
+ /// Change this URL’s query string.
+ pub fn set_query(&mut self, query: Option<&str>) {
+ let fragment = self.take_fragment();
+
+ // Remove any previous query
+ if let Some(start) = self.query_start.take() {
+ debug_assert!(self.byte_at(start) == b'?');
+ self.serialization.truncate(start as usize);
+ }
+ // Write the new query, if any
+ if let Some(input) = query {
+ self.query_start = Some(to_u32(self.serialization.len()).unwrap());
+ self.serialization.push('?');
+ let scheme_end = self.scheme_end;
+ self.mutate(|parser| parser.parse_query(scheme_end, parser::Input::new(input)));
+ }
+
+ self.restore_already_parsed_fragment(fragment);
+ }
+
+ /// Manipulate this URL’s query string, viewed as a sequence of name/value pairs
+ /// in `application/x-www-form-urlencoded` syntax.
+ ///
+ /// The return value has a method-chaining API:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// let mut url = Url::parse("https://example.net?lang=fr#nav").unwrap();
+ /// assert_eq!(url.query(), Some("lang=fr"));
+ ///
+ /// url.query_pairs_mut().append_pair("foo", "bar");
+ /// assert_eq!(url.query(), Some("lang=fr&foo=bar"));
+ /// assert_eq!(url.as_str(), "https://example.net/?lang=fr&foo=bar#nav");
+ ///
+ /// url.query_pairs_mut()
+ /// .clear()
+ /// .append_pair("foo", "bar & baz")
+ /// .append_pair("saisons", "Été+hiver");
+ /// assert_eq!(url.query(), Some("foo=bar+%26+baz&saisons=%C3%89t%C3%A9%2Bhiver"));
+ /// assert_eq!(url.as_str(),
+ /// "https://example.net/?foo=bar+%26+baz&saisons=%C3%89t%C3%A9%2Bhiver#nav");
+ /// ```
+ ///
+ /// Note: `url.query_pairs_mut().clear();` is equivalent to `url.set_query(Some(""))`,
+ /// not `url.set_query(None)`.
+ ///
+ /// The state of `Url` is unspecified if this return value is leaked without being dropped.
+ pub fn query_pairs_mut(&mut self) -> form_urlencoded::Serializer<UrlQuery> {
+ let fragment = self.take_fragment();
+
+ let query_start;
+ if let Some(start) = self.query_start {
+ debug_assert!(self.byte_at(start) == b'?');
+ query_start = start as usize;
+ } else {
+ query_start = self.serialization.len();
+ self.query_start = Some(to_u32(query_start).unwrap());
+ self.serialization.push('?');
+ }
+
+ let query = UrlQuery { url: self, fragment: fragment };
+ form_urlencoded::Serializer::for_suffix(query, query_start + "?".len())
+ }
+
+ fn take_after_path(&mut self) -> String {
+ match (self.query_start, self.fragment_start) {
+ (Some(i), _) | (None, Some(i)) => {
+ let after_path = self.slice(i..).to_owned();
+ self.serialization.truncate(i as usize);
+ after_path
+ },
+ (None, None) => String::new(),
+ }
+ }
+
+ /// Change this URL’s path.
+ pub fn set_path(&mut self, mut path: &str) {
+ let after_path = self.take_after_path();
+ let old_after_path_pos = to_u32(self.serialization.len()).unwrap();
+ let cannot_be_a_base = self.cannot_be_a_base();
+ let scheme_type = SchemeType::from(self.scheme());
+ self.serialization.truncate(self.path_start as usize);
+ self.mutate(|parser| {
+ if cannot_be_a_base {
+ if path.starts_with('/') {
+ parser.serialization.push_str("%2F");
+ path = &path[1..];
+ }
+ parser.parse_cannot_be_a_base_path(parser::Input::new(path));
+ } else {
+ let mut has_host = true; // FIXME
+ parser.parse_path_start(scheme_type, &mut has_host, parser::Input::new(path));
+ }
+ });
+ self.restore_after_path(old_after_path_pos, &after_path);
+ }
+
+ /// Return an object with methods to manipulate this URL’s path segments.
+ ///
+ /// Return `Err(())` if this URl is cannot-be-a-base.
+ pub fn path_segments_mut(&mut self) -> Result<PathSegmentsMut, ()> {
+ if self.cannot_be_a_base() {
+ Err(())
+ } else {
+ Ok(path_segments::new(self))
+ }
+ }
+
+ fn restore_after_path(&mut self, old_after_path_position: u32, after_path: &str) {
+ let new_after_path_position = to_u32(self.serialization.len()).unwrap();
+ let adjust = |index: &mut u32| {
+ *index -= old_after_path_position;
+ *index += new_after_path_position;
+ };
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+ self.serialization.push_str(after_path)
+ }
+
+ /// Change this URL’s port number.
+ ///
+ /// If this URL is cannot-be-a-base, does not have a host, or has the `file` scheme;
+ /// do nothing and return `Err`.
+ pub fn set_port(&mut self, mut port: Option<u16>) -> Result<(), ()> {
+ if !self.has_host() || self.scheme() == "file" {
+ return Err(())
+ }
+ if port.is_some() && port == parser::default_port(self.scheme()) {
+ port = None
+ }
+ self.set_port_internal(port);
+ Ok(())
+ }
+
+ fn set_port_internal(&mut self, port: Option<u16>) {
+ match (self.port, port) {
+ (None, None) => {}
+ (Some(_), None) => {
+ self.serialization.drain(self.host_end as usize .. self.path_start as usize);
+ let offset = self.path_start - self.host_end;
+ self.path_start = self.host_end;
+ if let Some(ref mut index) = self.query_start { *index -= offset }
+ if let Some(ref mut index) = self.fragment_start { *index -= offset }
+ }
+ (Some(old), Some(new)) if old == new => {}
+ (_, Some(new)) => {
+ let path_and_after = self.slice(self.path_start..).to_owned();
+ self.serialization.truncate(self.host_end as usize);
+ write!(&mut self.serialization, ":{}", new).unwrap();
+ let old_path_start = self.path_start;
+ let new_path_start = to_u32(self.serialization.len()).unwrap();
+ self.path_start = new_path_start;
+ let adjust = |index: &mut u32| {
+ *index -= old_path_start;
+ *index += new_path_start;
+ };
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+ self.serialization.push_str(&path_and_after);
+ }
+ }
+ self.port = port;
+ }
+
+ /// Change this URL’s host.
+ ///
+ /// If this URL is cannot-be-a-base or there is an error parsing the given `host`,
+ /// do nothing and return `Err`.
+ ///
+ /// Removing the host (calling this with `None`)
+ /// will also remove any username, password, and port number.
+ pub fn set_host(&mut self, host: Option<&str>) -> Result<(), ParseError> {
+ if self.cannot_be_a_base() {
+ return Err(ParseError::SetHostOnCannotBeABaseUrl)
+ }
+
+ if let Some(host) = host {
+ self.set_host_internal(try!(Host::parse(host)), None)
+ } else if self.has_host() {
+ debug_assert!(self.byte_at(self.scheme_end) == b':');
+ debug_assert!(self.byte_at(self.path_start) == b'/');
+ let new_path_start = self.scheme_end + 1;
+ self.serialization.drain(self.path_start as usize..new_path_start as usize);
+ let offset = self.path_start - new_path_start;
+ self.path_start = new_path_start;
+ self.username_end = new_path_start;
+ self.host_start = new_path_start;
+ self.host_end = new_path_start;
+ self.port = None;
+ if let Some(ref mut index) = self.query_start { *index -= offset }
+ if let Some(ref mut index) = self.fragment_start { *index -= offset }
+ }
+ Ok(())
+ }
+
+ /// opt_new_port: None means leave unchanged, Some(None) means remove any port number.
+ fn set_host_internal(&mut self, host: Host<String>, opt_new_port: Option<Option<u16>>) {
+ let old_suffix_pos = if opt_new_port.is_some() { self.path_start } else { self.host_end };
+ let suffix = self.slice(old_suffix_pos..).to_owned();
+ self.serialization.truncate(self.host_start as usize);
+ if !self.has_authority() {
+ debug_assert!(self.slice(self.scheme_end..self.host_start) == ":");
+ debug_assert!(self.username_end == self.host_start);
+ self.serialization.push('/');
+ self.serialization.push('/');
+ self.username_end += 2;
+ self.host_start += 2;
+ }
+ write!(&mut self.serialization, "{}", host).unwrap();
+ self.host_end = to_u32(self.serialization.len()).unwrap();
+ self.host = host.into();
+
+ if let Some(new_port) = opt_new_port {
+ self.port = new_port;
+ if let Some(port) = new_port {
+ write!(&mut self.serialization, ":{}", port).unwrap();
+ }
+ }
+ let new_suffix_pos = to_u32(self.serialization.len()).unwrap();
+ self.serialization.push_str(&suffix);
+
+ let adjust = |index: &mut u32| {
+ *index -= old_suffix_pos;
+ *index += new_suffix_pos;
+ };
+ adjust(&mut self.path_start);
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+ }
+
+ /// Change this URL’s host to the given IP address.
+ ///
+ /// If this URL is cannot-be-a-base, do nothing and return `Err`.
+ ///
+ /// Compared to `Url::set_host`, this skips the host parser.
+ pub fn set_ip_host(&mut self, address: IpAddr) -> Result<(), ()> {
+ if self.cannot_be_a_base() {
+ return Err(())
+ }
+
+ let address = match address {
+ IpAddr::V4(address) => Host::Ipv4(address),
+ IpAddr::V6(address) => Host::Ipv6(address),
+ };
+ self.set_host_internal(address, None);
+ Ok(())
+ }
+
+ /// Change this URL’s password.
+ ///
+ /// If this URL is cannot-be-a-base or does not have a host, do nothing and return `Err`.
+ pub fn set_password(&mut self, password: Option<&str>) -> Result<(), ()> {
+ if !self.has_host() {
+ return Err(())
+ }
+ if let Some(password) = password {
+ let host_and_after = self.slice(self.host_start..).to_owned();
+ self.serialization.truncate(self.username_end as usize);
+ self.serialization.push(':');
+ self.serialization.extend(utf8_percent_encode(password, USERINFO_ENCODE_SET));
+ self.serialization.push('@');
+
+ let old_host_start = self.host_start;
+ let new_host_start = to_u32(self.serialization.len()).unwrap();
+ let adjust = |index: &mut u32| {
+ *index -= old_host_start;
+ *index += new_host_start;
+ };
+ self.host_start = new_host_start;
+ adjust(&mut self.host_end);
+ adjust(&mut self.path_start);
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+
+ self.serialization.push_str(&host_and_after);
+ } else if self.byte_at(self.username_end) == b':' { // If there is a password to remove
+ let has_username_or_password = self.byte_at(self.host_start - 1) == b'@';
+ debug_assert!(has_username_or_password);
+ let username_start = self.scheme_end + 3;
+ let empty_username = username_start == self.username_end;
+ let start = self.username_end; // Remove the ':'
+ let end = if empty_username {
+ self.host_start // Remove the '@' as well
+ } else {
+ self.host_start - 1 // Keep the '@' to separate the username from the host
+ };
+ self.serialization.drain(start as usize .. end as usize);
+ let offset = end - start;
+ self.host_start -= offset;
+ self.host_end -= offset;
+ self.path_start -= offset;
+ if let Some(ref mut index) = self.query_start { *index -= offset }
+ if let Some(ref mut index) = self.fragment_start { *index -= offset }
+ }
+ Ok(())
+ }
+
+ /// Change this URL’s username.
+ ///
+ /// If this URL is cannot-be-a-base or does not have a host, do nothing and return `Err`.
+ pub fn set_username(&mut self, username: &str) -> Result<(), ()> {
+ if !self.has_host() {
+ return Err(())
+ }
+ let username_start = self.scheme_end + 3;
+ debug_assert!(self.slice(self.scheme_end..username_start) == "://");
+ if self.slice(username_start..self.username_end) == username {
+ return Ok(())
+ }
+ let after_username = self.slice(self.username_end..).to_owned();
+ self.serialization.truncate(username_start as usize);
+ self.serialization.extend(utf8_percent_encode(username, USERINFO_ENCODE_SET));
+
+ let mut removed_bytes = self.username_end;
+ self.username_end = to_u32(self.serialization.len()).unwrap();
+ let mut added_bytes = self.username_end;
+
+ let new_username_is_empty = self.username_end == username_start;
+ match (new_username_is_empty, after_username.chars().next()) {
+ (true, Some('@')) => {
+ removed_bytes += 1;
+ self.serialization.push_str(&after_username[1..]);
+ }
+ (false, Some('@')) | (_, Some(':')) | (true, _) => {
+ self.serialization.push_str(&after_username);
+ }
+ (false, _) => {
+ added_bytes += 1;
+ self.serialization.push('@');
+ self.serialization.push_str(&after_username);
+ }
+ }
+
+ let adjust = |index: &mut u32| {
+ *index -= removed_bytes;
+ *index += added_bytes;
+ };
+ adjust(&mut self.host_start);
+ adjust(&mut self.host_end);
+ adjust(&mut self.path_start);
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+ Ok(())
+ }
+
+ /// Change this URL’s scheme.
+ ///
+ /// Do nothing and return `Err` if:
+ /// * The new scheme is not in `[a-zA-Z][a-zA-Z0-9+.-]+`
+ /// * This URL is cannot-be-a-base and the new scheme is one of
+ /// `http`, `https`, `ws`, `wss`, `ftp`, or `gopher`
+ pub fn set_scheme(&mut self, scheme: &str) -> Result<(), ()> {
+ let mut parser = Parser::for_setter(String::new());
+ let remaining = try!(parser.parse_scheme(parser::Input::new(scheme)));
+ if !remaining.is_empty() ||
+ (!self.has_host() && SchemeType::from(&parser.serialization).is_special()) {
+ return Err(())
+ }
+ let old_scheme_end = self.scheme_end;
+ let new_scheme_end = to_u32(parser.serialization.len()).unwrap();
+ let adjust = |index: &mut u32| {
+ *index -= old_scheme_end;
+ *index += new_scheme_end;
+ };
+
+ self.scheme_end = new_scheme_end;
+ adjust(&mut self.username_end);
+ adjust(&mut self.host_start);
+ adjust(&mut self.host_end);
+ adjust(&mut self.path_start);
+ if let Some(ref mut index) = self.query_start { adjust(index) }
+ if let Some(ref mut index) = self.fragment_start { adjust(index) }
+
+ parser.serialization.push_str(self.slice(old_scheme_end..));
+ self.serialization = parser.serialization;
+ Ok(())
+ }
+
+ /// Convert a file name as `std::path::Path` into an URL in the `file` scheme.
+ ///
+ /// This returns `Err` if the given path is not absolute or,
+ /// on Windows, if the prefix is not a disk prefix (e.g. `C:`).
+ pub fn from_file_path<P: AsRef<Path>>(path: P) -> Result<Url, ()> {
+ let mut serialization = "file://".to_owned();
+ let path_start = serialization.len() as u32;
+ try!(path_to_file_url_segments(path.as_ref(), &mut serialization));
+ Ok(Url {
+ serialization: serialization,
+ scheme_end: "file".len() as u32,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: None,
+ fragment_start: None,
+ })
+ }
+
+ /// Convert a directory name as `std::path::Path` into an URL in the `file` scheme.
+ ///
+ /// This returns `Err` if the given path is not absolute or,
+ /// on Windows, if the prefix is not a disk prefix (e.g. `C:`).
+ ///
+ /// Compared to `from_file_path`, this ensure that URL’s the path has a trailing slash
+ /// so that the entire path is considered when using this URL as a base URL.
+ ///
+ /// For example:
+ ///
+ /// * `"index.html"` parsed with `Url::from_directory_path(Path::new("/var/www"))`
+ /// as the base URL is `file:///var/www/index.html`
+ /// * `"index.html"` parsed with `Url::from_file_path(Path::new("/var/www"))`
+ /// as the base URL is `file:///var/index.html`, which might not be what was intended.
+ ///
+ /// Note that `std::path` does not consider trailing slashes significant
+ /// and usually does not include them (e.g. in `Path::parent()`).
+ pub fn from_directory_path<P: AsRef<Path>>(path: P) -> Result<Url, ()> {
+ let mut url = try!(Url::from_file_path(path));
+ if !url.serialization.ends_with('/') {
+ url.serialization.push('/')
+ }
+ Ok(url)
+ }
+
+ /// Assuming the URL is in the `file` scheme or similar,
+ /// convert its path to an absolute `std::path::Path`.
+ ///
+ /// **Note:** This does not actually check the URL’s `scheme`,
+ /// and may give nonsensical results for other schemes.
+ /// It is the user’s responsibility to check the URL’s scheme before calling this.
+ ///
+ /// ```
+ /// # use url::Url;
+ /// # let url = Url::parse("file:///etc/passwd").unwrap();
+ /// let path = url.to_file_path();
+ /// ```
+ ///
+ /// Returns `Err` if the host is neither empty nor `"localhost"`,
+ /// or if `Path::new_opt()` returns `None`.
+ /// (That is, if the percent-decoded path contains a NUL byte or,
+ /// for a Windows path, is not UTF-8.)
+ #[inline]
+ pub fn to_file_path(&self) -> Result<PathBuf, ()> {
+ // FIXME: Figure out what to do w.r.t host.
+ if matches!(self.host(), None | Some(Host::Domain("localhost"))) {
+ if let Some(segments) = self.path_segments() {
+ return file_url_segments_to_pathbuf(segments)
+ }
+ }
+ Err(())
+ }
+
+ // Private helper methods:
+
+ #[inline]
+ fn slice<R>(&self, range: R) -> &str where R: RangeArg {
+ range.slice_of(&self.serialization)
+ }
+
+ #[inline]
+ fn byte_at(&self, i: u32) -> u8 {
+ self.serialization.as_bytes()[i as usize]
+ }
+}
+
+/// Return an error if `Url::host` or `Url::port_or_known_default` return `None`.
+impl ToSocketAddrs for Url {
+ type Iter = SocketAddrs;
+
+ fn to_socket_addrs(&self) -> io::Result<Self::Iter> {
+ try!(self.with_default_port(|_| Err(()))).to_socket_addrs()
+ }
+}
+
+/// Parse a string as an URL, without a base URL or encoding override.
+impl str::FromStr for Url {
+ type Err = ParseError;
+
+ #[inline]
+ fn from_str(input: &str) -> Result<Url, ::ParseError> {
+ Url::parse(input)
+ }
+}
+
+/// Display the serialization of this URL.
+impl fmt::Display for Url {
+ #[inline]
+ fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
+ fmt::Display::fmt(&self.serialization, formatter)
+ }
+}
+
+/// Debug the serialization of this URL.
+impl fmt::Debug for Url {
+ #[inline]
+ fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
+ fmt::Debug::fmt(&self.serialization, formatter)
+ }
+}
+
+/// URLs compare like their serialization.
+impl Eq for Url {}
+
+/// URLs compare like their serialization.
+impl PartialEq for Url {
+ #[inline]
+ fn eq(&self, other: &Self) -> bool {
+ self.serialization == other.serialization
+ }
+}
+
+/// URLs compare like their serialization.
+impl Ord for Url {
+ #[inline]
+ fn cmp(&self, other: &Self) -> cmp::Ordering {
+ self.serialization.cmp(&other.serialization)
+ }
+}
+
+/// URLs compare like their serialization.
+impl PartialOrd for Url {
+ #[inline]
+ fn partial_cmp(&self, other: &Self) -> Option<cmp::Ordering> {
+ self.serialization.partial_cmp(&other.serialization)
+ }
+}
+
+/// URLs hash like their serialization.
+impl hash::Hash for Url {
+ #[inline]
+ fn hash<H>(&self, state: &mut H) where H: hash::Hasher {
+ hash::Hash::hash(&self.serialization, state)
+ }
+}
+
+/// Return the serialization of this URL.
+impl AsRef<str> for Url {
+ #[inline]
+ fn as_ref(&self) -> &str {
+ &self.serialization
+ }
+}
+
+trait RangeArg {
+ fn slice_of<'a>(&self, s: &'a str) -> &'a str;
+}
+
+impl RangeArg for Range<u32> {
+ #[inline]
+ fn slice_of<'a>(&self, s: &'a str) -> &'a str {
+ &s[self.start as usize .. self.end as usize]
+ }
+}
+
+impl RangeArg for RangeFrom<u32> {
+ #[inline]
+ fn slice_of<'a>(&self, s: &'a str) -> &'a str {
+ &s[self.start as usize ..]
+ }
+}
+
+impl RangeArg for RangeTo<u32> {
+ #[inline]
+ fn slice_of<'a>(&self, s: &'a str) -> &'a str {
+ &s[.. self.end as usize]
+ }
+}
+
+#[cfg(feature="rustc-serialize")]
+impl rustc_serialize::Encodable for Url {
+ fn encode<S: rustc_serialize::Encoder>(&self, encoder: &mut S) -> Result<(), S::Error> {
+ encoder.emit_str(self.as_str())
+ }
+}
+
+
+#[cfg(feature="rustc-serialize")]
+impl rustc_serialize::Decodable for Url {
+ fn decode<D: rustc_serialize::Decoder>(decoder: &mut D) -> Result<Url, D::Error> {
+ Url::parse(&*try!(decoder.read_str())).map_err(|error| {
+ decoder.error(&format!("URL parsing error: {}", error))
+ })
+ }
+}
+
+/// Serializes this URL into a `serde` stream.
+///
+/// This implementation is only available if the `serde` Cargo feature is enabled.
+#[cfg(feature="serde")]
+impl serde::Serialize for Url {
+ fn serialize<S>(&self, serializer: &mut S) -> Result<(), S::Error> where S: serde::Serializer {
+ format!("{}", self).serialize(serializer)
+ }
+}
+
+/// Deserializes this URL from a `serde` stream.
+///
+/// This implementation is only available if the `serde` Cargo feature is enabled.
+#[cfg(feature="serde")]
+impl serde::Deserialize for Url {
+ fn deserialize<D>(deserializer: &mut D) -> Result<Url, D::Error> where D: serde::Deserializer {
+ let string_representation: String = try!(serde::Deserialize::deserialize(deserializer));
+ Ok(Url::parse(&string_representation).unwrap())
+ }
+}
+
+#[cfg(unix)]
+fn path_to_file_url_segments(path: &Path, serialization: &mut String) -> Result<(), ()> {
+ use std::os::unix::prelude::OsStrExt;
+ if !path.is_absolute() {
+ return Err(())
+ }
+ let mut empty = true;
+ // skip the root component
+ for component in path.components().skip(1) {
+ empty = false;
+ serialization.push('/');
+ serialization.extend(percent_encode(
+ component.as_os_str().as_bytes(), PATH_SEGMENT_ENCODE_SET));
+ }
+ if empty {
+ // An URL’s path must not be empty.
+ serialization.push('/');
+ }
+ Ok(())
+}
+
+#[cfg(windows)]
+fn path_to_file_url_segments(path: &Path, serialization: &mut String) -> Result<(), ()> {
+ path_to_file_url_segments_windows(path, serialization)
+}
+
+// Build this unconditionally to alleviate https://github.com/servo/rust-url/issues/102
+#[cfg_attr(not(windows), allow(dead_code))]
+fn path_to_file_url_segments_windows(path: &Path, serialization: &mut String) -> Result<(), ()> {
+ use std::path::{Prefix, Component};
+ if !path.is_absolute() {
+ return Err(())
+ }
+ let mut components = path.components();
+ let disk = match components.next() {
+ Some(Component::Prefix(ref p)) => match p.kind() {
+ Prefix::Disk(byte) => byte,
+ Prefix::VerbatimDisk(byte) => byte,
+ _ => return Err(()),
+ },
+
+ // FIXME: do something with UNC and other prefixes?
+ _ => return Err(())
+ };
+
+ // Start with the prefix, e.g. "C:"
+ serialization.push('/');
+ serialization.push(disk as char);
+ serialization.push(':');
+
+ for component in components {
+ if component == Component::RootDir { continue }
+ // FIXME: somehow work with non-unicode?
+ let component = try!(component.as_os_str().to_str().ok_or(()));
+ serialization.push('/');
+ serialization.extend(percent_encode(component.as_bytes(), PATH_SEGMENT_ENCODE_SET));
+ }
+ Ok(())
+}
+
+#[cfg(unix)]
+fn file_url_segments_to_pathbuf(segments: str::Split<char>) -> Result<PathBuf, ()> {
+ use std::ffi::OsStr;
+ use std::os::unix::prelude::OsStrExt;
+ use std::path::PathBuf;
+
+ let mut bytes = Vec::new();
+ for segment in segments {
+ bytes.push(b'/');
+ bytes.extend(percent_decode(segment.as_bytes()));
+ }
+ let os_str = OsStr::from_bytes(&bytes);
+ let path = PathBuf::from(os_str);
+ debug_assert!(path.is_absolute(),
+ "to_file_path() failed to produce an absolute Path");
+ Ok(path)
+}
+
+#[cfg(windows)]
+fn file_url_segments_to_pathbuf(segments: str::Split<char>) -> Result<PathBuf, ()> {
+ file_url_segments_to_pathbuf_windows(segments)
+}
+
+// Build this unconditionally to alleviate https://github.com/servo/rust-url/issues/102
+#[cfg_attr(not(windows), allow(dead_code))]
+fn file_url_segments_to_pathbuf_windows(mut segments: str::Split<char>) -> Result<PathBuf, ()> {
+ let first = try!(segments.next().ok_or(()));
+ if first.len() != 2 || !first.starts_with(parser::ascii_alpha)
+ || first.as_bytes()[1] != b':' {
+ return Err(())
+ }
+ let mut string = first.to_owned();
+ for segment in segments {
+ string.push('\\');
+
+ // Currently non-unicode windows paths cannot be represented
+ match String::from_utf8(percent_decode(segment.as_bytes()).collect()) {
+ Ok(s) => string.push_str(&s),
+ Err(..) => return Err(()),
+ }
+ }
+ let path = PathBuf::from(string);
+ debug_assert!(path.is_absolute(),
+ "to_file_path() failed to produce an absolute Path");
+ Ok(path)
+}
+
+fn io_error<T>(reason: &str) -> io::Result<T> {
+ Err(io::Error::new(io::ErrorKind::InvalidData, reason))
+}
+
+/// Implementation detail of `Url::query_pairs_mut`. Typically not used directly.
+pub struct UrlQuery<'a> {
+ url: &'a mut Url,
+ fragment: Option<String>,
+}
+
+impl<'a> Drop for UrlQuery<'a> {
+ fn drop(&mut self) {
+ self.url.restore_already_parsed_fragment(self.fragment.take())
+ }
+}
diff --git a/third_party/rust/url/src/origin.rs b/third_party/rust/url/src/origin.rs
new file mode 100644
index 000000000..2217c94fe
--- /dev/null
+++ b/third_party/rust/url/src/origin.rs
@@ -0,0 +1,114 @@
+// Copyright 2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+#[cfg(feature = "heapsize")] use heapsize::HeapSizeOf;
+use host::Host;
+use idna::domain_to_unicode;
+use parser::default_port;
+use std::sync::atomic::{AtomicUsize, ATOMIC_USIZE_INIT, Ordering};
+use Url;
+
+pub fn url_origin(url: &Url) -> Origin {
+ let scheme = url.scheme();
+ match scheme {
+ "blob" => {
+ let result = Url::parse(url.path());
+ match result {
+ Ok(ref url) => url_origin(url),
+ Err(_) => Origin::new_opaque()
+ }
+ },
+ "ftp" | "gopher" | "http" | "https" | "ws" | "wss" => {
+ Origin::Tuple(scheme.to_owned(), url.host().unwrap().to_owned(),
+ url.port_or_known_default().unwrap())
+ },
+ // TODO: Figure out what to do if the scheme is a file
+ "file" => Origin::new_opaque(),
+ _ => Origin::new_opaque()
+ }
+}
+
+/// The origin of an URL
+#[derive(PartialEq, Eq, Clone, Debug)]
+pub enum Origin {
+ /// A globally unique identifier
+ Opaque(OpaqueOrigin),
+
+ /// Consists of the URL's scheme, host and port
+ Tuple(String, Host<String>, u16)
+}
+
+#[cfg(feature = "heapsize")]
+impl HeapSizeOf for Origin {
+ fn heap_size_of_children(&self) -> usize {
+ match *self {
+ Origin::Tuple(ref scheme, ref host, _) => {
+ scheme.heap_size_of_children() +
+ host.heap_size_of_children()
+ },
+ _ => 0,
+ }
+ }
+}
+
+
+impl Origin {
+ /// Creates a new opaque origin that is only equal to itself.
+ pub fn new_opaque() -> Origin {
+ static COUNTER: AtomicUsize = ATOMIC_USIZE_INIT;
+ Origin::Opaque(OpaqueOrigin(COUNTER.fetch_add(1, Ordering::SeqCst)))
+ }
+
+ /// Return whether this origin is a (scheme, host, port) tuple
+ /// (as opposed to an opaque origin).
+ pub fn is_tuple(&self) -> bool {
+ matches!(*self, Origin::Tuple(..))
+ }
+
+ /// https://html.spec.whatwg.org/multipage/#ascii-serialisation-of-an-origin
+ pub fn ascii_serialization(&self) -> String {
+ match *self {
+ Origin::Opaque(_) => "null".to_owned(),
+ Origin::Tuple(ref scheme, ref host, port) => {
+ if default_port(scheme) == Some(port) {
+ format!("{}://{}", scheme, host)
+ } else {
+ format!("{}://{}:{}", scheme, host, port)
+ }
+ }
+ }
+ }
+
+ /// https://html.spec.whatwg.org/multipage/#unicode-serialisation-of-an-origin
+ pub fn unicode_serialization(&self) -> String {
+ match *self {
+ Origin::Opaque(_) => "null".to_owned(),
+ Origin::Tuple(ref scheme, ref host, port) => {
+ let host = match *host {
+ Host::Domain(ref domain) => {
+ let (domain, _errors) = domain_to_unicode(domain);
+ Host::Domain(domain)
+ }
+ _ => host.clone()
+ };
+ if default_port(scheme) == Some(port) {
+ format!("{}://{}", scheme, host)
+ } else {
+ format!("{}://{}:{}", scheme, host, port)
+ }
+ }
+ }
+ }
+}
+
+/// Opaque identifier for URLs that have file or other schemes
+#[derive(Eq, PartialEq, Clone, Debug)]
+pub struct OpaqueOrigin(usize);
+
+#[cfg(feature = "heapsize")]
+known_heap_size!(0, OpaqueOrigin);
diff --git a/third_party/rust/url/src/parser.rs b/third_party/rust/url/src/parser.rs
new file mode 100644
index 000000000..2945432fc
--- /dev/null
+++ b/third_party/rust/url/src/parser.rs
@@ -0,0 +1,1179 @@
+// Copyright 2013-2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+use std::ascii::AsciiExt;
+use std::error::Error;
+use std::fmt::{self, Formatter, Write};
+use std::str;
+
+use Url;
+use encoding::EncodingOverride;
+use host::{Host, HostInternal};
+use percent_encoding::{
+ utf8_percent_encode, percent_encode,
+ SIMPLE_ENCODE_SET, DEFAULT_ENCODE_SET, USERINFO_ENCODE_SET, QUERY_ENCODE_SET,
+ PATH_SEGMENT_ENCODE_SET
+};
+
+pub type ParseResult<T> = Result<T, ParseError>;
+
+macro_rules! simple_enum_error {
+ ($($name: ident => $description: expr,)+) => {
+ /// Errors that can occur during parsing.
+ #[derive(PartialEq, Eq, Clone, Copy, Debug)]
+ pub enum ParseError {
+ $(
+ $name,
+ )+
+ }
+
+ impl Error for ParseError {
+ fn description(&self) -> &str {
+ match *self {
+ $(
+ ParseError::$name => $description,
+ )+
+ }
+ }
+ }
+ }
+}
+
+simple_enum_error! {
+ EmptyHost => "empty host",
+ IdnaError => "invalid international domain name",
+ InvalidPort => "invalid port number",
+ InvalidIpv4Address => "invalid IPv4 address",
+ InvalidIpv6Address => "invalid IPv6 address",
+ InvalidDomainCharacter => "invalid domain character",
+ RelativeUrlWithoutBase => "relative URL without a base",
+ RelativeUrlWithCannotBeABaseBase => "relative URL with a cannot-be-a-base base",
+ SetHostOnCannotBeABaseUrl => "a cannot-be-a-base URL doesn’t have a host to set",
+ Overflow => "URLs more than 4 GB are not supported",
+}
+
+impl fmt::Display for ParseError {
+ fn fmt(&self, fmt: &mut Formatter) -> fmt::Result {
+ self.description().fmt(fmt)
+ }
+}
+
+impl From<::idna::uts46::Errors> for ParseError {
+ fn from(_: ::idna::uts46::Errors) -> ParseError { ParseError::IdnaError }
+}
+
+#[derive(Copy, Clone)]
+pub enum SchemeType {
+ File,
+ SpecialNotFile,
+ NotSpecial,
+}
+
+impl SchemeType {
+ pub fn is_special(&self) -> bool {
+ !matches!(*self, SchemeType::NotSpecial)
+ }
+
+ pub fn is_file(&self) -> bool {
+ matches!(*self, SchemeType::File)
+ }
+
+ pub fn from(s: &str) -> Self {
+ match s {
+ "http" | "https" | "ws" | "wss" | "ftp" | "gopher" => SchemeType::SpecialNotFile,
+ "file" => SchemeType::File,
+ _ => SchemeType::NotSpecial,
+ }
+ }
+}
+
+pub fn default_port(scheme: &str) -> Option<u16> {
+ match scheme {
+ "http" | "ws" => Some(80),
+ "https" | "wss" => Some(443),
+ "ftp" => Some(21),
+ "gopher" => Some(70),
+ _ => None,
+ }
+}
+
+#[derive(Clone)]
+pub struct Input<'i> {
+ chars: str::Chars<'i>,
+}
+
+impl<'i> Input<'i> {
+ pub fn new(input: &'i str) -> Self {
+ Input::with_log(input, None)
+ }
+
+ pub fn with_log(original_input: &'i str, log_syntax_violation: Option<&Fn(&'static str)>)
+ -> Self {
+ let input = original_input.trim_matches(c0_control_or_space);
+ if let Some(log) = log_syntax_violation {
+ if input.len() < original_input.len() {
+ log("leading or trailing control or space character are ignored in URLs")
+ }
+ if input.chars().any(|c| matches!(c, '\t' | '\n' | '\r')) {
+ log("tabs or newlines are ignored in URLs")
+ }
+ }
+ Input { chars: input.chars() }
+ }
+
+ #[inline]
+ pub fn is_empty(&self) -> bool {
+ self.clone().next().is_none()
+ }
+
+ #[inline]
+ fn starts_with<P: Pattern>(&self, p: P) -> bool {
+ p.split_prefix(&mut self.clone())
+ }
+
+ #[inline]
+ pub fn split_prefix<P: Pattern>(&self, p: P) -> Option<Self> {
+ let mut remaining = self.clone();
+ if p.split_prefix(&mut remaining) {
+ Some(remaining)
+ } else {
+ None
+ }
+ }
+
+ #[inline]
+ fn split_first(&self) -> (Option<char>, Self) {
+ let mut remaining = self.clone();
+ (remaining.next(), remaining)
+ }
+
+ #[inline]
+ fn count_matching<F: Fn(char) -> bool>(&self, f: F) -> (u32, Self) {
+ let mut count = 0;
+ let mut remaining = self.clone();
+ loop {
+ let mut input = remaining.clone();
+ if matches!(input.next(), Some(c) if f(c)) {
+ remaining = input;
+ count += 1;
+ } else {
+ return (count, remaining)
+ }
+ }
+ }
+
+ #[inline]
+ fn next_utf8(&mut self) -> Option<(char, &'i str)> {
+ loop {
+ let utf8 = self.chars.as_str();
+ match self.chars.next() {
+ Some(c) => {
+ if !matches!(c, '\t' | '\n' | '\r') {
+ return Some((c, &utf8[..c.len_utf8()]))
+ }
+ }
+ None => return None
+ }
+ }
+ }
+}
+
+pub trait Pattern {
+ fn split_prefix<'i>(self, input: &mut Input<'i>) -> bool;
+}
+
+impl Pattern for char {
+ fn split_prefix<'i>(self, input: &mut Input<'i>) -> bool { input.next() == Some(self) }
+}
+
+impl<'a> Pattern for &'a str {
+ fn split_prefix<'i>(self, input: &mut Input<'i>) -> bool {
+ for c in self.chars() {
+ if input.next() != Some(c) {
+ return false
+ }
+ }
+ true
+ }
+}
+
+impl<F: FnMut(char) -> bool> Pattern for F {
+ fn split_prefix<'i>(self, input: &mut Input<'i>) -> bool { input.next().map_or(false, self) }
+}
+
+impl<'i> Iterator for Input<'i> {
+ type Item = char;
+ fn next(&mut self) -> Option<char> {
+ self.chars.by_ref().filter(|&c| !matches!(c, '\t' | '\n' | '\r')).next()
+ }
+}
+
+pub struct Parser<'a> {
+ pub serialization: String,
+ pub base_url: Option<&'a Url>,
+ pub query_encoding_override: EncodingOverride,
+ pub log_syntax_violation: Option<&'a Fn(&'static str)>,
+ pub context: Context,
+}
+
+#[derive(PartialEq, Eq, Copy, Clone)]
+pub enum Context {
+ UrlParser,
+ Setter,
+ PathSegmentSetter,
+}
+
+impl<'a> Parser<'a> {
+ pub fn for_setter(serialization: String) -> Parser<'a> {
+ Parser {
+ serialization: serialization,
+ base_url: None,
+ query_encoding_override: EncodingOverride::utf8(),
+ log_syntax_violation: None,
+ context: Context::Setter,
+ }
+ }
+
+ fn syntax_violation(&self, reason: &'static str) {
+ if let Some(log) = self.log_syntax_violation {
+ log(reason)
+ }
+ }
+
+ fn syntax_violation_if<F: Fn() -> bool>(&self, reason: &'static str, test: F) {
+ // Skip test if not logging.
+ if let Some(log) = self.log_syntax_violation {
+ if test() {
+ log(reason)
+ }
+ }
+ }
+
+ /// https://url.spec.whatwg.org/#concept-basic-url-parser
+ pub fn parse_url(mut self, input: &str) -> ParseResult<Url> {
+ let input = Input::with_log(input, self.log_syntax_violation);
+ if let Ok(remaining) = self.parse_scheme(input.clone()) {
+ return self.parse_with_scheme(remaining)
+ }
+
+ // No-scheme state
+ if let Some(base_url) = self.base_url {
+ if input.starts_with('#') {
+ self.fragment_only(base_url, input)
+ } else if base_url.cannot_be_a_base() {
+ Err(ParseError::RelativeUrlWithCannotBeABaseBase)
+ } else {
+ let scheme_type = SchemeType::from(base_url.scheme());
+ if scheme_type.is_file() {
+ self.parse_file(input, Some(base_url))
+ } else {
+ self.parse_relative(input, scheme_type, base_url)
+ }
+ }
+ } else {
+ Err(ParseError::RelativeUrlWithoutBase)
+ }
+ }
+
+ pub fn parse_scheme<'i>(&mut self, mut input: Input<'i>) -> Result<Input<'i>, ()> {
+ if input.is_empty() || !input.starts_with(ascii_alpha) {
+ return Err(())
+ }
+ debug_assert!(self.serialization.is_empty());
+ while let Some(c) = input.next() {
+ match c {
+ 'a'...'z' | 'A'...'Z' | '0'...'9' | '+' | '-' | '.' => {
+ self.serialization.push(c.to_ascii_lowercase())
+ }
+ ':' => return Ok(input),
+ _ => {
+ self.serialization.clear();
+ return Err(())
+ }
+ }
+ }
+ // EOF before ':'
+ if self.context == Context::Setter {
+ Ok(input)
+ } else {
+ self.serialization.clear();
+ Err(())
+ }
+ }
+
+ fn parse_with_scheme(mut self, input: Input) -> ParseResult<Url> {
+ let scheme_end = try!(to_u32(self.serialization.len()));
+ let scheme_type = SchemeType::from(&self.serialization);
+ self.serialization.push(':');
+ match scheme_type {
+ SchemeType::File => {
+ self.syntax_violation_if("expected // after file:", || !input.starts_with("//"));
+ let base_file_url = self.base_url.and_then(|base| {
+ if base.scheme() == "file" { Some(base) } else { None }
+ });
+ self.serialization.clear();
+ self.parse_file(input, base_file_url)
+ }
+ SchemeType::SpecialNotFile => {
+ // special relative or authority state
+ let (slashes_count, remaining) = input.count_matching(|c| matches!(c, '/' | '\\'));
+ if let Some(base_url) = self.base_url {
+ if slashes_count < 2 &&
+ base_url.scheme() == &self.serialization[..scheme_end as usize] {
+ // "Cannot-be-a-base" URLs only happen with "not special" schemes.
+ debug_assert!(!base_url.cannot_be_a_base());
+ self.serialization.clear();
+ return self.parse_relative(input, scheme_type, base_url)
+ }
+ }
+ // special authority slashes state
+ self.syntax_violation_if("expected //", || {
+ input.clone().take_while(|&c| matches!(c, '/' | '\\'))
+ .collect::<String>() != "//"
+ });
+ self.after_double_slash(remaining, scheme_type, scheme_end)
+ }
+ SchemeType::NotSpecial => self.parse_non_special(input, scheme_type, scheme_end)
+ }
+ }
+
+ /// Scheme other than file, http, https, ws, ws, ftp, gopher.
+ fn parse_non_special(mut self, input: Input, scheme_type: SchemeType, scheme_end: u32)
+ -> ParseResult<Url> {
+ // path or authority state (
+ if let Some(input) = input.split_prefix("//") {
+ return self.after_double_slash(input, scheme_type, scheme_end)
+ }
+ // Anarchist URL (no authority)
+ let path_start = try!(to_u32(self.serialization.len()));
+ let username_end = path_start;
+ let host_start = path_start;
+ let host_end = path_start;
+ let host = HostInternal::None;
+ let port = None;
+ let remaining = if let Some(input) = input.split_prefix('/') {
+ let path_start = self.serialization.len();
+ self.serialization.push('/');
+ self.parse_path(scheme_type, &mut false, path_start, input)
+ } else {
+ self.parse_cannot_be_a_base_path(input)
+ };
+ self.with_query_and_fragment(scheme_end, username_end, host_start,
+ host_end, host, port, path_start, remaining)
+ }
+
+ fn parse_file(mut self, input: Input, mut base_file_url: Option<&Url>) -> ParseResult<Url> {
+ // file state
+ debug_assert!(self.serialization.is_empty());
+ let (first_char, input_after_first_char) = input.split_first();
+ match first_char {
+ None => {
+ if let Some(base_url) = base_file_url {
+ // Copy everything except the fragment
+ let before_fragment = match base_url.fragment_start {
+ Some(i) => &base_url.serialization[..i as usize],
+ None => &*base_url.serialization,
+ };
+ self.serialization.push_str(before_fragment);
+ Ok(Url {
+ serialization: self.serialization,
+ fragment_start: None,
+ ..*base_url
+ })
+ } else {
+ self.serialization.push_str("file:///");
+ let scheme_end = "file".len() as u32;
+ let path_start = "file://".len() as u32;
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: None,
+ fragment_start: None,
+ })
+ }
+ },
+ Some('?') => {
+ if let Some(base_url) = base_file_url {
+ // Copy everything up to the query string
+ let before_query = match (base_url.query_start, base_url.fragment_start) {
+ (None, None) => &*base_url.serialization,
+ (Some(i), _) |
+ (None, Some(i)) => base_url.slice(..i)
+ };
+ self.serialization.push_str(before_query);
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(base_url.scheme_end, input));
+ Ok(Url {
+ serialization: self.serialization,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ ..*base_url
+ })
+ } else {
+ self.serialization.push_str("file:///");
+ let scheme_end = "file".len() as u32;
+ let path_start = "file://".len() as u32;
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(scheme_end, input));
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ })
+ }
+ },
+ Some('#') => {
+ if let Some(base_url) = base_file_url {
+ self.fragment_only(base_url, input)
+ } else {
+ self.serialization.push_str("file:///");
+ let scheme_end = "file".len() as u32;
+ let path_start = "file://".len() as u32;
+ let fragment_start = "file:///".len() as u32;
+ self.parse_fragment(input_after_first_char);
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: None,
+ fragment_start: Some(fragment_start),
+ })
+ }
+ }
+ Some('/') | Some('\\') => {
+ self.syntax_violation_if("backslash", || first_char == Some('\\'));
+ // file slash state
+ let (next_char, input_after_next_char) = input_after_first_char.split_first();
+ self.syntax_violation_if("backslash", || next_char == Some('\\'));
+ if matches!(next_char, Some('/') | Some('\\')) {
+ // file host state
+ self.serialization.push_str("file://");
+ let scheme_end = "file".len() as u32;
+ let host_start = "file://".len() as u32;
+ let (path_start, host, remaining) =
+ try!(self.parse_file_host(input_after_next_char));
+ let host_end = try!(to_u32(self.serialization.len()));
+ let mut has_host = !matches!(host, HostInternal::None);
+ let remaining = if path_start {
+ self.parse_path_start(SchemeType::File, &mut has_host, remaining)
+ } else {
+ let path_start = self.serialization.len();
+ self.serialization.push('/');
+ self.parse_path(SchemeType::File, &mut has_host, path_start, remaining)
+ };
+ // FIXME: deal with has_host
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(scheme_end, remaining));
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: host_start,
+ host_start: host_start,
+ host_end: host_end,
+ host: host,
+ port: None,
+ path_start: host_end,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ })
+ } else {
+ self.serialization.push_str("file:///");
+ let scheme_end = "file".len() as u32;
+ let path_start = "file://".len();
+ if let Some(base_url) = base_file_url {
+ let first_segment = base_url.path_segments().unwrap().next().unwrap();
+ // FIXME: *normalized* drive letter
+ if is_windows_drive_letter(first_segment) {
+ self.serialization.push_str(first_segment);
+ self.serialization.push('/');
+ }
+ }
+ let remaining = self.parse_path(
+ SchemeType::File, &mut false, path_start, input_after_first_char);
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(scheme_end, remaining));
+ let path_start = path_start as u32;
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ })
+ }
+ }
+ _ => {
+ if starts_with_windows_drive_letter_segment(&input) {
+ base_file_url = None;
+ }
+ if let Some(base_url) = base_file_url {
+ let before_query = match (base_url.query_start, base_url.fragment_start) {
+ (None, None) => &*base_url.serialization,
+ (Some(i), _) |
+ (None, Some(i)) => base_url.slice(..i)
+ };
+ self.serialization.push_str(before_query);
+ self.pop_path(SchemeType::File, base_url.path_start as usize);
+ let remaining = self.parse_path(
+ SchemeType::File, &mut true, base_url.path_start as usize, input);
+ self.with_query_and_fragment(
+ base_url.scheme_end, base_url.username_end, base_url.host_start,
+ base_url.host_end, base_url.host, base_url.port, base_url.path_start, remaining)
+ } else {
+ self.serialization.push_str("file:///");
+ let scheme_end = "file".len() as u32;
+ let path_start = "file://".len();
+ let remaining = self.parse_path(
+ SchemeType::File, &mut false, path_start, input);
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(scheme_end, remaining));
+ let path_start = path_start as u32;
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: path_start,
+ host_start: path_start,
+ host_end: path_start,
+ host: HostInternal::None,
+ port: None,
+ path_start: path_start,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ })
+ }
+ }
+ }
+ }
+
+ fn parse_relative(mut self, input: Input, scheme_type: SchemeType, base_url: &Url)
+ -> ParseResult<Url> {
+ // relative state
+ debug_assert!(self.serialization.is_empty());
+ let (first_char, input_after_first_char) = input.split_first();
+ match first_char {
+ None => {
+ // Copy everything except the fragment
+ let before_fragment = match base_url.fragment_start {
+ Some(i) => &base_url.serialization[..i as usize],
+ None => &*base_url.serialization,
+ };
+ self.serialization.push_str(before_fragment);
+ Ok(Url {
+ serialization: self.serialization,
+ fragment_start: None,
+ ..*base_url
+ })
+ },
+ Some('?') => {
+ // Copy everything up to the query string
+ let before_query = match (base_url.query_start, base_url.fragment_start) {
+ (None, None) => &*base_url.serialization,
+ (Some(i), _) |
+ (None, Some(i)) => base_url.slice(..i)
+ };
+ self.serialization.push_str(before_query);
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(base_url.scheme_end, input));
+ Ok(Url {
+ serialization: self.serialization,
+ query_start: query_start,
+ fragment_start: fragment_start,
+ ..*base_url
+ })
+ },
+ Some('#') => self.fragment_only(base_url, input),
+ Some('/') | Some('\\') => {
+ let (slashes_count, remaining) = input.count_matching(|c| matches!(c, '/' | '\\'));
+ if slashes_count >= 2 {
+ self.syntax_violation_if("expected //", || {
+ input.clone().take_while(|&c| matches!(c, '/' | '\\'))
+ .collect::<String>() != "//"
+ });
+ let scheme_end = base_url.scheme_end;
+ debug_assert!(base_url.byte_at(scheme_end) == b':');
+ self.serialization.push_str(base_url.slice(..scheme_end + 1));
+ return self.after_double_slash(remaining, scheme_type, scheme_end)
+ }
+ let path_start = base_url.path_start;
+ debug_assert!(base_url.byte_at(path_start) == b'/');
+ self.serialization.push_str(base_url.slice(..path_start + 1));
+ let remaining = self.parse_path(
+ scheme_type, &mut true, path_start as usize, input_after_first_char);
+ self.with_query_and_fragment(
+ base_url.scheme_end, base_url.username_end, base_url.host_start,
+ base_url.host_end, base_url.host, base_url.port, base_url.path_start, remaining)
+ }
+ _ => {
+ let before_query = match (base_url.query_start, base_url.fragment_start) {
+ (None, None) => &*base_url.serialization,
+ (Some(i), _) |
+ (None, Some(i)) => base_url.slice(..i)
+ };
+ self.serialization.push_str(before_query);
+ // FIXME spec says just "remove last entry", not the "pop" algorithm
+ self.pop_path(scheme_type, base_url.path_start as usize);
+ let remaining = self.parse_path(
+ scheme_type, &mut true, base_url.path_start as usize, input);
+ self.with_query_and_fragment(
+ base_url.scheme_end, base_url.username_end, base_url.host_start,
+ base_url.host_end, base_url.host, base_url.port, base_url.path_start, remaining)
+ }
+ }
+ }
+
+ fn after_double_slash(mut self, input: Input, scheme_type: SchemeType, scheme_end: u32)
+ -> ParseResult<Url> {
+ self.serialization.push('/');
+ self.serialization.push('/');
+ // authority state
+ let (username_end, remaining) = try!(self.parse_userinfo(input, scheme_type));
+ // host state
+ let host_start = try!(to_u32(self.serialization.len()));
+ let (host_end, host, port, remaining) =
+ try!(self.parse_host_and_port(remaining, scheme_end, scheme_type));
+ // path state
+ let path_start = try!(to_u32(self.serialization.len()));
+ let remaining = self.parse_path_start(
+ scheme_type, &mut true, remaining);
+ self.with_query_and_fragment(scheme_end, username_end, host_start,
+ host_end, host, port, path_start, remaining)
+ }
+
+ /// Return (username_end, remaining)
+ fn parse_userinfo<'i>(&mut self, mut input: Input<'i>, scheme_type: SchemeType)
+ -> ParseResult<(u32, Input<'i>)> {
+ let mut last_at = None;
+ let mut remaining = input.clone();
+ let mut char_count = 0;
+ while let Some(c) = remaining.next() {
+ match c {
+ '@' => {
+ if last_at.is_some() {
+ self.syntax_violation("unencoded @ sign in username or password")
+ } else {
+ self.syntax_violation(
+ "embedding authentification information (username or password) \
+ in an URL is not recommended")
+ }
+ last_at = Some((char_count, remaining.clone()))
+ },
+ '/' | '?' | '#' => break,
+ '\\' if scheme_type.is_special() => break,
+ _ => (),
+ }
+ char_count += 1;
+ }
+ let (mut userinfo_char_count, remaining) = match last_at {
+ None => return Ok((try!(to_u32(self.serialization.len())), input)),
+ Some((0, remaining)) => return Ok((try!(to_u32(self.serialization.len())), remaining)),
+ Some(x) => x
+ };
+
+ let mut username_end = None;
+ while userinfo_char_count > 0 {
+ let (c, utf8_c) = input.next_utf8().unwrap();
+ userinfo_char_count -= 1;
+ if c == ':' && username_end.is_none() {
+ // Start parsing password
+ username_end = Some(try!(to_u32(self.serialization.len())));
+ self.serialization.push(':');
+ } else {
+ self.check_url_code_point(c, &input);
+ self.serialization.extend(utf8_percent_encode(utf8_c, USERINFO_ENCODE_SET));
+ }
+ }
+ let username_end = match username_end {
+ Some(i) => i,
+ None => try!(to_u32(self.serialization.len())),
+ };
+ self.serialization.push('@');
+ Ok((username_end, remaining))
+ }
+
+ fn parse_host_and_port<'i>(&mut self, input: Input<'i>,
+ scheme_end: u32, scheme_type: SchemeType)
+ -> ParseResult<(u32, HostInternal, Option<u16>, Input<'i>)> {
+ let (host, remaining) = try!(
+ Parser::parse_host(input, scheme_type));
+ write!(&mut self.serialization, "{}", host).unwrap();
+ let host_end = try!(to_u32(self.serialization.len()));
+ let (port, remaining) = if let Some(remaining) = remaining.split_prefix(':') {
+ let scheme = || default_port(&self.serialization[..scheme_end as usize]);
+ try!(Parser::parse_port(remaining, scheme, self.context))
+ } else {
+ (None, remaining)
+ };
+ if let Some(port) = port {
+ write!(&mut self.serialization, ":{}", port).unwrap()
+ }
+ Ok((host_end, host.into(), port, remaining))
+ }
+
+ pub fn parse_host<'i>(mut input: Input<'i>, scheme_type: SchemeType)
+ -> ParseResult<(Host<String>, Input<'i>)> {
+ // Undo the Input abstraction here to avoid allocating in the common case
+ // where the host part of the input does not contain any tab or newline
+ let input_str = input.chars.as_str();
+ let mut inside_square_brackets = false;
+ let mut has_ignored_chars = false;
+ let mut non_ignored_chars = 0;
+ let mut bytes = 0;
+ for c in input_str.chars() {
+ match c {
+ ':' if !inside_square_brackets => break,
+ '\\' if scheme_type.is_special() => break,
+ '/' | '?' | '#' => break,
+ '\t' | '\n' | '\r' => {
+ has_ignored_chars = true;
+ }
+ '[' => {
+ inside_square_brackets = true;
+ non_ignored_chars += 1
+ }
+ ']' => {
+ inside_square_brackets = false;
+ non_ignored_chars += 1
+ }
+ _ => non_ignored_chars += 1
+ }
+ bytes += c.len_utf8();
+ }
+ let replaced: String;
+ let host_str;
+ {
+ let host_input = input.by_ref().take(non_ignored_chars);
+ if has_ignored_chars {
+ replaced = host_input.collect();
+ host_str = &*replaced
+ } else {
+ for _ in host_input {}
+ host_str = &input_str[..bytes]
+ }
+ }
+ if scheme_type.is_special() && host_str.is_empty() {
+ return Err(ParseError::EmptyHost)
+ }
+ let host = try!(Host::parse(host_str));
+ Ok((host, input))
+ }
+
+ pub fn parse_file_host<'i>(&mut self, input: Input<'i>)
+ -> ParseResult<(bool, HostInternal, Input<'i>)> {
+ // Undo the Input abstraction here to avoid allocating in the common case
+ // where the host part of the input does not contain any tab or newline
+ let input_str = input.chars.as_str();
+ let mut has_ignored_chars = false;
+ let mut non_ignored_chars = 0;
+ let mut bytes = 0;
+ for c in input_str.chars() {
+ match c {
+ '/' | '\\' | '?' | '#' => break,
+ '\t' | '\n' | '\r' => has_ignored_chars = true,
+ _ => non_ignored_chars += 1,
+ }
+ bytes += c.len_utf8();
+ }
+ let replaced: String;
+ let host_str;
+ let mut remaining = input.clone();
+ {
+ let host_input = remaining.by_ref().take(non_ignored_chars);
+ if has_ignored_chars {
+ replaced = host_input.collect();
+ host_str = &*replaced
+ } else {
+ for _ in host_input {}
+ host_str = &input_str[..bytes]
+ }
+ }
+ if is_windows_drive_letter(host_str) {
+ return Ok((false, HostInternal::None, input))
+ }
+ let host = if host_str.is_empty() {
+ HostInternal::None
+ } else {
+ match try!(Host::parse(host_str)) {
+ Host::Domain(ref d) if d == "localhost" => HostInternal::None,
+ host => {
+ write!(&mut self.serialization, "{}", host).unwrap();
+ host.into()
+ }
+ }
+ };
+ Ok((true, host, remaining))
+ }
+
+ pub fn parse_port<'i, P>(mut input: Input<'i>, default_port: P,
+ context: Context)
+ -> ParseResult<(Option<u16>, Input<'i>)>
+ where P: Fn() -> Option<u16> {
+ let mut port: u32 = 0;
+ let mut has_any_digit = false;
+ while let (Some(c), remaining) = input.split_first() {
+ if let Some(digit) = c.to_digit(10) {
+ port = port * 10 + digit;
+ if port > ::std::u16::MAX as u32 {
+ return Err(ParseError::InvalidPort)
+ }
+ has_any_digit = true;
+ } else if context == Context::UrlParser && !matches!(c, '/' | '\\' | '?' | '#') {
+ return Err(ParseError::InvalidPort)
+ } else {
+ break
+ }
+ input = remaining;
+ }
+ let mut opt_port = Some(port as u16);
+ if !has_any_digit || opt_port == default_port() {
+ opt_port = None;
+ }
+ return Ok((opt_port, input))
+ }
+
+ pub fn parse_path_start<'i>(&mut self, scheme_type: SchemeType, has_host: &mut bool,
+ mut input: Input<'i>)
+ -> Input<'i> {
+ // Path start state
+ match input.split_first() {
+ (Some('/'), remaining) => input = remaining,
+ (Some('\\'), remaining) => if scheme_type.is_special() {
+ self.syntax_violation("backslash");
+ input = remaining
+ },
+ _ => {}
+ }
+ let path_start = self.serialization.len();
+ self.serialization.push('/');
+ self.parse_path(scheme_type, has_host, path_start, input)
+ }
+
+ pub fn parse_path<'i>(&mut self, scheme_type: SchemeType, has_host: &mut bool,
+ path_start: usize, mut input: Input<'i>)
+ -> Input<'i> {
+ // Relative path state
+ debug_assert!(self.serialization.ends_with("/"));
+ loop {
+ let segment_start = self.serialization.len();
+ let mut ends_with_slash = false;
+ loop {
+ let input_before_c = input.clone();
+ let (c, utf8_c) = if let Some(x) = input.next_utf8() { x } else { break };
+ match c {
+ '/' if self.context != Context::PathSegmentSetter => {
+ ends_with_slash = true;
+ break
+ },
+ '\\' if self.context != Context::PathSegmentSetter &&
+ scheme_type.is_special() => {
+ self.syntax_violation("backslash");
+ ends_with_slash = true;
+ break
+ },
+ '?' | '#' if self.context == Context::UrlParser => {
+ input = input_before_c;
+ break
+ },
+ _ => {
+ self.check_url_code_point(c, &input);
+ if c == '%' {
+ let after_percent_sign = input.clone();
+ if matches!(input.next(), Some('2')) &&
+ matches!(input.next(), Some('E') | Some('e')) {
+ self.serialization.push('.');
+ continue
+ }
+ input = after_percent_sign
+ }
+ if self.context == Context::PathSegmentSetter {
+ self.serialization.extend(utf8_percent_encode(
+ utf8_c, PATH_SEGMENT_ENCODE_SET));
+ } else {
+ self.serialization.extend(utf8_percent_encode(
+ utf8_c, DEFAULT_ENCODE_SET));
+ }
+ }
+ }
+ }
+ match &self.serialization[segment_start..] {
+ ".." => {
+ debug_assert!(self.serialization.as_bytes()[segment_start - 1] == b'/');
+ self.serialization.truncate(segment_start - 1); // Truncate "/.."
+ self.pop_path(scheme_type, path_start);
+ if !self.serialization[path_start..].ends_with("/") {
+ self.serialization.push('/')
+ }
+ },
+ "." => {
+ self.serialization.truncate(segment_start);
+ },
+ _ => {
+ if scheme_type.is_file() && is_windows_drive_letter(
+ &self.serialization[path_start + 1..]
+ ) {
+ if self.serialization.ends_with('|') {
+ self.serialization.pop();
+ self.serialization.push(':');
+ }
+ if *has_host {
+ self.syntax_violation("file: with host and Windows drive letter");
+ *has_host = false; // FIXME account for this in callers
+ }
+ }
+ if ends_with_slash {
+ self.serialization.push('/')
+ }
+ }
+ }
+ if !ends_with_slash {
+ break
+ }
+ }
+ input
+ }
+
+ /// https://url.spec.whatwg.org/#pop-a-urls-path
+ fn pop_path(&mut self, scheme_type: SchemeType, path_start: usize) {
+ if self.serialization.len() > path_start {
+ let slash_position = self.serialization[path_start..].rfind('/').unwrap();
+ // + 1 since rfind returns the position before the slash.
+ let segment_start = path_start + slash_position + 1;
+ // Don’t pop a Windows drive letter
+ // FIXME: *normalized* Windows drive letter
+ if !(
+ scheme_type.is_file() &&
+ is_windows_drive_letter(&self.serialization[segment_start..])
+ ) {
+ self.serialization.truncate(segment_start);
+ }
+ }
+
+ }
+
+ pub fn parse_cannot_be_a_base_path<'i>(&mut self, mut input: Input<'i>) -> Input<'i> {
+ loop {
+ let input_before_c = input.clone();
+ match input.next_utf8() {
+ Some(('?', _)) | Some(('#', _)) if self.context == Context::UrlParser => {
+ return input_before_c
+ }
+ Some((c, utf8_c)) => {
+ self.check_url_code_point(c, &input);
+ self.serialization.extend(utf8_percent_encode(
+ utf8_c, SIMPLE_ENCODE_SET));
+ }
+ None => return input
+ }
+ }
+ }
+
+ fn with_query_and_fragment(mut self, scheme_end: u32, username_end: u32,
+ host_start: u32, host_end: u32, host: HostInternal,
+ port: Option<u16>, path_start: u32, remaining: Input)
+ -> ParseResult<Url> {
+ let (query_start, fragment_start) =
+ try!(self.parse_query_and_fragment(scheme_end, remaining));
+ Ok(Url {
+ serialization: self.serialization,
+ scheme_end: scheme_end,
+ username_end: username_end,
+ host_start: host_start,
+ host_end: host_end,
+ host: host,
+ port: port,
+ path_start: path_start,
+ query_start: query_start,
+ fragment_start: fragment_start
+ })
+ }
+
+ /// Return (query_start, fragment_start)
+ fn parse_query_and_fragment(&mut self, scheme_end: u32, mut input: Input)
+ -> ParseResult<(Option<u32>, Option<u32>)> {
+ let mut query_start = None;
+ match input.next() {
+ Some('#') => {}
+ Some('?') => {
+ query_start = Some(try!(to_u32(self.serialization.len())));
+ self.serialization.push('?');
+ let remaining = self.parse_query(scheme_end, input);
+ if let Some(remaining) = remaining {
+ input = remaining
+ } else {
+ return Ok((query_start, None))
+ }
+ }
+ None => return Ok((None, None)),
+ _ => panic!("Programming error. parse_query_and_fragment() called without ? or # {:?}")
+ }
+
+ let fragment_start = try!(to_u32(self.serialization.len()));
+ self.serialization.push('#');
+ self.parse_fragment(input);
+ Ok((query_start, Some(fragment_start)))
+ }
+
+ pub fn parse_query<'i>(&mut self, scheme_end: u32, mut input: Input<'i>)
+ -> Option<Input<'i>> {
+ let mut query = String::new(); // FIXME: use a streaming decoder instead
+ let mut remaining = None;
+ while let Some(c) = input.next() {
+ if c == '#' && self.context == Context::UrlParser {
+ remaining = Some(input);
+ break
+ } else {
+ self.check_url_code_point(c, &input);
+ query.push(c);
+ }
+ }
+
+ let encoding = match &self.serialization[..scheme_end as usize] {
+ "http" | "https" | "file" | "ftp" | "gopher" => self.query_encoding_override,
+ _ => EncodingOverride::utf8(),
+ };
+ let query_bytes = encoding.encode(query.into());
+ self.serialization.extend(percent_encode(&query_bytes, QUERY_ENCODE_SET));
+ remaining
+ }
+
+ fn fragment_only(mut self, base_url: &Url, mut input: Input) -> ParseResult<Url> {
+ let before_fragment = match base_url.fragment_start {
+ Some(i) => base_url.slice(..i),
+ None => &*base_url.serialization,
+ };
+ debug_assert!(self.serialization.is_empty());
+ self.serialization.reserve(before_fragment.len() + input.chars.as_str().len());
+ self.serialization.push_str(before_fragment);
+ self.serialization.push('#');
+ let next = input.next();
+ debug_assert!(next == Some('#'));
+ self.parse_fragment(input);
+ Ok(Url {
+ serialization: self.serialization,
+ fragment_start: Some(try!(to_u32(before_fragment.len()))),
+ ..*base_url
+ })
+ }
+
+ pub fn parse_fragment(&mut self, mut input: Input) {
+ while let Some(c) = input.next() {
+ if c == '\0' {
+ self.syntax_violation("NULL characters are ignored in URL fragment identifiers")
+ } else {
+ self.check_url_code_point(c, &input);
+ self.serialization.push(c); // No percent-encoding here.
+ }
+ }
+ }
+
+ fn check_url_code_point(&self, c: char, input: &Input) {
+ if let Some(log) = self.log_syntax_violation {
+ if c == '%' {
+ let mut input = input.clone();
+ if !matches!((input.next(), input.next()), (Some(a), Some(b))
+ if is_ascii_hex_digit(a) && is_ascii_hex_digit(b)) {
+ log("expected 2 hex digits after %")
+ }
+ } else if !is_url_code_point(c) {
+ log("non-URL code point")
+ }
+ }
+ }
+}
+
+#[inline]
+fn is_ascii_hex_digit(c: char) -> bool {
+ matches!(c, 'a'...'f' | 'A'...'F' | '0'...'9')
+}
+
+// Non URL code points:
+// U+0000 to U+0020 (space)
+// " # % < > [ \ ] ^ ` { | }
+// U+007F to U+009F
+// surrogates
+// U+FDD0 to U+FDEF
+// Last two of each plane: U+__FFFE to U+__FFFF for __ in 00 to 10 hex
+#[inline]
+fn is_url_code_point(c: char) -> bool {
+ matches!(c,
+ 'a'...'z' |
+ 'A'...'Z' |
+ '0'...'9' |
+ '!' | '$' | '&' | '\'' | '(' | ')' | '*' | '+' | ',' | '-' |
+ '.' | '/' | ':' | ';' | '=' | '?' | '@' | '_' | '~' |
+ '\u{A0}'...'\u{D7FF}' | '\u{E000}'...'\u{FDCF}' | '\u{FDF0}'...'\u{FFFD}' |
+ '\u{10000}'...'\u{1FFFD}' | '\u{20000}'...'\u{2FFFD}' |
+ '\u{30000}'...'\u{3FFFD}' | '\u{40000}'...'\u{4FFFD}' |
+ '\u{50000}'...'\u{5FFFD}' | '\u{60000}'...'\u{6FFFD}' |
+ '\u{70000}'...'\u{7FFFD}' | '\u{80000}'...'\u{8FFFD}' |
+ '\u{90000}'...'\u{9FFFD}' | '\u{A0000}'...'\u{AFFFD}' |
+ '\u{B0000}'...'\u{BFFFD}' | '\u{C0000}'...'\u{CFFFD}' |
+ '\u{D0000}'...'\u{DFFFD}' | '\u{E1000}'...'\u{EFFFD}' |
+ '\u{F0000}'...'\u{FFFFD}' | '\u{100000}'...'\u{10FFFD}')
+}
+
+/// https://url.spec.whatwg.org/#c0-controls-and-space
+#[inline]
+fn c0_control_or_space(ch: char) -> bool {
+ ch <= ' ' // U+0000 to U+0020
+}
+
+/// https://url.spec.whatwg.org/#ascii-alpha
+#[inline]
+pub fn ascii_alpha(ch: char) -> bool {
+ matches!(ch, 'a'...'z' | 'A'...'Z')
+}
+
+#[inline]
+pub fn to_u32(i: usize) -> ParseResult<u32> {
+ if i <= ::std::u32::MAX as usize {
+ Ok(i as u32)
+ } else {
+ Err(ParseError::Overflow)
+ }
+}
+
+/// Wether the scheme is file:, the path has a single segment, and that segment
+/// is a Windows drive letter
+fn is_windows_drive_letter(segment: &str) -> bool {
+ segment.len() == 2
+ && starts_with_windows_drive_letter(segment)
+}
+
+fn starts_with_windows_drive_letter(s: &str) -> bool {
+ ascii_alpha(s.as_bytes()[0] as char)
+ && matches!(s.as_bytes()[1], b':' | b'|')
+}
+
+fn starts_with_windows_drive_letter_segment(input: &Input) -> bool {
+ let mut input = input.clone();
+ matches!((input.next(), input.next(), input.next()), (Some(a), Some(b), Some(c))
+ if ascii_alpha(a) && matches!(b, ':' | '|') && matches!(c, '/' | '\\' | '?' | '#'))
+}
diff --git a/third_party/rust/url/src/path_segments.rs b/third_party/rust/url/src/path_segments.rs
new file mode 100644
index 000000000..437a84ee7
--- /dev/null
+++ b/third_party/rust/url/src/path_segments.rs
@@ -0,0 +1,187 @@
+// Copyright 2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+use parser::{self, SchemeType, to_u32};
+use std::str;
+use Url;
+
+/// Exposes methods to manipulate the path of an URL that is not cannot-be-base.
+///
+/// The path always starts with a `/` slash, and is made of slash-separated segments.
+/// There is always at least one segment (which may be the empty string).
+///
+/// Examples:
+///
+/// ```rust
+/// # use url::Url;
+/// let mut url = Url::parse("mailto:me@example.com").unwrap();
+/// assert!(url.path_segments_mut().is_err());
+///
+/// let mut url = Url::parse("http://example.net/foo/index.html").unwrap();
+/// url.path_segments_mut().unwrap().pop().push("img").push("2/100%.png");
+/// assert_eq!(url.as_str(), "http://example.net/foo/img/2%2F100%25.png");
+/// ```
+pub struct PathSegmentsMut<'a> {
+ url: &'a mut Url,
+ after_first_slash: usize,
+ after_path: String,
+ old_after_path_position: u32,
+}
+
+// Not re-exported outside the crate
+pub fn new(url: &mut Url) -> PathSegmentsMut {
+ let after_path = url.take_after_path();
+ let old_after_path_position = to_u32(url.serialization.len()).unwrap();
+ debug_assert!(url.byte_at(url.path_start) == b'/');
+ PathSegmentsMut {
+ after_first_slash: url.path_start as usize + "/".len(),
+ url: url,
+ old_after_path_position: old_after_path_position,
+ after_path: after_path,
+ }
+}
+
+impl<'a> Drop for PathSegmentsMut<'a> {
+ fn drop(&mut self) {
+ self.url.restore_after_path(self.old_after_path_position, &self.after_path)
+ }
+}
+
+impl<'a> PathSegmentsMut<'a> {
+ /// Remove all segments in the path, leaving the minimal `url.path() == "/"`.
+ ///
+ /// Returns `&mut Self` so that method calls can be chained.
+ ///
+ /// Example:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// let mut url = Url::parse("https://github.com/servo/rust-url/").unwrap();
+ /// url.path_segments_mut().unwrap().clear().push("logout");
+ /// assert_eq!(url.as_str(), "https://github.com/logout");
+ /// ```
+ pub fn clear(&mut self) -> &mut Self {
+ self.url.serialization.truncate(self.after_first_slash);
+ self
+ }
+
+ /// Remove the last segment of this URL’s path if it is empty,
+ /// except if these was only one segment to begin with.
+ ///
+ /// In other words, remove one path trailing slash, if any,
+ /// unless it is also the initial slash (so this does nothing if `url.path() == "/")`.
+ ///
+ /// Returns `&mut Self` so that method calls can be chained.
+ ///
+ /// Example:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// let mut url = Url::parse("https://github.com/servo/rust-url/").unwrap();
+ /// url.path_segments_mut().unwrap().push("pulls");
+ /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url//pulls");
+ ///
+ /// let mut url = Url::parse("https://github.com/servo/rust-url/").unwrap();
+ /// url.path_segments_mut().unwrap().pop_if_empty().push("pulls");
+ /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/pulls");
+ /// ```
+ pub fn pop_if_empty(&mut self) -> &mut Self {
+ if self.url.serialization[self.after_first_slash..].ends_with('/') {
+ self.url.serialization.pop();
+ }
+ self
+ }
+
+ /// Remove the last segment of this URL’s path.
+ ///
+ /// If the path only has one segment, make it empty such that `url.path() == "/"`.
+ ///
+ /// Returns `&mut Self` so that method calls can be chained.
+ pub fn pop(&mut self) -> &mut Self {
+ let last_slash = self.url.serialization[self.after_first_slash..].rfind('/').unwrap_or(0);
+ self.url.serialization.truncate(self.after_first_slash + last_slash);
+ self
+ }
+
+ /// Append the given segment at the end of this URL’s path.
+ ///
+ /// See the documentation for `.extend()`.
+ ///
+ /// Returns `&mut Self` so that method calls can be chained.
+ pub fn push(&mut self, segment: &str) -> &mut Self {
+ self.extend(Some(segment))
+ }
+
+ /// Append each segment from the given iterator at the end of this URL’s path.
+ ///
+ /// Each segment is percent-encoded like in `Url::parse` or `Url::join`,
+ /// except that `%` and `/` characters are also encoded (to `%25` and `%2F`).
+ /// This is unlike `Url::parse` where `%` is left as-is in case some of the input
+ /// is already percent-encoded, and `/` denotes a path segment separator.)
+ ///
+ /// Note that, in addition to slashes between new segments,
+ /// this always adds a slash between the existing path and the new segments
+ /// *except* if the existing path is `"/"`.
+ /// If the previous last segment was empty (if the path had a trailing slash)
+ /// the path after `.extend()` will contain two consecutive slashes.
+ /// If that is undesired, call `.pop_if_empty()` first.
+ ///
+ /// To obtain a behavior similar to `Url::join`, call `.pop()` unconditionally first.
+ ///
+ /// Returns `&mut Self` so that method calls can be chained.
+ ///
+ /// Example:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// let mut url = Url::parse("https://github.com/").unwrap();
+ /// let org = "servo";
+ /// let repo = "rust-url";
+ /// let issue_number = "188";
+ /// url.path_segments_mut().unwrap().extend(&[org, repo, "issues", issue_number]);
+ /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/issues/188");
+ /// ```
+ ///
+ /// In order to make sure that parsing the serialization of an URL gives the same URL,
+ /// a segment is ignored if it is `"."` or `".."`:
+ ///
+ /// ```rust
+ /// # use url::Url;
+ /// let mut url = Url::parse("https://github.com/servo").unwrap();
+ /// url.path_segments_mut().unwrap().extend(&["..", "rust-url", ".", "pulls"]);
+ /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/pulls");
+ /// ```
+ pub fn extend<I>(&mut self, segments: I) -> &mut Self
+ where I: IntoIterator, I::Item: AsRef<str> {
+ let scheme_type = SchemeType::from(self.url.scheme());
+ let path_start = self.url.path_start as usize;
+ self.url.mutate(|parser| {
+ parser.context = parser::Context::PathSegmentSetter;
+ for segment in segments {
+ let segment = segment.as_ref();
+ if matches!(segment, "." | "..") {
+ continue
+ }
+ if parser.serialization.len() > path_start + 1 {
+ parser.serialization.push('/');
+ }
+ let mut has_host = true; // FIXME account for this?
+ parser.parse_path(scheme_type, &mut has_host, path_start,
+ parser::Input::new(segment));
+ }
+ });
+ self
+ }
+
+ /// For internal testing, not part of the public API.
+ #[doc(hidden)]
+ pub fn assert_url_invariants(&mut self) -> &mut Self {
+ self.url.assert_invariants();
+ self
+ }
+}
diff --git a/third_party/rust/url/src/percent_encoding.rs b/third_party/rust/url/src/percent_encoding.rs
new file mode 100644
index 000000000..a4fb6177a
--- /dev/null
+++ b/third_party/rust/url/src/percent_encoding.rs
@@ -0,0 +1,344 @@
+// Copyright 2013-2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+use encoding;
+use std::ascii::AsciiExt;
+use std::borrow::Cow;
+use std::fmt;
+use std::slice;
+use std::str;
+
+/// Represents a set of characters / bytes that should be percent-encoded.
+///
+/// See [encode sets specification](http://url.spec.whatwg.org/#simple-encode-set).
+///
+/// Different characters need to be encoded in different parts of an URL.
+/// For example, a literal `?` question mark in an URL’s path would indicate
+/// the start of the query string.
+/// A question mark meant to be part of the path therefore needs to be percent-encoded.
+/// In the query string however, a question mark does not have any special meaning
+/// and does not need to be percent-encoded.
+///
+/// A few sets are defined in this module.
+/// Use the [`define_encode_set!`](../macro.define_encode_set!.html) macro to define different ones.
+pub trait EncodeSet: Clone {
+ /// Called with UTF-8 bytes rather than code points.
+ /// Should return true for all non-ASCII bytes.
+ fn contains(&self, byte: u8) -> bool;
+}
+
+/// Define a new struct
+/// that implements the [`EncodeSet`](percent_encoding/trait.EncodeSet.html) trait,
+/// for use in [`percent_decode()`](percent_encoding/fn.percent_encode.html)
+/// and related functions.
+///
+/// Parameters are characters to include in the set in addition to those of the base set.
+/// See [encode sets specification](http://url.spec.whatwg.org/#simple-encode-set).
+///
+/// Example
+/// =======
+///
+/// ```rust
+/// #[macro_use] extern crate url;
+/// use url::percent_encoding::{utf8_percent_encode, SIMPLE_ENCODE_SET};
+/// define_encode_set! {
+/// /// This encode set is used in the URL parser for query strings.
+/// pub QUERY_ENCODE_SET = [SIMPLE_ENCODE_SET] | {' ', '"', '#', '<', '>'}
+/// }
+/// # fn main() {
+/// assert_eq!(utf8_percent_encode("foo bar", QUERY_ENCODE_SET).collect::<String>(), "foo%20bar");
+/// # }
+/// ```
+#[macro_export]
+macro_rules! define_encode_set {
+ ($(#[$attr: meta])* pub $name: ident = [$base_set: expr] | {$($ch: pat),*}) => {
+ $(#[$attr])*
+ #[derive(Copy, Clone)]
+ #[allow(non_camel_case_types)]
+ pub struct $name;
+
+ impl $crate::percent_encoding::EncodeSet for $name {
+ #[inline]
+ fn contains(&self, byte: u8) -> bool {
+ match byte as char {
+ $(
+ $ch => true,
+ )*
+ _ => $base_set.contains(byte)
+ }
+ }
+ }
+ }
+}
+
+/// This encode set is used for the path of cannot-be-a-base URLs.
+#[derive(Copy, Clone)]
+#[allow(non_camel_case_types)]
+pub struct SIMPLE_ENCODE_SET;
+
+impl EncodeSet for SIMPLE_ENCODE_SET {
+ #[inline]
+ fn contains(&self, byte: u8) -> bool {
+ byte < 0x20 || byte > 0x7E
+ }
+}
+
+define_encode_set! {
+ /// This encode set is used in the URL parser for query strings.
+ pub QUERY_ENCODE_SET = [SIMPLE_ENCODE_SET] | {' ', '"', '#', '<', '>'}
+}
+
+define_encode_set! {
+ /// This encode set is used for path components.
+ pub DEFAULT_ENCODE_SET = [QUERY_ENCODE_SET] | {'`', '?', '{', '}'}
+}
+
+define_encode_set! {
+ /// This encode set is used for on '/'-separated path segment
+ pub PATH_SEGMENT_ENCODE_SET = [DEFAULT_ENCODE_SET] | {'%', '/'}
+}
+
+define_encode_set! {
+ /// This encode set is used for username and password.
+ pub USERINFO_ENCODE_SET = [DEFAULT_ENCODE_SET] | {
+ '/', ':', ';', '=', '@', '[', '\\', ']', '^', '|'
+ }
+}
+
+/// Return the percent-encoding of the given bytes.
+///
+/// This is unconditional, unlike `percent_encode()` which uses an encode set.
+pub fn percent_encode_byte(byte: u8) -> &'static str {
+ let index = usize::from(byte) * 3;
+ &"\
+ %00%01%02%03%04%05%06%07%08%09%0A%0B%0C%0D%0E%0F\
+ %10%11%12%13%14%15%16%17%18%19%1A%1B%1C%1D%1E%1F\
+ %20%21%22%23%24%25%26%27%28%29%2A%2B%2C%2D%2E%2F\
+ %30%31%32%33%34%35%36%37%38%39%3A%3B%3C%3D%3E%3F\
+ %40%41%42%43%44%45%46%47%48%49%4A%4B%4C%4D%4E%4F\
+ %50%51%52%53%54%55%56%57%58%59%5A%5B%5C%5D%5E%5F\
+ %60%61%62%63%64%65%66%67%68%69%6A%6B%6C%6D%6E%6F\
+ %70%71%72%73%74%75%76%77%78%79%7A%7B%7C%7D%7E%7F\
+ %80%81%82%83%84%85%86%87%88%89%8A%8B%8C%8D%8E%8F\
+ %90%91%92%93%94%95%96%97%98%99%9A%9B%9C%9D%9E%9F\
+ %A0%A1%A2%A3%A4%A5%A6%A7%A8%A9%AA%AB%AC%AD%AE%AF\
+ %B0%B1%B2%B3%B4%B5%B6%B7%B8%B9%BA%BB%BC%BD%BE%BF\
+ %C0%C1%C2%C3%C4%C5%C6%C7%C8%C9%CA%CB%CC%CD%CE%CF\
+ %D0%D1%D2%D3%D4%D5%D6%D7%D8%D9%DA%DB%DC%DD%DE%DF\
+ %E0%E1%E2%E3%E4%E5%E6%E7%E8%E9%EA%EB%EC%ED%EE%EF\
+ %F0%F1%F2%F3%F4%F5%F6%F7%F8%F9%FA%FB%FC%FD%FE%FF\
+ "[index..index + 3]
+}
+
+/// Percent-encode the given bytes with the given encode set.
+///
+/// The encode set define which bytes (in addition to non-ASCII and controls)
+/// need to be percent-encoded.
+/// The choice of this set depends on context.
+/// For example, `?` needs to be encoded in an URL path but not in a query string.
+///
+/// The return value is an iterator of `&str` slices (so it has a `.collect::<String>()` method)
+/// that also implements `Display` and `Into<Cow<str>>`.
+/// The latter returns `Cow::Borrowed` when none of the bytes in `input`
+/// are in the given encode set.
+#[inline]
+pub fn percent_encode<E: EncodeSet>(input: &[u8], encode_set: E) -> PercentEncode<E> {
+ PercentEncode {
+ bytes: input,
+ encode_set: encode_set,
+ }
+}
+
+/// Percent-encode the UTF-8 encoding of the given string.
+///
+/// See `percent_encode()` for how to use the return value.
+#[inline]
+pub fn utf8_percent_encode<E: EncodeSet>(input: &str, encode_set: E) -> PercentEncode<E> {
+ percent_encode(input.as_bytes(), encode_set)
+}
+
+/// The return type of `percent_encode()` and `utf8_percent_encode()`.
+#[derive(Clone)]
+pub struct PercentEncode<'a, E: EncodeSet> {
+ bytes: &'a [u8],
+ encode_set: E,
+}
+
+impl<'a, E: EncodeSet> Iterator for PercentEncode<'a, E> {
+ type Item = &'a str;
+
+ fn next(&mut self) -> Option<&'a str> {
+ if let Some((&first_byte, remaining)) = self.bytes.split_first() {
+ if self.encode_set.contains(first_byte) {
+ self.bytes = remaining;
+ Some(percent_encode_byte(first_byte))
+ } else {
+ assert!(first_byte.is_ascii());
+ for (i, &byte) in remaining.iter().enumerate() {
+ if self.encode_set.contains(byte) {
+ // 1 for first_byte + i for previous iterations of this loop
+ let (unchanged_slice, remaining) = self.bytes.split_at(1 + i);
+ self.bytes = remaining;
+ return Some(unsafe { str::from_utf8_unchecked(unchanged_slice) })
+ } else {
+ assert!(byte.is_ascii());
+ }
+ }
+ let unchanged_slice = self.bytes;
+ self.bytes = &[][..];
+ Some(unsafe { str::from_utf8_unchecked(unchanged_slice) })
+ }
+ } else {
+ None
+ }
+ }
+
+ fn size_hint(&self) -> (usize, Option<usize>) {
+ if self.bytes.is_empty() {
+ (0, Some(0))
+ } else {
+ (1, Some(self.bytes.len()))
+ }
+ }
+}
+
+impl<'a, E: EncodeSet> fmt::Display for PercentEncode<'a, E> {
+ fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
+ for c in (*self).clone() {
+ try!(formatter.write_str(c))
+ }
+ Ok(())
+ }
+}
+
+impl<'a, E: EncodeSet> From<PercentEncode<'a, E>> for Cow<'a, str> {
+ fn from(mut iter: PercentEncode<'a, E>) -> Self {
+ match iter.next() {
+ None => "".into(),
+ Some(first) => {
+ match iter.next() {
+ None => first.into(),
+ Some(second) => {
+ let mut string = first.to_owned();
+ string.push_str(second);
+ string.extend(iter);
+ string.into()
+ }
+ }
+ }
+ }
+ }
+}
+
+/// Percent-decode the given bytes.
+///
+/// The return value is an iterator of decoded `u8` bytes
+/// that also implements `Into<Cow<u8>>`
+/// (which returns `Cow::Borrowed` when `input` contains no percent-encoded sequence)
+/// and has `decode_utf8()` and `decode_utf8_lossy()` methods.
+#[inline]
+pub fn percent_decode<'a>(input: &'a [u8]) -> PercentDecode<'a> {
+ PercentDecode {
+ bytes: input.iter()
+ }
+}
+
+/// The return type of `percent_decode()`.
+#[derive(Clone)]
+pub struct PercentDecode<'a> {
+ bytes: slice::Iter<'a, u8>,
+}
+
+fn after_percent_sign(iter: &mut slice::Iter<u8>) -> Option<u8> {
+ let initial_iter = iter.clone();
+ let h = iter.next().and_then(|&b| (b as char).to_digit(16));
+ let l = iter.next().and_then(|&b| (b as char).to_digit(16));
+ if let (Some(h), Some(l)) = (h, l) {
+ Some(h as u8 * 0x10 + l as u8)
+ } else {
+ *iter = initial_iter;
+ None
+ }
+}
+
+impl<'a> Iterator for PercentDecode<'a> {
+ type Item = u8;
+
+ fn next(&mut self) -> Option<u8> {
+ self.bytes.next().map(|&byte| {
+ if byte == b'%' {
+ after_percent_sign(&mut self.bytes).unwrap_or(byte)
+ } else {
+ byte
+ }
+ })
+ }
+
+ fn size_hint(&self) -> (usize, Option<usize>) {
+ let bytes = self.bytes.len();
+ (bytes / 3, Some(bytes))
+ }
+}
+
+impl<'a> From<PercentDecode<'a>> for Cow<'a, [u8]> {
+ fn from(iter: PercentDecode<'a>) -> Self {
+ match iter.if_any() {
+ Some(vec) => Cow::Owned(vec),
+ None => Cow::Borrowed(iter.bytes.as_slice()),
+ }
+ }
+}
+
+impl<'a> PercentDecode<'a> {
+ /// If the percent-decoding is different from the input, return it as a new bytes vector.
+ pub fn if_any(&self) -> Option<Vec<u8>> {
+ let mut bytes_iter = self.bytes.clone();
+ while bytes_iter.find(|&&b| b == b'%').is_some() {
+ if let Some(decoded_byte) = after_percent_sign(&mut bytes_iter) {
+ let initial_bytes = self.bytes.as_slice();
+ let unchanged_bytes_len = initial_bytes.len() - bytes_iter.len() - 3;
+ let mut decoded = initial_bytes[..unchanged_bytes_len].to_owned();
+ decoded.push(decoded_byte);
+ decoded.extend(PercentDecode {
+ bytes: bytes_iter
+ });
+ return Some(decoded)
+ }
+ }
+ // Nothing to decode
+ None
+ }
+
+ /// Decode the result of percent-decoding as UTF-8.
+ ///
+ /// This is return `Err` when the percent-decoded bytes are not well-formed in UTF-8.
+ pub fn decode_utf8(self) -> Result<Cow<'a, str>, str::Utf8Error> {
+ match self.clone().into() {
+ Cow::Borrowed(bytes) => {
+ match str::from_utf8(bytes) {
+ Ok(s) => Ok(s.into()),
+ Err(e) => Err(e),
+ }
+ }
+ Cow::Owned(bytes) => {
+ match String::from_utf8(bytes) {
+ Ok(s) => Ok(s.into()),
+ Err(e) => Err(e.utf8_error()),
+ }
+ }
+ }
+ }
+
+ /// Decode the result of percent-decoding as UTF-8, lossily.
+ ///
+ /// Invalid UTF-8 percent-encoded byte sequences will be replaced � U+FFFD,
+ /// the replacement character.
+ pub fn decode_utf8_lossy(self) -> Cow<'a, str> {
+ encoding::decode_utf8_lossy(self.clone().into())
+ }
+}
diff --git a/third_party/rust/url/src/quirks.rs b/third_party/rust/url/src/quirks.rs
new file mode 100644
index 000000000..9a7537f47
--- /dev/null
+++ b/third_party/rust/url/src/quirks.rs
@@ -0,0 +1,217 @@
+// Copyright 2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+//! Getters and setters for URL components implemented per https://url.spec.whatwg.org/#api
+//!
+//! Unless you need to be interoperable with web browsers,
+//! you probably want to use `Url` method instead.
+
+use {Url, Position, Host, ParseError, idna};
+use parser::{Parser, SchemeType, default_port, Context, Input};
+
+/// https://url.spec.whatwg.org/#dom-url-domaintoascii
+pub fn domain_to_ascii(domain: &str) -> String {
+ match Host::parse(domain) {
+ Ok(Host::Domain(domain)) => domain,
+ _ => String::new(),
+ }
+}
+
+/// https://url.spec.whatwg.org/#dom-url-domaintounicode
+pub fn domain_to_unicode(domain: &str) -> String {
+ match Host::parse(domain) {
+ Ok(Host::Domain(ref domain)) => {
+ let (unicode, _errors) = idna::domain_to_unicode(domain);
+ unicode
+ }
+ _ => String::new(),
+ }
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-href
+pub fn href(url: &Url) -> &str {
+ url.as_str()
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-href
+pub fn set_href(url: &mut Url, value: &str) -> Result<(), ParseError> {
+ *url = try!(Url::parse(value));
+ Ok(())
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-origin
+pub fn origin(url: &Url) -> String {
+ url.origin().unicode_serialization()
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-protocol
+#[inline]
+pub fn protocol(url: &Url) -> &str {
+ &url.as_str()[..url.scheme().len() + ":".len()]
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-protocol
+pub fn set_protocol(url: &mut Url, mut new_protocol: &str) -> Result<(), ()> {
+ // The scheme state in the spec ignores everything after the first `:`,
+ // but `set_scheme` errors if there is more.
+ if let Some(position) = new_protocol.find(':') {
+ new_protocol = &new_protocol[..position];
+ }
+ url.set_scheme(new_protocol)
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-username
+#[inline]
+pub fn username(url: &Url) -> &str {
+ url.username()
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-username
+pub fn set_username(url: &mut Url, new_username: &str) -> Result<(), ()> {
+ url.set_username(new_username)
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-password
+#[inline]
+pub fn password(url: &Url) -> &str {
+ url.password().unwrap_or("")
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-password
+pub fn set_password(url: &mut Url, new_password: &str) -> Result<(), ()> {
+ url.set_password(if new_password.is_empty() { None } else { Some(new_password) })
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-host
+#[inline]
+pub fn host(url: &Url) -> &str {
+ &url[Position::BeforeHost..Position::AfterPort]
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-host
+pub fn set_host(url: &mut Url, new_host: &str) -> Result<(), ()> {
+ if url.cannot_be_a_base() {
+ return Err(())
+ }
+ let host;
+ let opt_port;
+ {
+ let scheme = url.scheme();
+ let result = Parser::parse_host(Input::new(new_host), SchemeType::from(scheme));
+ match result {
+ Ok((h, remaining)) => {
+ host = h;
+ opt_port = if let Some(remaining) = remaining.split_prefix(':') {
+ Parser::parse_port(remaining, || default_port(scheme), Context::Setter)
+ .ok().map(|(port, _remaining)| port)
+ } else {
+ None
+ };
+ }
+ Err(_) => return Err(())
+ }
+ }
+ url.set_host_internal(host, opt_port);
+ Ok(())
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-hostname
+#[inline]
+pub fn hostname(url: &Url) -> &str {
+ url.host_str().unwrap_or("")
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-hostname
+pub fn set_hostname(url: &mut Url, new_hostname: &str) -> Result<(), ()> {
+ if url.cannot_be_a_base() {
+ return Err(())
+ }
+ let result = Parser::parse_host(Input::new(new_hostname), SchemeType::from(url.scheme()));
+ if let Ok((host, _remaining)) = result {
+ url.set_host_internal(host, None);
+ Ok(())
+ } else {
+ Err(())
+ }
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-port
+#[inline]
+pub fn port(url: &Url) -> &str {
+ &url[Position::BeforePort..Position::AfterPort]
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-port
+pub fn set_port(url: &mut Url, new_port: &str) -> Result<(), ()> {
+ let result;
+ {
+ // has_host implies !cannot_be_a_base
+ let scheme = url.scheme();
+ if !url.has_host() || scheme == "file" {
+ return Err(())
+ }
+ result = Parser::parse_port(Input::new(new_port), || default_port(scheme), Context::Setter)
+ }
+ if let Ok((new_port, _remaining)) = result {
+ url.set_port_internal(new_port);
+ Ok(())
+ } else {
+ Err(())
+ }
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-pathname
+#[inline]
+pub fn pathname(url: &Url) -> &str {
+ url.path()
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-pathname
+pub fn set_pathname(url: &mut Url, new_pathname: &str) {
+ if !url.cannot_be_a_base() {
+ url.set_path(new_pathname)
+ }
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-search
+pub fn search(url: &Url) -> &str {
+ trim(&url[Position::AfterPath..Position::AfterQuery])
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-search
+pub fn set_search(url: &mut Url, new_search: &str) {
+ url.set_query(match new_search {
+ "" => None,
+ _ if new_search.starts_with('?') => Some(&new_search[1..]),
+ _ => Some(new_search),
+ })
+}
+
+/// Getter for https://url.spec.whatwg.org/#dom-url-hash
+pub fn hash(url: &Url) -> &str {
+ trim(&url[Position::AfterQuery..])
+}
+
+/// Setter for https://url.spec.whatwg.org/#dom-url-hash
+pub fn set_hash(url: &mut Url, new_hash: &str) {
+ if url.scheme() != "javascript" {
+ url.set_fragment(match new_hash {
+ "" => None,
+ _ if new_hash.starts_with('#') => Some(&new_hash[1..]),
+ _ => Some(new_hash),
+ })
+ }
+}
+
+fn trim(s: &str) -> &str {
+ if s.len() == 1 {
+ ""
+ } else {
+ s
+ }
+}
diff --git a/third_party/rust/url/src/slicing.rs b/third_party/rust/url/src/slicing.rs
new file mode 100644
index 000000000..926f3c796
--- /dev/null
+++ b/third_party/rust/url/src/slicing.rs
@@ -0,0 +1,182 @@
+// Copyright 2016 The rust-url developers.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+use std::ops::{Range, RangeFrom, RangeTo, RangeFull, Index};
+use Url;
+
+impl Index<RangeFull> for Url {
+ type Output = str;
+ fn index(&self, _: RangeFull) -> &str {
+ &self.serialization
+ }
+}
+
+impl Index<RangeFrom<Position>> for Url {
+ type Output = str;
+ fn index(&self, range: RangeFrom<Position>) -> &str {
+ &self.serialization[self.index(range.start)..]
+ }
+}
+
+impl Index<RangeTo<Position>> for Url {
+ type Output = str;
+ fn index(&self, range: RangeTo<Position>) -> &str {
+ &self.serialization[..self.index(range.end)]
+ }
+}
+
+impl Index<Range<Position>> for Url {
+ type Output = str;
+ fn index(&self, range: Range<Position>) -> &str {
+ &self.serialization[self.index(range.start)..self.index(range.end)]
+ }
+}
+
+/// Indicates a position within a URL based on its components.
+///
+/// A range of positions can be used for slicing `Url`:
+///
+/// ```rust
+/// # use url::{Url, Position};
+/// # fn something(some_url: Url) {
+/// let serialization: &str = &some_url[..];
+/// let serialization_without_fragment: &str = &some_url[..Position::AfterQuery];
+/// let authority: &str = &some_url[Position::BeforeUsername..Position::AfterPort];
+/// let data_url_payload: &str = &some_url[Position::BeforePath..Position::AfterQuery];
+/// let scheme_relative: &str = &some_url[Position::BeforeUsername..];
+/// # }
+/// ```
+///
+/// In a pseudo-grammar (where `[`…`]?` makes a sub-sequence optional),
+/// URL components and delimiters that separate them are:
+///
+/// ```notrust
+/// url =
+/// scheme ":"
+/// [ "//" [ username [ ":" password ]? "@" ]? host [ ":" port ]? ]?
+/// path [ "?" query ]? [ "#" fragment ]?
+/// ```
+///
+/// When a given component is not present,
+/// its "before" and "after" position are the same
+/// (so that `&some_url[BeforeFoo..AfterFoo]` is the empty string)
+/// and component ordering is preserved
+/// (so that a missing query "is between" a path and a fragment).
+///
+/// The end of a component and the start of the next are either the same or separate
+/// by a delimiter.
+/// (Not that the initial `/` of a path is considered part of the path here, not a delimiter.)
+/// For example, `&url[..BeforeFragment]` would include a `#` delimiter (if present in `url`),
+/// so `&url[..AfterQuery]` might be desired instead.
+///
+/// `BeforeScheme` and `AfterFragment` are always the start and end of the entire URL,
+/// so `&url[BeforeScheme..X]` is the same as `&url[..X]`
+/// and `&url[X..AfterFragment]` is the same as `&url[X..]`.
+#[derive(Copy, Clone, Debug)]
+pub enum Position {
+ BeforeScheme,
+ AfterScheme,
+ BeforeUsername,
+ AfterUsername,
+ BeforePassword,
+ AfterPassword,
+ BeforeHost,
+ AfterHost,
+ BeforePort,
+ AfterPort,
+ BeforePath,
+ AfterPath,
+ BeforeQuery,
+ AfterQuery,
+ BeforeFragment,
+ AfterFragment
+}
+
+impl Url {
+ #[inline]
+ fn index(&self, position: Position) -> usize {
+ match position {
+ Position::BeforeScheme => 0,
+
+ Position::AfterScheme => self.scheme_end as usize,
+
+ Position::BeforeUsername => if self.has_authority() {
+ self.scheme_end as usize + "://".len()
+ } else {
+ debug_assert!(self.byte_at(self.scheme_end) == b':');
+ debug_assert!(self.scheme_end + ":".len() as u32 == self.username_end);
+ self.scheme_end as usize + ":".len()
+ },
+
+ Position::AfterUsername => self.username_end as usize,
+
+ Position::BeforePassword => if self.has_authority() &&
+ self.byte_at(self.username_end) == b':' {
+ self.username_end as usize + ":".len()
+ } else {
+ debug_assert!(self.username_end == self.host_start);
+ self.username_end as usize
+ },
+
+ Position::AfterPassword => if self.has_authority() &&
+ self.byte_at(self.username_end) == b':' {
+ debug_assert!(self.byte_at(self.host_start - "@".len() as u32) == b'@');
+ self.host_start as usize - "@".len()
+ } else {
+ debug_assert!(self.username_end == self.host_start);
+ self.host_start as usize
+ },
+
+ Position::BeforeHost => self.host_start as usize,
+
+ Position::AfterHost => self.host_end as usize,
+
+ Position::BeforePort => if self.port.is_some() {
+ debug_assert!(self.byte_at(self.host_end) == b':');
+ self.host_end as usize + ":".len()
+ } else {
+ self.host_end as usize
+ },
+
+ Position::AfterPort => self.path_start as usize,
+
+ Position::BeforePath => self.path_start as usize,
+
+ Position::AfterPath => match (self.query_start, self.fragment_start) {
+ (Some(q), _) => q as usize,
+ (None, Some(f)) => f as usize,
+ (None, None) => self.serialization.len(),
+ },
+
+ Position::BeforeQuery => match (self.query_start, self.fragment_start) {
+ (Some(q), _) => {
+ debug_assert!(self.byte_at(q) == b'?');
+ q as usize + "?".len()
+ }
+ (None, Some(f)) => f as usize,
+ (None, None) => self.serialization.len(),
+ },
+
+ Position::AfterQuery => match self.fragment_start {
+ None => self.serialization.len(),
+ Some(f) => f as usize,
+ },
+
+ Position::BeforeFragment => match self.fragment_start {
+ Some(f) => {
+ debug_assert!(self.byte_at(f) == b'#');
+ f as usize + "#".len()
+ }
+ None => self.serialization.len(),
+ },
+
+ Position::AfterFragment => self.serialization.len(),
+ }
+ }
+}
+
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-27 20:53:18 +0000