1 | // Copyright 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 | use crate::parser::{self, to_u32, SchemeType}; |
10 | use crate::Url; |
11 | use alloc::string::String; |
12 | use core::str; |
13 | |
14 | /// Exposes methods to manipulate the path of an URL that is not cannot-be-base. |
15 | /// |
16 | /// The path always starts with a `/` slash, and is made of slash-separated segments. |
17 | /// There is always at least one segment (which may be the empty string). |
18 | /// |
19 | /// Examples: |
20 | /// |
21 | /// ```rust |
22 | /// use url::Url; |
23 | /// |
24 | /// # #[cfg (feature = "std" )] |
25 | /// # use std::error::Error; |
26 | /// # #[cfg (not(feature = "std" ))] |
27 | /// # use core::error::Error; |
28 | /// |
29 | /// # fn run() -> Result<(), Box<dyn Error>> { |
30 | /// let mut url = Url::parse("mailto:me@example.com" )?; |
31 | /// assert!(url.path_segments_mut().is_err()); |
32 | /// |
33 | /// let mut url = Url::parse("http://example.net/foo/index.html" )?; |
34 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
35 | /// .pop().push("img" ).push("2/100%.png" ); |
36 | /// assert_eq!(url.as_str(), "http://example.net/foo/img/2%2F100%25.png" ); |
37 | /// # Ok(()) |
38 | /// # } |
39 | /// # run().unwrap(); |
40 | /// ``` |
41 | #[derive (Debug)] |
42 | pub struct PathSegmentsMut<'a> { |
43 | url: &'a mut Url, |
44 | after_first_slash: usize, |
45 | after_path: String, |
46 | old_after_path_position: u32, |
47 | } |
48 | |
49 | // Not re-exported outside the crate |
50 | pub fn new(url: &mut Url) -> PathSegmentsMut<'_> { |
51 | let after_path: String = url.take_after_path(); |
52 | let old_after_path_position: u32 = to_u32(url.serialization.len()).unwrap(); |
53 | // Special urls always have a non empty path |
54 | if SchemeType::from(url.scheme()).is_special() { |
55 | debug_assert!(url.byte_at(url.path_start) == b'/' ); |
56 | } else { |
57 | debug_assert!( |
58 | url.serialization.len() == url.path_start as usize |
59 | || url.byte_at(url.path_start) == b'/' |
60 | ); |
61 | } |
62 | PathSegmentsMut { |
63 | after_first_slash: url.path_start as usize + "/" .len(), |
64 | url, |
65 | old_after_path_position, |
66 | after_path, |
67 | } |
68 | } |
69 | |
70 | impl<'a> Drop for PathSegmentsMut<'a> { |
71 | fn drop(&mut self) { |
72 | self.url |
73 | .restore_after_path(self.old_after_path_position, &self.after_path) |
74 | } |
75 | } |
76 | |
77 | impl<'a> PathSegmentsMut<'a> { |
78 | /// Remove all segments in the path, leaving the minimal `url.path() == "/"`. |
79 | /// |
80 | /// Returns `&mut Self` so that method calls can be chained. |
81 | /// |
82 | /// Example: |
83 | /// |
84 | /// ```rust |
85 | /// use url::Url; |
86 | /// |
87 | /// # #[cfg (feature = "std" )] |
88 | /// # use std::error::Error; |
89 | /// # #[cfg (not(feature = "std" ))] |
90 | /// # use core::error::Error; |
91 | /// |
92 | /// # fn run() -> Result<(), Box<dyn Error>> { |
93 | /// let mut url = Url::parse("https://github.com/servo/rust-url/" )?; |
94 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
95 | /// .clear().push("logout" ); |
96 | /// assert_eq!(url.as_str(), "https://github.com/logout" ); |
97 | /// # Ok(()) |
98 | /// # } |
99 | /// # run().unwrap(); |
100 | /// ``` |
101 | pub fn clear(&mut self) -> &mut Self { |
102 | self.url.serialization.truncate(self.after_first_slash); |
103 | self |
104 | } |
105 | |
106 | /// Remove the last segment of this URL’s path if it is empty, |
107 | /// except if these was only one segment to begin with. |
108 | /// |
109 | /// In other words, remove one path trailing slash, if any, |
110 | /// unless it is also the initial slash (so this does nothing if `url.path() == "/")`. |
111 | /// |
112 | /// Returns `&mut Self` so that method calls can be chained. |
113 | /// |
114 | /// Example: |
115 | /// |
116 | /// ```rust |
117 | /// use url::Url; |
118 | /// |
119 | /// # #[cfg (feature = "std" )] |
120 | /// # use std::error::Error; |
121 | /// # #[cfg (not(feature = "std" ))] |
122 | /// # use core::error::Error; |
123 | /// |
124 | /// # fn run() -> Result<(), Box<dyn Error>> { |
125 | /// let mut url = Url::parse("https://github.com/servo/rust-url/" )?; |
126 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
127 | /// .push("pulls" ); |
128 | /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url//pulls" ); |
129 | /// |
130 | /// let mut url = Url::parse("https://github.com/servo/rust-url/" )?; |
131 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
132 | /// .pop_if_empty().push("pulls" ); |
133 | /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/pulls" ); |
134 | /// # Ok(()) |
135 | /// # } |
136 | /// # run().unwrap(); |
137 | /// ``` |
138 | pub fn pop_if_empty(&mut self) -> &mut Self { |
139 | if self.after_first_slash >= self.url.serialization.len() { |
140 | return self; |
141 | } |
142 | if self.url.serialization[self.after_first_slash..].ends_with('/' ) { |
143 | self.url.serialization.pop(); |
144 | } |
145 | self |
146 | } |
147 | |
148 | /// Remove the last segment of this URL’s path. |
149 | /// |
150 | /// If the path only has one segment, make it empty such that `url.path() == "/"`. |
151 | /// |
152 | /// Returns `&mut Self` so that method calls can be chained. |
153 | pub fn pop(&mut self) -> &mut Self { |
154 | if self.after_first_slash >= self.url.serialization.len() { |
155 | return self; |
156 | } |
157 | let last_slash = self.url.serialization[self.after_first_slash..] |
158 | .rfind('/' ) |
159 | .unwrap_or(0); |
160 | self.url |
161 | .serialization |
162 | .truncate(self.after_first_slash + last_slash); |
163 | self |
164 | } |
165 | |
166 | /// Append the given segment at the end of this URL’s path. |
167 | /// |
168 | /// See the documentation for `.extend()`. |
169 | /// |
170 | /// Returns `&mut Self` so that method calls can be chained. |
171 | pub fn push(&mut self, segment: &str) -> &mut Self { |
172 | self.extend(Some(segment)) |
173 | } |
174 | |
175 | /// Append each segment from the given iterator at the end of this URL’s path. |
176 | /// |
177 | /// Each segment is percent-encoded like in `Url::parse` or `Url::join`, |
178 | /// except that `%` and `/` characters are also encoded (to `%25` and `%2F`). |
179 | /// This is unlike `Url::parse` where `%` is left as-is in case some of the input |
180 | /// is already percent-encoded, and `/` denotes a path segment separator.) |
181 | /// |
182 | /// Note that, in addition to slashes between new segments, |
183 | /// this always adds a slash between the existing path and the new segments |
184 | /// *except* if the existing path is `"/"`. |
185 | /// If the previous last segment was empty (if the path had a trailing slash) |
186 | /// the path after `.extend()` will contain two consecutive slashes. |
187 | /// If that is undesired, call `.pop_if_empty()` first. |
188 | /// |
189 | /// To obtain a behavior similar to `Url::join`, call `.pop()` unconditionally first. |
190 | /// |
191 | /// Returns `&mut Self` so that method calls can be chained. |
192 | /// |
193 | /// Example: |
194 | /// |
195 | /// ```rust |
196 | /// use url::Url; |
197 | /// |
198 | /// # #[cfg (feature = "std" )] |
199 | /// # use std::error::Error; |
200 | /// # #[cfg (not(feature = "std" ))] |
201 | /// # use core::error::Error; |
202 | /// |
203 | /// # fn run() -> Result<(), Box<dyn Error>> { |
204 | /// let mut url = Url::parse("https://github.com/" )?; |
205 | /// let org = "servo" ; |
206 | /// let repo = "rust-url" ; |
207 | /// let issue_number = "188" ; |
208 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
209 | /// .extend(&[org, repo, "issues" , issue_number]); |
210 | /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/issues/188" ); |
211 | /// # Ok(()) |
212 | /// # } |
213 | /// # run().unwrap(); |
214 | /// ``` |
215 | /// |
216 | /// In order to make sure that parsing the serialization of an URL gives the same URL, |
217 | /// a segment is ignored if it is `"."` or `".."`: |
218 | /// |
219 | /// ```rust |
220 | /// use url::Url; |
221 | /// |
222 | /// # #[cfg (feature = "std" )] |
223 | /// # use std::error::Error; |
224 | /// # #[cfg (not(feature = "std" ))] |
225 | /// # use core::error::Error; |
226 | /// |
227 | /// # fn run() -> Result<(), Box<dyn Error>> { |
228 | /// let mut url = Url::parse("https://github.com/servo" )?; |
229 | /// url.path_segments_mut().map_err(|_| "cannot be base" )? |
230 | /// .extend(&[".." , "rust-url" , "." , "pulls" ]); |
231 | /// assert_eq!(url.as_str(), "https://github.com/servo/rust-url/pulls" ); |
232 | /// # Ok(()) |
233 | /// # } |
234 | /// # run().unwrap(); |
235 | /// ``` |
236 | pub fn extend<I>(&mut self, segments: I) -> &mut Self |
237 | where |
238 | I: IntoIterator, |
239 | I::Item: AsRef<str>, |
240 | { |
241 | let scheme_type = SchemeType::from(self.url.scheme()); |
242 | let path_start = self.url.path_start as usize; |
243 | self.url.mutate(|parser| { |
244 | parser.context = parser::Context::PathSegmentSetter; |
245 | for segment in segments { |
246 | let segment = segment.as_ref(); |
247 | if matches!(segment, "." | ".." ) { |
248 | continue; |
249 | } |
250 | if parser.serialization.len() > path_start + 1 |
251 | // Non special url's path might still be empty |
252 | || parser.serialization.len() == path_start |
253 | { |
254 | parser.serialization.push('/' ); |
255 | } |
256 | let mut has_host = true; // FIXME account for this? |
257 | parser.parse_path( |
258 | scheme_type, |
259 | &mut has_host, |
260 | path_start, |
261 | parser::Input::new_no_trim(segment), |
262 | ); |
263 | } |
264 | }); |
265 | self |
266 | } |
267 | } |
268 | |