]>
Commit | Line | Data |
---|---|---|
2c00a5a8 XL |
1 | #![allow(missing_docs)] // FIXME: Document this |
2 | ||
ea8adc8c | 3 | pub mod fs; |
2c00a5a8 | 4 | mod string; |
f035d41b | 5 | pub(crate) mod toml_ext; |
dc9dc135 | 6 | use crate::errors::Error; |
83c7162d | 7 | use regex::Regex; |
ea8adc8c | 8 | |
f035d41b | 9 | use pulldown_cmark::{html, CodeBlockKind, CowStr, Event, Options, Parser, Tag}; |
83c7162d | 10 | |
ea8adc8c | 11 | use std::borrow::Cow; |
416331ca XL |
12 | use std::fmt::Write; |
13 | use std::path::Path; | |
ea8adc8c | 14 | |
e74abb32 XL |
15 | pub use self::string::{ |
16 | take_anchored_lines, take_lines, take_rustdoc_include_anchored_lines, | |
17 | take_rustdoc_include_lines, | |
18 | }; | |
ea8adc8c | 19 | |
83c7162d | 20 | /// Replaces multiple consecutive whitespace characters with a single space character. |
dc9dc135 | 21 | pub fn collapse_whitespace(text: &str) -> Cow<'_, str> { |
83c7162d XL |
22 | lazy_static! { |
23 | static ref RE: Regex = Regex::new(r"\s\s+").unwrap(); | |
24 | } | |
25 | RE.replace_all(text, " ") | |
26 | } | |
27 | ||
9fa01778 XL |
28 | /// Convert the given string to a valid HTML element ID. |
29 | /// The only restriction is that the ID must not contain any ASCII whitespace. | |
83c7162d | 30 | pub fn normalize_id(content: &str) -> String { |
9fa01778 | 31 | content |
83c7162d XL |
32 | .chars() |
33 | .filter_map(|ch| { | |
34 | if ch.is_alphanumeric() || ch == '_' || ch == '-' { | |
35 | Some(ch.to_ascii_lowercase()) | |
36 | } else if ch.is_whitespace() { | |
37 | Some('-') | |
38 | } else { | |
39 | None | |
40 | } | |
dc9dc135 XL |
41 | }) |
42 | .collect::<String>() | |
83c7162d XL |
43 | } |
44 | ||
45 | /// Generate an ID for use with anchors which is derived from a "normalised" | |
46 | /// string. | |
47 | pub fn id_from_content(content: &str) -> String { | |
48 | let mut content = content.to_string(); | |
49 | ||
50 | // Skip any tags or html-encoded stuff | |
9fa01778 XL |
51 | const REPL_SUB: &[&str] = &[ |
52 | "<em>", | |
53 | "</em>", | |
54 | "<code>", | |
55 | "</code>", | |
56 | "<strong>", | |
57 | "</strong>", | |
58 | "<", | |
59 | ">", | |
60 | "&", | |
61 | "'", | |
62 | """, | |
63 | ]; | |
83c7162d XL |
64 | for sub in REPL_SUB { |
65 | content = content.replace(sub, ""); | |
66 | } | |
67 | ||
68 | // Remove spaces and hashes indicating a header | |
dc9dc135 | 69 | let trimmed = content.trim().trim_start_matches('#').trim(); |
83c7162d XL |
70 | |
71 | normalize_id(trimmed) | |
72 | } | |
73 | ||
416331ca XL |
74 | /// Fix links to the correct location. |
75 | /// | |
76 | /// This adjusts links, such as turning `.md` extensions to `.html`. | |
77 | /// | |
78 | /// `path` is the path to the page being rendered relative to the root of the | |
79 | /// book. This is used for the `print.html` page so that links on the print | |
80 | /// page go to the original location. Normal page rendering sets `path` to | |
81 | /// None. Ideally, print page links would link to anchors on the print page, | |
82 | /// but that is very difficult. | |
83 | fn adjust_links<'a>(event: Event<'a>, path: Option<&Path>) -> Event<'a> { | |
9fa01778 | 84 | lazy_static! { |
dc9dc135 | 85 | static ref SCHEME_LINK: Regex = Regex::new(r"^[a-z][a-z0-9+.-]*:").unwrap(); |
9fa01778 XL |
86 | static ref MD_LINK: Regex = Regex::new(r"(?P<link>.*)\.md(?P<anchor>#.*)?").unwrap(); |
87 | } | |
88 | ||
416331ca XL |
89 | fn fix<'a>(dest: CowStr<'a>, path: Option<&Path>) -> CowStr<'a> { |
90 | if dest.starts_with('#') { | |
91 | // Fragment-only link. | |
92 | if let Some(path) = path { | |
93 | let mut base = path.display().to_string(); | |
94 | if base.ends_with(".md") { | |
95 | base.replace_range(base.len() - 3.., ".html"); | |
96 | } | |
97 | return format!("{}{}", base, dest).into(); | |
98 | } else { | |
99 | return dest; | |
100 | } | |
101 | } | |
dc9dc135 XL |
102 | // Don't modify links with schemes like `https`. |
103 | if !SCHEME_LINK.is_match(&dest) { | |
104 | // This is a relative link, adjust it as necessary. | |
105 | let mut fixed_link = String::new(); | |
416331ca XL |
106 | if let Some(path) = path { |
107 | let base = path | |
108 | .parent() | |
109 | .expect("path can't be empty") | |
110 | .to_str() | |
111 | .expect("utf-8 paths only"); | |
112 | if !base.is_empty() { | |
113 | write!(fixed_link, "{}/", base).unwrap(); | |
114 | } | |
dc9dc135 | 115 | } |
9fa01778 | 116 | |
dc9dc135 XL |
117 | if let Some(caps) = MD_LINK.captures(&dest) { |
118 | fixed_link.push_str(&caps["link"]); | |
119 | fixed_link.push_str(".html"); | |
120 | if let Some(anchor) = caps.name("anchor") { | |
121 | fixed_link.push_str(anchor.as_str()); | |
9fa01778 | 122 | } |
dc9dc135 XL |
123 | } else { |
124 | fixed_link.push_str(&dest); | |
125 | }; | |
126 | return CowStr::from(fixed_link); | |
127 | } | |
128 | dest | |
129 | } | |
9fa01778 | 130 | |
416331ca XL |
131 | fn fix_html<'a>(html: CowStr<'a>, path: Option<&Path>) -> CowStr<'a> { |
132 | // This is a terrible hack, but should be reasonably reliable. Nobody | |
133 | // should ever parse a tag with a regex. However, there isn't anything | |
134 | // in Rust that I know of that is suitable for handling partial html | |
135 | // fragments like those generated by pulldown_cmark. | |
136 | // | |
137 | // There are dozens of HTML tags/attributes that contain paths, so | |
138 | // feel free to add more tags if desired; these are the only ones I | |
139 | // care about right now. | |
140 | lazy_static! { | |
141 | static ref HTML_LINK: Regex = | |
142 | Regex::new(r#"(<(?:a|img) [^>]*?(?:src|href)=")([^"]+?)""#).unwrap(); | |
143 | } | |
144 | ||
145 | HTML_LINK | |
146 | .replace_all(&html, |caps: ®ex::Captures<'_>| { | |
147 | let fixed = fix(caps[2].into(), path); | |
148 | format!("{}{}\"", &caps[1], fixed) | |
149 | }) | |
150 | .into_owned() | |
151 | .into() | |
152 | } | |
153 | ||
dc9dc135 XL |
154 | match event { |
155 | Event::Start(Tag::Link(link_type, dest, title)) => { | |
416331ca | 156 | Event::Start(Tag::Link(link_type, fix(dest, path), title)) |
dc9dc135 XL |
157 | } |
158 | Event::Start(Tag::Image(link_type, dest, title)) => { | |
416331ca | 159 | Event::Start(Tag::Image(link_type, fix(dest, path), title)) |
9fa01778 | 160 | } |
416331ca | 161 | Event::Html(html) => Event::Html(fix_html(html, path)), |
9fa01778 XL |
162 | _ => event, |
163 | } | |
164 | } | |
165 | ||
2c00a5a8 | 166 | /// Wrapper around the pulldown-cmark parser for rendering markdown to HTML. |
ea8adc8c | 167 | pub fn render_markdown(text: &str, curly_quotes: bool) -> String { |
416331ca | 168 | render_markdown_with_path(text, curly_quotes, None) |
9fa01778 XL |
169 | } |
170 | ||
dc9dc135 | 171 | pub fn new_cmark_parser(text: &str) -> Parser<'_> { |
ea8adc8c | 172 | let mut opts = Options::empty(); |
dc9dc135 XL |
173 | opts.insert(Options::ENABLE_TABLES); |
174 | opts.insert(Options::ENABLE_FOOTNOTES); | |
175 | opts.insert(Options::ENABLE_STRIKETHROUGH); | |
176 | opts.insert(Options::ENABLE_TASKLISTS); | |
177 | Parser::new_ext(text, opts) | |
178 | } | |
ea8adc8c | 179 | |
416331ca | 180 | pub fn render_markdown_with_path(text: &str, curly_quotes: bool, path: Option<&Path>) -> String { |
dc9dc135 XL |
181 | let mut s = String::with_capacity(text.len() * 3 / 2); |
182 | let p = new_cmark_parser(text); | |
ea8adc8c | 183 | let mut converter = EventQuoteConverter::new(curly_quotes); |
9fa01778 XL |
184 | let events = p |
185 | .map(clean_codeblock_headers) | |
416331ca | 186 | .map(|event| adjust_links(event, path)) |
9fa01778 | 187 | .map(|event| converter.convert(event)); |
ea8adc8c XL |
188 | |
189 | html::push_html(&mut s, events); | |
190 | s | |
191 | } | |
192 | ||
193 | struct EventQuoteConverter { | |
194 | enabled: bool, | |
195 | convert_text: bool, | |
196 | } | |
197 | ||
198 | impl EventQuoteConverter { | |
199 | fn new(enabled: bool) -> Self { | |
2c00a5a8 | 200 | EventQuoteConverter { |
9fa01778 | 201 | enabled, |
2c00a5a8 XL |
202 | convert_text: true, |
203 | } | |
ea8adc8c XL |
204 | } |
205 | ||
206 | fn convert<'a>(&mut self, event: Event<'a>) -> Event<'a> { | |
207 | if !self.enabled { | |
208 | return event; | |
209 | } | |
210 | ||
211 | match event { | |
dc9dc135 | 212 | Event::Start(Tag::CodeBlock(_)) => { |
ea8adc8c XL |
213 | self.convert_text = false; |
214 | event | |
2c00a5a8 | 215 | } |
dc9dc135 | 216 | Event::End(Tag::CodeBlock(_)) => { |
ea8adc8c XL |
217 | self.convert_text = true; |
218 | event | |
2c00a5a8 XL |
219 | } |
220 | Event::Text(ref text) if self.convert_text => { | |
dc9dc135 | 221 | Event::Text(CowStr::from(convert_quotes_to_curly(text))) |
2c00a5a8 | 222 | } |
ea8adc8c XL |
223 | _ => event, |
224 | } | |
225 | } | |
226 | } | |
227 | ||
dc9dc135 | 228 | fn clean_codeblock_headers(event: Event<'_>) -> Event<'_> { |
ea8adc8c | 229 | match event { |
f035d41b | 230 | Event::Start(Tag::CodeBlock(CodeBlockKind::Fenced(ref info))) => { |
2c00a5a8 | 231 | let info: String = info.chars().filter(|ch| !ch.is_whitespace()).collect(); |
ea8adc8c | 232 | |
f035d41b | 233 | Event::Start(Tag::CodeBlock(CodeBlockKind::Fenced(CowStr::from(info)))) |
2c00a5a8 | 234 | } |
ea8adc8c XL |
235 | _ => event, |
236 | } | |
237 | } | |
238 | ||
ea8adc8c XL |
239 | fn convert_quotes_to_curly(original_text: &str) -> String { |
240 | // We'll consider the start to be "whitespace". | |
241 | let mut preceded_by_whitespace = true; | |
242 | ||
9fa01778 XL |
243 | original_text |
244 | .chars() | |
245 | .map(|original_char| { | |
246 | let converted_char = match original_char { | |
247 | '\'' => { | |
248 | if preceded_by_whitespace { | |
249 | '‘' | |
250 | } else { | |
251 | '’' | |
252 | } | |
2c00a5a8 | 253 | } |
9fa01778 XL |
254 | '"' => { |
255 | if preceded_by_whitespace { | |
256 | '“' | |
257 | } else { | |
258 | '”' | |
259 | } | |
2c00a5a8 | 260 | } |
9fa01778 XL |
261 | _ => original_char, |
262 | }; | |
2c00a5a8 | 263 | |
9fa01778 | 264 | preceded_by_whitespace = original_char.is_whitespace(); |
2c00a5a8 | 265 | |
9fa01778 | 266 | converted_char |
dc9dc135 XL |
267 | }) |
268 | .collect() | |
2c00a5a8 | 269 | } |
ea8adc8c | 270 | |
2c00a5a8 XL |
271 | /// Prints a "backtrace" of some `Error`. |
272 | pub fn log_backtrace(e: &Error) { | |
273 | error!("Error: {}", e); | |
ea8adc8c | 274 | |
f035d41b | 275 | for cause in e.chain().skip(1) { |
2c00a5a8 XL |
276 | error!("\tCaused By: {}", cause); |
277 | } | |
ea8adc8c XL |
278 | } |
279 | ||
280 | #[cfg(test)] | |
281 | mod tests { | |
282 | mod render_markdown { | |
283 | use super::super::render_markdown; | |
284 | ||
9fa01778 XL |
285 | #[test] |
286 | fn preserves_external_links() { | |
287 | assert_eq!( | |
288 | render_markdown("[example](https://www.rust-lang.org/)", false), | |
289 | "<p><a href=\"https://www.rust-lang.org/\">example</a></p>\n" | |
290 | ); | |
291 | } | |
292 | ||
293 | #[test] | |
294 | fn it_can_adjust_markdown_links() { | |
295 | assert_eq!( | |
296 | render_markdown("[example](example.md)", false), | |
297 | "<p><a href=\"example.html\">example</a></p>\n" | |
298 | ); | |
299 | assert_eq!( | |
300 | render_markdown("[example_anchor](example.md#anchor)", false), | |
301 | "<p><a href=\"example.html#anchor\">example_anchor</a></p>\n" | |
302 | ); | |
303 | ||
304 | // this anchor contains 'md' inside of it | |
305 | assert_eq!( | |
306 | render_markdown("[phantom data](foo.html#phantomdata)", false), | |
307 | "<p><a href=\"foo.html#phantomdata\">phantom data</a></p>\n" | |
308 | ); | |
309 | } | |
310 | ||
ea8adc8c XL |
311 | #[test] |
312 | fn it_can_keep_quotes_straight() { | |
313 | assert_eq!(render_markdown("'one'", false), "<p>'one'</p>\n"); | |
314 | } | |
315 | ||
316 | #[test] | |
317 | fn it_can_make_quotes_curly_except_when_they_are_in_code() { | |
318 | let input = r#" | |
319 | 'one' | |
320 | ``` | |
321 | 'two' | |
322 | ``` | |
323 | `'three'` 'four'"#; | |
324 | let expected = r#"<p>‘one’</p> | |
325 | <pre><code>'two' | |
326 | </code></pre> | |
327 | <p><code>'three'</code> ‘four’</p> | |
328 | "#; | |
329 | assert_eq!(render_markdown(input, true), expected); | |
330 | } | |
331 | ||
332 | #[test] | |
333 | fn whitespace_outside_of_codeblock_header_is_preserved() { | |
334 | let input = r#" | |
335 | some text with spaces | |
336 | ```rust | |
337 | fn main() { | |
338 | // code inside is unchanged | |
339 | } | |
340 | ``` | |
341 | more text with spaces | |
342 | "#; | |
343 | ||
344 | let expected = r#"<p>some text with spaces</p> | |
345 | <pre><code class="language-rust">fn main() { | |
346 | // code inside is unchanged | |
347 | } | |
348 | </code></pre> | |
349 | <p>more text with spaces</p> | |
350 | "#; | |
351 | assert_eq!(render_markdown(input, false), expected); | |
352 | assert_eq!(render_markdown(input, true), expected); | |
353 | } | |
354 | ||
355 | #[test] | |
356 | fn rust_code_block_properties_are_passed_as_space_delimited_class() { | |
357 | let input = r#" | |
358 | ```rust,no_run,should_panic,property_3 | |
359 | ``` | |
360 | "#; | |
361 | ||
f9f354fc | 362 | let expected = r#"<pre><code class="language-rust,no_run,should_panic,property_3"></code></pre> |
ea8adc8c XL |
363 | "#; |
364 | assert_eq!(render_markdown(input, false), expected); | |
365 | assert_eq!(render_markdown(input, true), expected); | |
366 | } | |
367 | ||
368 | #[test] | |
369 | fn rust_code_block_properties_with_whitespace_are_passed_as_space_delimited_class() { | |
370 | let input = r#" | |
371 | ```rust, no_run,,,should_panic , ,property_3 | |
372 | ``` | |
373 | "#; | |
374 | ||
f9f354fc | 375 | let expected = r#"<pre><code class="language-rust,no_run,,,should_panic,,property_3"></code></pre> |
ea8adc8c XL |
376 | "#; |
377 | assert_eq!(render_markdown(input, false), expected); | |
378 | assert_eq!(render_markdown(input, true), expected); | |
379 | } | |
380 | ||
381 | #[test] | |
382 | fn rust_code_block_without_properties_has_proper_html_class() { | |
383 | let input = r#" | |
2c00a5a8 | 384 | ```rust |
ea8adc8c XL |
385 | ``` |
386 | "#; | |
387 | ||
388 | let expected = r#"<pre><code class="language-rust"></code></pre> | |
389 | "#; | |
390 | assert_eq!(render_markdown(input, false), expected); | |
391 | assert_eq!(render_markdown(input, true), expected); | |
392 | ||
393 | let input = r#" | |
394 | ```rust | |
395 | ``` | |
396 | "#; | |
397 | assert_eq!(render_markdown(input, false), expected); | |
398 | assert_eq!(render_markdown(input, true), expected); | |
ea8adc8c XL |
399 | } |
400 | } | |
401 | ||
83c7162d XL |
402 | mod html_munging { |
403 | use super::super::{id_from_content, normalize_id}; | |
404 | ||
405 | #[test] | |
406 | fn it_generates_anchors() { | |
9fa01778 XL |
407 | assert_eq!( |
408 | id_from_content("## Method-call expressions"), | |
409 | "method-call-expressions" | |
410 | ); | |
411 | assert_eq!(id_from_content("## **Bold** title"), "bold-title"); | |
412 | assert_eq!(id_from_content("## `Code` title"), "code-title"); | |
413 | } | |
414 | ||
415 | #[test] | |
416 | fn it_generates_anchors_from_non_ascii_initial() { | |
417 | assert_eq!( | |
418 | id_from_content("## `--passes`: add more rustdoc passes"), | |
419 | "--passes-add-more-rustdoc-passes" | |
420 | ); | |
421 | assert_eq!( | |
422 | id_from_content("## 中文標題 CJK title"), | |
423 | "中文標題-cjk-title" | |
424 | ); | |
425 | assert_eq!(id_from_content("## Über"), "Über"); | |
83c7162d XL |
426 | } |
427 | ||
428 | #[test] | |
429 | fn it_normalizes_ids() { | |
9fa01778 XL |
430 | assert_eq!( |
431 | normalize_id("`--passes`: add more rustdoc passes"), | |
432 | "--passes-add-more-rustdoc-passes" | |
433 | ); | |
434 | assert_eq!( | |
435 | normalize_id("Method-call 🐙 expressions \u{1f47c}"), | |
436 | "method-call--expressions-" | |
437 | ); | |
438 | assert_eq!(normalize_id("_-_12345"), "_-_12345"); | |
439 | assert_eq!(normalize_id("12345"), "12345"); | |
440 | assert_eq!(normalize_id("中文"), "中文"); | |
441 | assert_eq!(normalize_id("にほんご"), "にほんご"); | |
442 | assert_eq!(normalize_id("한국어"), "한국어"); | |
83c7162d XL |
443 | assert_eq!(normalize_id(""), ""); |
444 | } | |
445 | } | |
446 | ||
ea8adc8c XL |
447 | mod convert_quotes_to_curly { |
448 | use super::super::convert_quotes_to_curly; | |
449 | ||
450 | #[test] | |
451 | fn it_converts_single_quotes() { | |
e74abb32 | 452 | assert_eq!(convert_quotes_to_curly("'one', 'two'"), "‘one’, ‘two’"); |
ea8adc8c XL |
453 | } |
454 | ||
455 | #[test] | |
456 | fn it_converts_double_quotes() { | |
e74abb32 | 457 | assert_eq!(convert_quotes_to_curly(r#""one", "two""#), "“one”, “two”"); |
ea8adc8c XL |
458 | } |
459 | ||
460 | #[test] | |
461 | fn it_treats_tab_as_whitespace() { | |
462 | assert_eq!(convert_quotes_to_curly("\t'one'"), "\t‘one’"); | |
463 | } | |
464 | } | |
465 | } |