percent_encoding/
percent_encoding.rs

1// Copyright 2013-2016 The rust-url developers.
2//
3// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
4// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
5// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
6// option. This file may not be copied, modified, or distributed
7// except according to those terms.
8
9//! URLs use special characters to indicate the parts of the request.
10//! For example, a `?` question mark marks the end of a path and the start of a query string.
11//! In order for that character to exist inside a path, it needs to be encoded differently.
12//!
13//! Percent encoding replaces reserved characters with the `%` escape character
14//! followed by a byte value as two hexadecimal digits.
15//! For example, an ASCII space is replaced with `%20`.
16//!
17//! When encoding, the set of characters that can (and should, for readability) be left alone
18//! depends on the context.
19//! The `?` question mark mentioned above is not a separator when used literally
20//! inside of a query string, and therefore does not need to be encoded.
21//! The [`AsciiSet`] parameter of [`percent_encode`] and [`utf8_percent_encode`]
22//! lets callers configure this.
23//!
24//! This crate deliberately does not provide many different sets.
25//! Users should consider in what context the encoded string will be used,
26//! read relevant specifications, and define their own set.
27//! This is done by using the `add` method of an existing set.
28//!
29//! # Examples
30//!
31//! ```
32//! use percent_encoding::{utf8_percent_encode, AsciiSet, CONTROLS};
33//!
34//! /// https://url.spec.whatwg.org/#fragment-percent-encode-set
35//! const FRAGMENT: &AsciiSet = &CONTROLS.add(b' ').add(b'"').add(b'<').add(b'>').add(b'`');
36//!
37//! assert_eq!(utf8_percent_encode("foo <bar>", FRAGMENT).to_string(), "foo%20%3Cbar%3E");
38//! ```
39#![no_std]
40
41// For forwards compatibility
42#[cfg(feature = "std")]
43extern crate std as _;
44
45#[cfg(feature = "alloc")]
46extern crate alloc;
47
48#[cfg(feature = "alloc")]
49use alloc::{
50    borrow::{Cow, ToOwned},
51    string::String,
52    vec::Vec,
53};
54use core::{fmt, slice, str};
55
56pub use self::ascii_set::{AsciiSet, CONTROLS, NON_ALPHANUMERIC};
57
58mod ascii_set;
59
60/// Return the percent-encoding of the given byte.
61///
62/// This is unconditional, unlike `percent_encode()` which has an `AsciiSet` parameter.
63///
64/// # Examples
65///
66/// ```
67/// use percent_encoding::percent_encode_byte;
68///
69/// assert_eq!("foo bar".bytes().map(percent_encode_byte).collect::<String>(),
70///            "%66%6F%6F%20%62%61%72");
71/// ```
72#[inline]
73pub fn percent_encode_byte(byte: u8) -> &'static str {
74    static ENC_TABLE: &[u8; 768] = b"\
75      %00%01%02%03%04%05%06%07%08%09%0A%0B%0C%0D%0E%0F\
76      %10%11%12%13%14%15%16%17%18%19%1A%1B%1C%1D%1E%1F\
77      %20%21%22%23%24%25%26%27%28%29%2A%2B%2C%2D%2E%2F\
78      %30%31%32%33%34%35%36%37%38%39%3A%3B%3C%3D%3E%3F\
79      %40%41%42%43%44%45%46%47%48%49%4A%4B%4C%4D%4E%4F\
80      %50%51%52%53%54%55%56%57%58%59%5A%5B%5C%5D%5E%5F\
81      %60%61%62%63%64%65%66%67%68%69%6A%6B%6C%6D%6E%6F\
82      %70%71%72%73%74%75%76%77%78%79%7A%7B%7C%7D%7E%7F\
83      %80%81%82%83%84%85%86%87%88%89%8A%8B%8C%8D%8E%8F\
84      %90%91%92%93%94%95%96%97%98%99%9A%9B%9C%9D%9E%9F\
85      %A0%A1%A2%A3%A4%A5%A6%A7%A8%A9%AA%AB%AC%AD%AE%AF\
86      %B0%B1%B2%B3%B4%B5%B6%B7%B8%B9%BA%BB%BC%BD%BE%BF\
87      %C0%C1%C2%C3%C4%C5%C6%C7%C8%C9%CA%CB%CC%CD%CE%CF\
88      %D0%D1%D2%D3%D4%D5%D6%D7%D8%D9%DA%DB%DC%DD%DE%DF\
89      %E0%E1%E2%E3%E4%E5%E6%E7%E8%E9%EA%EB%EC%ED%EE%EF\
90      %F0%F1%F2%F3%F4%F5%F6%F7%F8%F9%FA%FB%FC%FD%FE%FF\
91      ";
92
93    let index = usize::from(byte) * 3;
94    // SAFETY: ENC_TABLE is ascii-only, so any subset of it should be
95    // ascii-only too, which is valid utf8.
96    unsafe { str::from_utf8_unchecked(&ENC_TABLE[index..index + 3]) }
97}
98
99/// Percent-encode the given bytes with the given set.
100///
101/// Non-ASCII bytes and bytes in `ascii_set` are encoded.
102///
103/// The return type:
104///
105/// * Implements `Iterator<Item = &str>` and therefore has a `.collect::<String>()` method,
106/// * Implements `Display` and therefore has a `.to_string()` method,
107/// * Implements `Into<Cow<str>>` borrowing `input` when none of its bytes are encoded.
108///
109/// # Examples
110///
111/// ```
112/// use percent_encoding::{percent_encode, NON_ALPHANUMERIC};
113///
114/// assert_eq!(percent_encode(b"foo bar?", NON_ALPHANUMERIC).to_string(), "foo%20bar%3F");
115/// ```
116#[inline]
117pub fn percent_encode<'a>(input: &'a [u8], ascii_set: &'static AsciiSet) -> PercentEncode<'a> {
118    PercentEncode {
119        bytes: input,
120        ascii_set,
121    }
122}
123
124/// Percent-encode the UTF-8 encoding of the given string.
125///
126/// See [`percent_encode`] regarding the return type.
127///
128/// # Examples
129///
130/// ```
131/// use percent_encoding::{utf8_percent_encode, NON_ALPHANUMERIC};
132///
133/// assert_eq!(utf8_percent_encode("foo bar?", NON_ALPHANUMERIC).to_string(), "foo%20bar%3F");
134/// ```
135#[inline]
136pub fn utf8_percent_encode<'a>(input: &'a str, ascii_set: &'static AsciiSet) -> PercentEncode<'a> {
137    percent_encode(input.as_bytes(), ascii_set)
138}
139
140/// The return type of [`percent_encode`] and [`utf8_percent_encode`].
141#[derive(Debug, Clone, PartialEq, Eq)]
142pub struct PercentEncode<'a> {
143    bytes: &'a [u8],
144    ascii_set: &'static AsciiSet,
145}
146
147impl<'a> Iterator for PercentEncode<'a> {
148    type Item = &'a str;
149
150    fn next(&mut self) -> Option<&'a str> {
151        if let Some((&first_byte, remaining)) = self.bytes.split_first() {
152            if self.ascii_set.should_percent_encode(first_byte) {
153                self.bytes = remaining;
154                Some(percent_encode_byte(first_byte))
155            } else {
156                // The unsafe blocks here are appropriate because the bytes are
157                // confirmed as a subset of UTF-8 in should_percent_encode.
158                for (i, &byte) in remaining.iter().enumerate() {
159                    if self.ascii_set.should_percent_encode(byte) {
160                        // 1 for first_byte + i for previous iterations of this loop
161                        let (unchanged_slice, remaining) = self.bytes.split_at(1 + i);
162                        self.bytes = remaining;
163                        return Some(unsafe { str::from_utf8_unchecked(unchanged_slice) });
164                    }
165                }
166                let unchanged_slice = self.bytes;
167                self.bytes = &[][..];
168                Some(unsafe { str::from_utf8_unchecked(unchanged_slice) })
169            }
170        } else {
171            None
172        }
173    }
174
175    fn size_hint(&self) -> (usize, Option<usize>) {
176        if self.bytes.is_empty() {
177            (0, Some(0))
178        } else {
179            (1, Some(self.bytes.len()))
180        }
181    }
182}
183
184impl fmt::Display for PercentEncode<'_> {
185    fn fmt(&self, formatter: &mut fmt::Formatter<'_>) -> fmt::Result {
186        for c in (*self).clone() {
187            formatter.write_str(c)?
188        }
189        Ok(())
190    }
191}
192
193#[cfg(feature = "alloc")]
194impl<'a> From<PercentEncode<'a>> for Cow<'a, str> {
195    fn from(mut iter: PercentEncode<'a>) -> Self {
196        match iter.next() {
197            None => "".into(),
198            Some(first) => match iter.next() {
199                None => first.into(),
200                Some(second) => {
201                    let mut string = first.to_owned();
202                    string.push_str(second);
203                    string.extend(iter);
204                    string.into()
205                }
206            },
207        }
208    }
209}
210
211/// Percent-decode the given string.
212///
213/// <https://url.spec.whatwg.org/#string-percent-decode>
214///
215/// See [`percent_decode`] regarding the return type.
216#[inline]
217pub fn percent_decode_str(input: &str) -> PercentDecode<'_> {
218    percent_decode(input.as_bytes())
219}
220
221/// Percent-decode the given bytes.
222///
223/// <https://url.spec.whatwg.org/#percent-decode>
224///
225/// Any sequence of `%` followed by two hexadecimal digits is decoded.
226/// The return type:
227///
228/// * Implements `Into<Cow<u8>>` borrowing `input` when it contains no percent-encoded sequence,
229/// * Implements `Iterator<Item = u8>` and therefore has a `.collect::<Vec<u8>>()` method,
230/// * Has `decode_utf8()` and `decode_utf8_lossy()` methods.
231///
232/// # Examples
233///
234/// ```
235/// use percent_encoding::percent_decode;
236///
237/// assert_eq!(percent_decode(b"foo%20bar%3f").decode_utf8().unwrap(), "foo bar?");
238/// ```
239#[inline]
240pub fn percent_decode(input: &[u8]) -> PercentDecode<'_> {
241    PercentDecode {
242        bytes: input.iter(),
243    }
244}
245
246/// The return type of [`percent_decode`].
247#[derive(Clone, Debug)]
248pub struct PercentDecode<'a> {
249    bytes: slice::Iter<'a, u8>,
250}
251
252fn after_percent_sign(iter: &mut slice::Iter<'_, u8>) -> Option<u8> {
253    let mut cloned_iter = iter.clone();
254    let h = char::from(*cloned_iter.next()?).to_digit(16)?;
255    let l = char::from(*cloned_iter.next()?).to_digit(16)?;
256    *iter = cloned_iter;
257    Some(h as u8 * 0x10 + l as u8)
258}
259
260impl Iterator for PercentDecode<'_> {
261    type Item = u8;
262
263    fn next(&mut self) -> Option<u8> {
264        self.bytes.next().map(|&byte| {
265            if byte == b'%' {
266                after_percent_sign(&mut self.bytes).unwrap_or(byte)
267            } else {
268                byte
269            }
270        })
271    }
272
273    fn size_hint(&self) -> (usize, Option<usize>) {
274        let bytes = self.bytes.len();
275        ((bytes + 2) / 3, Some(bytes))
276    }
277}
278
279#[cfg(feature = "alloc")]
280impl<'a> From<PercentDecode<'a>> for Cow<'a, [u8]> {
281    fn from(iter: PercentDecode<'a>) -> Self {
282        match iter.if_any() {
283            Some(vec) => Cow::Owned(vec),
284            None => Cow::Borrowed(iter.bytes.as_slice()),
285        }
286    }
287}
288
289impl<'a> PercentDecode<'a> {
290    /// If the percent-decoding is different from the input, return it as a new bytes vector.
291    #[cfg(feature = "alloc")]
292    fn if_any(&self) -> Option<Vec<u8>> {
293        let mut bytes_iter = self.bytes.clone();
294        while bytes_iter.any(|&b| b == b'%') {
295            if let Some(decoded_byte) = after_percent_sign(&mut bytes_iter) {
296                let initial_bytes = self.bytes.as_slice();
297                let unchanged_bytes_len = initial_bytes.len() - bytes_iter.len() - 3;
298                let mut decoded = initial_bytes[..unchanged_bytes_len].to_owned();
299                decoded.push(decoded_byte);
300                decoded.extend(PercentDecode {
301                    bytes: bytes_iter,
302                });
303                return Some(decoded);
304            }
305        }
306        // Nothing to decode
307        None
308    }
309
310    /// Decode the result of percent-decoding as UTF-8.
311    ///
312    /// This is return `Err` when the percent-decoded bytes are not well-formed in UTF-8.
313    #[cfg(feature = "alloc")]
314    pub fn decode_utf8(self) -> Result<Cow<'a, str>, str::Utf8Error> {
315        match self.clone().into() {
316            Cow::Borrowed(bytes) => match str::from_utf8(bytes) {
317                Ok(s) => Ok(s.into()),
318                Err(e) => Err(e),
319            },
320            Cow::Owned(bytes) => match String::from_utf8(bytes) {
321                Ok(s) => Ok(s.into()),
322                Err(e) => Err(e.utf8_error()),
323            },
324        }
325    }
326
327    /// Decode the result of percent-decoding as UTF-8, lossily.
328    ///
329    /// Invalid UTF-8 percent-encoded byte sequences will be replaced � U+FFFD,
330    /// the replacement character.
331    #[cfg(feature = "alloc")]
332    pub fn decode_utf8_lossy(self) -> Cow<'a, str> {
333        decode_utf8_lossy(self.clone().into())
334    }
335}
336
337// std::ptr::addr_eq was stabilized in rust 1.76. Once we upgrade
338// the MSRV we can remove this lint override.
339#[cfg(feature = "alloc")]
340#[allow(ambiguous_wide_pointer_comparisons)]
341fn decode_utf8_lossy(input: Cow<'_, [u8]>) -> Cow<'_, str> {
342    // Note: This function is duplicated in `form_urlencoded/src/query_encoding.rs`.
343    match input {
344        Cow::Borrowed(bytes) => String::from_utf8_lossy(bytes),
345        Cow::Owned(bytes) => {
346            match String::from_utf8_lossy(&bytes) {
347                Cow::Borrowed(utf8) => {
348                    // If from_utf8_lossy returns a Cow::Borrowed, then we can
349                    // be sure our original bytes were valid UTF-8. This is because
350                    // if the bytes were invalid UTF-8 from_utf8_lossy would have
351                    // to allocate a new owned string to back the Cow so it could
352                    // replace invalid bytes with a placeholder.
353
354                    // First we do a debug_assert to confirm our description above.
355                    let raw_utf8: *const [u8] = utf8.as_bytes();
356                    debug_assert!(core::ptr::eq(raw_utf8, &*bytes));
357
358                    // Given we know the original input bytes are valid UTF-8,
359                    // and we have ownership of those bytes, we re-use them and
360                    // return a Cow::Owned here.
361                    Cow::Owned(unsafe { String::from_utf8_unchecked(bytes) })
362                }
363                Cow::Owned(s) => Cow::Owned(s),
364            }
365        }
366    }
367}
368
369#[cfg(test)]
370mod tests {
371
372    use super::*;
373
374    #[test]
375    fn percent_encode_byte() {
376        for i in 0..=0xFF {
377            let encoded = super::percent_encode_byte(i);
378            assert_eq!(encoded, alloc::format!("%{:02X}", i));
379        }
380    }
381
382    #[test]
383    fn percent_encode_accepts_ascii_set_ref() {
384        let encoded = percent_encode(b"foo bar?", &AsciiSet::EMPTY);
385        assert_eq!(encoded.collect::<String>(), "foo bar?");
386    }
387
388    #[test]
389    fn percent_encode_collect() {
390        let encoded = percent_encode(b"foo bar?", NON_ALPHANUMERIC);
391        assert_eq!(encoded.collect::<String>(), String::from("foo%20bar%3F"));
392
393        let encoded = percent_encode(b"\x00\x01\x02\x03", CONTROLS);
394        assert_eq!(encoded.collect::<String>(), String::from("%00%01%02%03"));
395    }
396
397    #[test]
398    fn percent_encode_display() {
399        let encoded = percent_encode(b"foo bar?", NON_ALPHANUMERIC);
400        assert_eq!(alloc::format!("{}", encoded), "foo%20bar%3F");
401    }
402
403    #[test]
404    fn percent_encode_cow() {
405        let encoded = percent_encode(b"foo bar?", NON_ALPHANUMERIC);
406        assert_eq!(Cow::from(encoded), "foo%20bar%3F");
407    }
408
409    #[test]
410    fn utf8_percent_encode_accepts_ascii_set_ref() {
411        let encoded = super::utf8_percent_encode("foo bar?", &AsciiSet::EMPTY);
412        assert_eq!(encoded.collect::<String>(), "foo bar?");
413    }
414
415    #[test]
416    fn utf8_percent_encode() {
417        assert_eq!(
418            super::utf8_percent_encode("foo bar?", NON_ALPHANUMERIC),
419            percent_encode(b"foo bar?", NON_ALPHANUMERIC)
420        );
421    }
422
423    #[test]
424    fn percent_decode() {
425        assert_eq!(super::percent_decode(b"foo%20bar%3f").decode_utf8().unwrap(), "foo bar?");
426    }
427
428    #[test]
429    fn percent_decode_str() {
430        assert_eq!(super::percent_decode_str("foo%20bar%3f").decode_utf8().unwrap(), "foo bar?");
431    }
432
433    #[test]
434    fn percent_decode_collect() {
435        let decoded = super::percent_decode(b"foo%20bar%3f");
436        assert_eq!(decoded.collect::<Vec<u8>>(), b"foo bar?");
437    }
438
439    #[test]
440    fn percent_decode_cow() {
441        let decoded = super::percent_decode(b"foo%20bar%3f");
442        assert_eq!(Cow::from(decoded), Cow::Owned::<[u8]>(b"foo bar?".to_vec()));
443
444        let decoded = super::percent_decode(b"foo bar?");
445        assert_eq!(Cow::from(decoded), Cow::Borrowed(b"foo bar?"));
446    }
447
448    #[test]
449    fn percent_decode_invalid_utf8() {
450        // Invalid UTF-8 sequence
451        let decoded = super::percent_decode(b"%00%9F%92%96").decode_utf8().unwrap_err();
452        assert_eq!(decoded.valid_up_to(), 1);
453        assert_eq!(decoded.error_len(), Some(1));
454    }
455
456    #[test]
457    fn percent_decode_utf8_lossy() {
458        assert_eq!(super::percent_decode(b"%F0%9F%92%96").decode_utf8_lossy(), "💖");
459    }
460
461    #[test]
462    fn percent_decode_utf8_lossy_invalid_utf8() {
463        assert_eq!(super::percent_decode(b"%00%9F%92%96").decode_utf8_lossy(), "\u{0}���");
464    }
465}
percent_encoding/percent_encoding.rs

percent_encoding/
percent_encoding.rs
