selma 0.0.2-arm64-darwin

Sign up to get free protection for your applications and to get access to all the features.
@@ -0,0 +1,578 @@
1
+ use std::{borrow::BorrowMut, cell::RefMut, collections::HashMap};
2
+
3
+ use lol_html::html_content::{Comment, ContentType, Doctype, Element, EndTag};
4
+ use magnus::{
5
+ class, exception, function, method, scan_args, Error, Module, Object, RArray, RHash, RModule,
6
+ Value,
7
+ };
8
+
9
+ use crate::tags::Tag;
10
+
11
+ #[derive(Clone, Debug)]
12
+ struct ElementSanitizer {
13
+ allowed_attrs: Vec<String>,
14
+ required_attrs: Vec<String>,
15
+ allowed_classes: Vec<String>,
16
+ protocol_sanitizers: HashMap<String, Vec<String>>,
17
+ }
18
+
19
+ #[derive(Clone, Debug)]
20
+ pub struct Sanitizer {
21
+ flags: [u8; Tag::TAG_COUNT],
22
+ allowed_attrs: Vec<String>,
23
+ allowed_classes: Vec<String>,
24
+ element_sanitizers: HashMap<String, ElementSanitizer>,
25
+
26
+ pub escape_tagfilter: bool,
27
+ pub allow_comments: bool,
28
+ pub allow_doctype: bool,
29
+ config: RHash,
30
+ }
31
+
32
+ #[derive(Clone, Debug)]
33
+ #[magnus::wrap(class = "Selma::Sanitizer")]
34
+ pub struct SelmaSanitizer(std::cell::RefCell<Sanitizer>);
35
+
36
+ impl SelmaSanitizer {
37
+ const SELMA_SANITIZER_ALLOW: u8 = (1 << 0);
38
+ const SELMA_SANITIZER_ESCAPE_TAGFILTER: u8 = (1 << 1);
39
+ const SELMA_SANITIZER_REMOVE_CONTENTS: u8 = (1 << 2);
40
+ const SELMA_SANITIZER_WRAP_WHITESPACE: u8 = (1 << 3);
41
+
42
+ pub fn new(arguments: &[Value]) -> Result<Self, Error> {
43
+ let args = scan_args::scan_args::<(), (Option<RHash>,), (), (), (), ()>(arguments)?;
44
+ let (opt_config,): (Option<RHash>,) = args.optional;
45
+
46
+ let config = match opt_config {
47
+ Some(config) => config,
48
+ // TODO: this seems like a hack to fix?
49
+ None => magnus::eval::<RHash>(r#"Selma::Sanitizer::Config::DEFAULT"#).unwrap(),
50
+ };
51
+
52
+ let mut element_sanitizers = HashMap::new();
53
+ Tag::html_tags().iter().for_each(|html_tag| {
54
+ let es = ElementSanitizer {
55
+ allowed_attrs: vec![],
56
+ allowed_classes: vec![],
57
+ required_attrs: vec![],
58
+
59
+ protocol_sanitizers: HashMap::new(),
60
+ };
61
+ element_sanitizers.insert(Tag::element_name_from_enum(html_tag).to_string(), es);
62
+ });
63
+
64
+ Ok(Self(std::cell::RefCell::new(Sanitizer {
65
+ flags: [0; Tag::TAG_COUNT],
66
+ allowed_attrs: vec![],
67
+ allowed_classes: vec![],
68
+ element_sanitizers,
69
+
70
+ escape_tagfilter: true,
71
+ allow_comments: false,
72
+ allow_doctype: true,
73
+ config,
74
+ })))
75
+ }
76
+
77
+ fn get_config(&self) -> Result<RHash, Error> {
78
+ let binding = self.0.borrow();
79
+
80
+ Ok(binding.config)
81
+ }
82
+
83
+ /// Toggle a sanitizer option on or off.
84
+ fn set_flag(&self, tag_name: String, flag: u8, set: bool) {
85
+ let tag = Tag::tag_from_tag_name(tag_name.as_str());
86
+ if set {
87
+ self.0.borrow_mut().flags[tag.index] |= flag;
88
+ } else {
89
+ self.0.borrow_mut().flags[tag.index] &= !flag;
90
+ }
91
+ }
92
+
93
+ /// Toggles all sanitization options on or off.
94
+ fn set_all_flags(&self, flag: u8, set: bool) {
95
+ if set {
96
+ Tag::html_tags().iter().enumerate().for_each(|(iter, _)| {
97
+ self.0.borrow_mut().flags[iter] |= flag;
98
+ });
99
+ } else {
100
+ Tag::html_tags().iter().enumerate().for_each(|(iter, _)| {
101
+ self.0.borrow_mut().flags[iter] &= flag;
102
+ });
103
+ }
104
+ }
105
+
106
+ /// Whether or not to keep dangerous HTML tags.
107
+ fn set_escape_tagfilter(&self, allow: bool) -> bool {
108
+ self.0.borrow_mut().escape_tagfilter = allow;
109
+ allow
110
+ }
111
+
112
+ pub fn escape_tagfilter(&self, e: &mut Element) -> bool {
113
+ if self.0.borrow().escape_tagfilter {
114
+ let tag = Tag::tag_from_element(e);
115
+ if Tag::is_tag_escapeworthy(tag) {
116
+ e.remove();
117
+ return true;
118
+ }
119
+ }
120
+
121
+ false
122
+ }
123
+
124
+ pub fn get_escape_tagfilter(&self) -> bool {
125
+ self.0.borrow().escape_tagfilter
126
+ }
127
+
128
+ /// Whether or not to keep HTML comments.
129
+ fn set_allow_comments(&self, allow: bool) -> bool {
130
+ self.0.borrow_mut().allow_comments = allow;
131
+ allow
132
+ }
133
+
134
+ pub fn get_allow_comments(&self) -> bool {
135
+ self.0.borrow().allow_comments
136
+ }
137
+
138
+ pub fn remove_comment(&self, c: &mut Comment) {
139
+ c.remove();
140
+ }
141
+
142
+ /// Whether or not to keep HTML doctype.
143
+ fn set_allow_doctype(&self, allow: bool) -> bool {
144
+ self.0.borrow_mut().allow_doctype = allow;
145
+ allow
146
+ }
147
+
148
+ /// Whether or not to keep HTML doctype.
149
+ pub fn get_allow_doctype(&self) -> bool {
150
+ self.0.borrow().allow_doctype
151
+ }
152
+
153
+ pub fn remove_doctype(&self, d: &mut Doctype) {
154
+ d.remove();
155
+ }
156
+
157
+ fn set_allowed_attribute(&self, eln: Value, attr_name: String, allow: bool) -> bool {
158
+ let mut binding = self.0.borrow_mut();
159
+
160
+ let element_name = eln.to_r_string().unwrap().to_string().unwrap();
161
+ if element_name == "all" {
162
+ let allowed_attrs = &mut binding.allowed_attrs;
163
+ Self::set_allowed(allowed_attrs, &attr_name, allow);
164
+ } else {
165
+ let element_sanitizer = Self::get_mut_element_sanitizer(&mut binding, &element_name);
166
+
167
+ element_sanitizer.allowed_attrs.push(attr_name);
168
+ }
169
+
170
+ allow
171
+ }
172
+
173
+ fn set_allowed_class(&self, element_name: String, class_name: String, allow: bool) -> bool {
174
+ let mut binding = self.0.borrow_mut();
175
+ if element_name == "all" {
176
+ let allowed_classes = &mut binding.allowed_classes;
177
+ Self::set_allowed(allowed_classes, &class_name, allow);
178
+ } else {
179
+ let element_sanitizer = Self::get_mut_element_sanitizer(&mut binding, &element_name);
180
+
181
+ let allowed_classes = element_sanitizer.allowed_classes.borrow_mut();
182
+ Self::set_allowed(allowed_classes, &class_name, allow)
183
+ }
184
+ allow
185
+ }
186
+
187
+ fn set_allowed_protocols(&self, element_name: String, attr_name: String, allow_list: RArray) {
188
+ let mut binding = self.0.borrow_mut();
189
+
190
+ let element_sanitizer = Self::get_mut_element_sanitizer(&mut binding, &element_name);
191
+
192
+ let protocol_sanitizers = element_sanitizer.protocol_sanitizers.borrow_mut();
193
+
194
+ for opt_allowed_protocol in allow_list.each() {
195
+ let allowed_protocol = opt_allowed_protocol.unwrap();
196
+ let protocol_list = protocol_sanitizers.get_mut(&attr_name);
197
+ if allowed_protocol.is_kind_of(class::string()) {
198
+ match protocol_list {
199
+ None => {
200
+ protocol_sanitizers
201
+ .insert(attr_name.to_string(), vec![allowed_protocol.to_string()]);
202
+ }
203
+ Some(protocol_list) => protocol_list.push(allowed_protocol.to_string()),
204
+ }
205
+ } else if allowed_protocol.is_kind_of(class::symbol())
206
+ && allowed_protocol.inspect() == ":relative"
207
+ {
208
+ match protocol_list {
209
+ None => {
210
+ protocol_sanitizers.insert(
211
+ attr_name.to_string(),
212
+ vec!["#".to_string(), "/".to_string()],
213
+ );
214
+ }
215
+ Some(protocol_list) => {
216
+ protocol_list.push("#".to_string());
217
+ protocol_list.push("/".to_string());
218
+ }
219
+ }
220
+ }
221
+ }
222
+ }
223
+
224
+ fn set_allowed(set: &mut Vec<String>, attr_name: &String, allow: bool) {
225
+ if allow {
226
+ set.push(attr_name.to_string());
227
+ } else if set.contains(attr_name) {
228
+ set.swap_remove(set.iter().position(|x| x == attr_name).unwrap());
229
+ }
230
+ }
231
+
232
+ pub fn sanitize_attributes(&self, element: &mut Element) {
233
+ let binding = self.0.borrow_mut();
234
+ let tag = Tag::tag_from_element(element);
235
+ let element_sanitizer = Self::get_element_sanitizer(&binding, &element.tag_name());
236
+
237
+ // FIXME: This is a hack to get around the fact that we can't borrow
238
+ let attribute_map: HashMap<String, String> = element
239
+ .attributes()
240
+ .iter()
241
+ .map(|a| (a.name(), a.value()))
242
+ .collect();
243
+
244
+ for (attr_name, attr_val) in attribute_map.iter() {
245
+ // you can actually embed <!-- ... --> inside
246
+ // an HTML tag to pass malicious data. If this is
247
+ // encountered, remove the entire element to be safe.
248
+ if attr_name.starts_with("<!--") {
249
+ Self::force_remove_element(self, element);
250
+ return;
251
+ }
252
+
253
+ // first, trim leading spaces and unescape any encodings
254
+ let trimmed = attr_val.trim_start();
255
+ let x = escapist::unescape_html(trimmed.as_bytes());
256
+ let unescaped_attr_val = String::from_utf8_lossy(&x).to_string();
257
+
258
+ if !Self::should_keep_attribute(
259
+ &binding,
260
+ element,
261
+ element_sanitizer,
262
+ attr_name,
263
+ &unescaped_attr_val,
264
+ ) {
265
+ element.remove_attribute(attr_name);
266
+ } else {
267
+ // Prevent the use of `<meta>` elements that set a charset other than UTF-8,
268
+ // since output is always UTF-8.
269
+ if Tag::is_meta(tag) {
270
+ if attr_name == "charset" && unescaped_attr_val != "utf-8" {
271
+ element.set_attribute(attr_name, "utf-8");
272
+ }
273
+ } else if !unescaped_attr_val.is_empty() {
274
+ let mut buf = String::new();
275
+ // ...then, escape any special characters, for security
276
+ if attr_name == "href" {
277
+ // FIXME: gross--------------vvvv
278
+ escapist::escape_href(&mut buf, unescaped_attr_val.to_string().as_str());
279
+ } else {
280
+ escapist::escape_html(&mut buf, unescaped_attr_val.to_string().as_str());
281
+ };
282
+
283
+ element.set_attribute(attr_name, &buf);
284
+ }
285
+ }
286
+ }
287
+
288
+ let required = &element_sanitizer.required_attrs;
289
+ if required.contains(&"*".to_string()) {
290
+ return;
291
+ }
292
+ for attr in element.attributes().iter() {
293
+ let attr_name = &attr.name();
294
+ if required.contains(attr_name) {
295
+ return;
296
+ }
297
+ }
298
+ }
299
+
300
+ fn should_keep_attribute(
301
+ binding: &RefMut<Sanitizer>,
302
+ element: &mut Element,
303
+ element_sanitizer: &ElementSanitizer,
304
+ attr_name: &String,
305
+ attr_val: &String,
306
+ ) -> bool {
307
+ let mut allowed: bool = false;
308
+ let element_allowed_attrs = element_sanitizer.allowed_attrs.contains(attr_name);
309
+ let sanitizer_allowed_attrs = binding.allowed_attrs.contains(attr_name);
310
+
311
+ if element_allowed_attrs {
312
+ allowed = true;
313
+ }
314
+
315
+ if !allowed && sanitizer_allowed_attrs {
316
+ allowed = true;
317
+ }
318
+
319
+ if !allowed {
320
+ return false;
321
+ }
322
+
323
+ let protocol_sanitizer_values = element_sanitizer.protocol_sanitizers.get(attr_name);
324
+ match protocol_sanitizer_values {
325
+ None => {
326
+ // has a protocol, but no sanitization list
327
+ if !attr_val.is_empty() && Self::has_protocol(attr_val) {
328
+ return false;
329
+ }
330
+ }
331
+ Some(protocol_sanitizer_values) => {
332
+ if !attr_val.is_empty()
333
+ && !Self::has_allowed_protocol(protocol_sanitizer_values, attr_val)
334
+ {
335
+ return false;
336
+ }
337
+ }
338
+ }
339
+
340
+ if attr_name == "class"
341
+ && !Self::sanitize_class_attribute(
342
+ binding,
343
+ element,
344
+ element_sanitizer,
345
+ attr_name,
346
+ attr_val,
347
+ )
348
+ .unwrap()
349
+ {
350
+ return false;
351
+ }
352
+
353
+ true
354
+ }
355
+
356
+ fn has_protocol(attr_val: &str) -> bool {
357
+ attr_val.contains("://")
358
+ }
359
+
360
+ fn has_allowed_protocol(protocols_allowed: &[String], attr_val: &String) -> bool {
361
+ // FIXME: is there a more idiomatic way to do this?
362
+ let mut pos: usize = 0;
363
+ let mut chars = attr_val.chars();
364
+ let len = attr_val.len();
365
+
366
+ for (i, c) in attr_val.chars().enumerate() {
367
+ if c != ':' && c != '/' && c != '#' && pos + 1 < len {
368
+ pos = i + 1;
369
+ } else {
370
+ break;
371
+ }
372
+ }
373
+
374
+ let char = chars.nth(pos).unwrap();
375
+
376
+ if char == '/' {
377
+ return protocols_allowed.contains(&"/".to_string());
378
+ }
379
+
380
+ if char == '#' {
381
+ return protocols_allowed.contains(&"#".to_string());
382
+ }
383
+
384
+ // Allow protocol name to be case-insensitive
385
+ let protocol = attr_val[0..pos].to_lowercase();
386
+
387
+ protocols_allowed.contains(&protocol.to_lowercase())
388
+ }
389
+
390
+ fn sanitize_class_attribute(
391
+ binding: &RefMut<Sanitizer>,
392
+ element: &mut Element,
393
+ element_sanitizer: &ElementSanitizer,
394
+ attr_name: &str,
395
+ attr_val: &str,
396
+ ) -> Result<bool, Error> {
397
+ let allowed_global = &binding.allowed_classes;
398
+
399
+ let mut valid_classes: Vec<String> = vec![];
400
+
401
+ let allowed_local = &element_sanitizer.allowed_classes;
402
+
403
+ // No class filters, so everything goes through
404
+ if allowed_global.is_empty() && allowed_local.is_empty() {
405
+ return Ok(true);
406
+ }
407
+
408
+ let attr_value = attr_val.trim_start();
409
+ attr_value
410
+ .split_whitespace()
411
+ .map(|s| s.to_string())
412
+ .for_each(|class| {
413
+ if allowed_global.contains(&class) || allowed_local.contains(&class) {
414
+ valid_classes.push(class);
415
+ }
416
+ });
417
+
418
+ if valid_classes.is_empty() {
419
+ return Ok(false);
420
+ }
421
+
422
+ match element.set_attribute(attr_name, valid_classes.join(" ").as_str()) {
423
+ Ok(_) => Ok(true),
424
+ Err(err) => Err(Error::new(
425
+ exception::runtime_error(),
426
+ format!("AttributeNameError: {}", err),
427
+ )),
428
+ }
429
+ }
430
+
431
+ pub fn allow_element(&self, element: &mut Element) -> bool {
432
+ let tag = Tag::tag_from_element(element);
433
+ let flags: u8 = self.0.borrow().flags[tag.index];
434
+
435
+ (flags & Self::SELMA_SANITIZER_ALLOW) == 0
436
+ }
437
+
438
+ pub fn try_remove_element(&self, element: &mut Element) -> bool {
439
+ let tag = Tag::tag_from_element(element);
440
+ let flags: u8 = self.0.borrow().flags[tag.index];
441
+
442
+ let should_remove = !element.removed() && self.allow_element(element);
443
+
444
+ if should_remove {
445
+ if Tag::has_text_content(tag) {
446
+ Self::remove_element(
447
+ element,
448
+ tag.self_closing,
449
+ Self::SELMA_SANITIZER_REMOVE_CONTENTS,
450
+ );
451
+ } else {
452
+ Self::remove_element(element, tag.self_closing, flags);
453
+ }
454
+
455
+ Self::check_if_end_tag_needs_removal(element);
456
+ } else {
457
+ // anything in <iframe> must be removed, if it's kept
458
+ if Tag::is_iframe(tag) {
459
+ if self.0.borrow().flags[tag.index] != 0 {
460
+ element.set_inner_content(" ", ContentType::Text);
461
+ } else {
462
+ element.set_inner_content("", ContentType::Text);
463
+ }
464
+ }
465
+ }
466
+
467
+ should_remove
468
+ }
469
+
470
+ fn remove_element(element: &mut Element, self_closing: bool, flags: u8) {
471
+ let wrap_whitespace = (flags & Self::SELMA_SANITIZER_WRAP_WHITESPACE) != 0;
472
+ let remove_contents = (flags & Self::SELMA_SANITIZER_REMOVE_CONTENTS) != 0;
473
+
474
+ if remove_contents {
475
+ element.remove();
476
+ } else {
477
+ if wrap_whitespace {
478
+ if self_closing {
479
+ element.after(" ", ContentType::Text);
480
+ } else {
481
+ element.before(" ", ContentType::Text);
482
+ element.after(" ", ContentType::Text);
483
+ }
484
+ }
485
+ element.remove_and_keep_content();
486
+ }
487
+ }
488
+
489
+ pub fn force_remove_element(&self, element: &mut Element) {
490
+ let tag = Tag::tag_from_element(element);
491
+ let self_closing = tag.self_closing;
492
+ Self::remove_element(element, self_closing, Self::SELMA_SANITIZER_REMOVE_CONTENTS);
493
+ Self::check_if_end_tag_needs_removal(element);
494
+ }
495
+
496
+ fn check_if_end_tag_needs_removal(element: &mut Element) {
497
+ if element.removed() && !Tag::tag_from_element(element).self_closing {
498
+ element
499
+ .on_end_tag(move |end| {
500
+ Self::remove_end_tag(end);
501
+ Ok(())
502
+ })
503
+ .unwrap();
504
+ }
505
+ }
506
+
507
+ fn remove_end_tag(end_tag: &mut EndTag) {
508
+ end_tag.remove();
509
+ }
510
+
511
+ fn get_element_sanitizer<'a>(
512
+ binding: &'a RefMut<Sanitizer>,
513
+ element_name: &str,
514
+ ) -> &'a ElementSanitizer {
515
+ binding.element_sanitizers.get(element_name).unwrap()
516
+ }
517
+
518
+ fn get_mut_element_sanitizer<'a>(
519
+ binding: &'a mut Sanitizer,
520
+ element_name: &str,
521
+ ) -> &'a mut ElementSanitizer {
522
+ binding.element_sanitizers.get_mut(element_name).unwrap()
523
+ }
524
+ }
525
+
526
+ pub fn init(m_selma: RModule) -> Result<(), Error> {
527
+ let c_sanitizer = m_selma.define_class("Sanitizer", Default::default())?;
528
+
529
+ c_sanitizer.define_singleton_method("new", function!(SelmaSanitizer::new, -1))?;
530
+ c_sanitizer.define_method("config", method!(SelmaSanitizer::get_config, 0))?;
531
+
532
+ c_sanitizer.define_method("set_flag", method!(SelmaSanitizer::set_flag, 3))?;
533
+ c_sanitizer.define_method("set_all_flags", method!(SelmaSanitizer::set_all_flags, 2))?;
534
+
535
+ c_sanitizer.define_method(
536
+ "set_escape_tagfilter",
537
+ method!(SelmaSanitizer::set_escape_tagfilter, 1),
538
+ )?;
539
+ c_sanitizer.define_method(
540
+ "escape_tagfilter",
541
+ method!(SelmaSanitizer::get_escape_tagfilter, 0),
542
+ )?;
543
+
544
+ c_sanitizer.define_method(
545
+ "set_allow_comments",
546
+ method!(SelmaSanitizer::set_allow_comments, 1),
547
+ )?;
548
+ c_sanitizer.define_method(
549
+ "allow_comments",
550
+ method!(SelmaSanitizer::get_allow_comments, 0),
551
+ )?;
552
+
553
+ c_sanitizer.define_method(
554
+ "set_allow_doctype",
555
+ method!(SelmaSanitizer::set_allow_doctype, 1),
556
+ )?;
557
+ c_sanitizer.define_method(
558
+ "allow_doctype",
559
+ method!(SelmaSanitizer::get_allow_doctype, 0),
560
+ )?;
561
+
562
+ c_sanitizer.define_method(
563
+ "set_allowed_attribute",
564
+ method!(SelmaSanitizer::set_allowed_attribute, 3),
565
+ )?;
566
+
567
+ c_sanitizer.define_method(
568
+ "set_allowed_class",
569
+ method!(SelmaSanitizer::set_allowed_class, 3),
570
+ )?;
571
+
572
+ c_sanitizer.define_method(
573
+ "set_allowed_protocols",
574
+ method!(SelmaSanitizer::set_allowed_protocols, 3),
575
+ )?;
576
+
577
+ Ok(())
578
+ }
@@ -0,0 +1,115 @@
1
+ use magnus::{exception, function, scan_args, Error, Module, Object, RModule, Value};
2
+
3
+ #[derive(Clone, Debug)]
4
+ #[magnus::wrap(class = "Selma::Selector")]
5
+ pub struct SelmaSelector {
6
+ match_element: Option<String>,
7
+ match_text_within: Option<String>,
8
+ ignore_text_within: Option<Vec<String>>,
9
+ }
10
+
11
+ impl SelmaSelector {
12
+ fn new(args: &[Value]) -> Result<Self, Error> {
13
+ let (match_element, match_text_within, rb_ignore_text_within) =
14
+ Self::scan_parse_args(args)?;
15
+
16
+ if match_element.is_none() && match_text_within.is_none() {
17
+ return Err(Error::new(
18
+ exception::arg_error(),
19
+ "Neither `match_element` nor `match_text_within` option given",
20
+ ));
21
+ }
22
+
23
+ // FIXME: not excited about this double parse work (`element!` does it too),
24
+ // but at least we can bail ASAP if the CSS is invalid
25
+ if match_element.is_some() {
26
+ let css = match_element.as_ref().unwrap();
27
+ if css.parse::<lol_html::Selector>().is_err() {
28
+ return Err(Error::new(
29
+ exception::arg_error(),
30
+ format!("Could not parse `match_element` (`{}`) as valid CSS", css),
31
+ ));
32
+ }
33
+ }
34
+
35
+ if match_text_within.is_some() {
36
+ let css = match_text_within.as_ref().unwrap();
37
+ if css.parse::<lol_html::Selector>().is_err() {
38
+ return Err(Error::new(
39
+ exception::arg_error(),
40
+ format!(
41
+ "Could not parse `match_text_within` (`{}`) as valid CSS",
42
+ css
43
+ ),
44
+ ));
45
+ }
46
+ }
47
+
48
+ let ignore_text_within = match rb_ignore_text_within {
49
+ None => None,
50
+ Some(rb_ignore_text_within) => {
51
+ let mut ignore_text_within = vec![];
52
+ rb_ignore_text_within.iter().for_each(|i| {
53
+ // TODO: test this against malice
54
+ let ignore_text_within_tag_name = i.to_string();
55
+ ignore_text_within.push(ignore_text_within_tag_name);
56
+ });
57
+ Some(ignore_text_within)
58
+ }
59
+ };
60
+
61
+ Ok(Self {
62
+ match_element,
63
+ match_text_within,
64
+ ignore_text_within,
65
+ })
66
+ }
67
+
68
+ #[allow(clippy::let_unit_value)]
69
+ fn scan_parse_args(
70
+ args: &[Value],
71
+ ) -> Result<(Option<String>, Option<String>, Option<Vec<String>>), Error> {
72
+ let args = scan_args::scan_args(args)?;
73
+ let _: () = args.required;
74
+ let _: () = args.optional;
75
+ let _: () = args.splat;
76
+ let _: () = args.trailing;
77
+ let _: () = args.block;
78
+
79
+ let kw = scan_args::get_kwargs::<
80
+ _,
81
+ (),
82
+ (Option<String>, Option<String>, Option<Vec<String>>),
83
+ (),
84
+ >(
85
+ args.keywords,
86
+ &[],
87
+ &["match_element", "match_text_within", "ignore_text_within"],
88
+ )?;
89
+ let (match_element, match_text_within, rb_ignore_text_within) = kw.optional;
90
+
91
+ Ok((match_element, match_text_within, rb_ignore_text_within))
92
+ }
93
+
94
+ pub fn match_element(&self) -> Option<String> {
95
+ self.match_element.clone()
96
+ }
97
+
98
+ pub fn match_text_within(&self) -> Option<String> {
99
+ self.match_text_within.clone()
100
+ }
101
+
102
+ pub fn ignore_text_within(&self) -> Option<Vec<String>> {
103
+ self.ignore_text_within.clone()
104
+ }
105
+ }
106
+
107
+ pub fn init(m_selma: RModule) -> Result<(), Error> {
108
+ let c_selector = m_selma
109
+ .define_class("Selector", Default::default())
110
+ .expect("cannot define class Selma::Selector");
111
+
112
+ c_selector.define_singleton_method("new", function!(SelmaSelector::new, -1))?;
113
+
114
+ Ok(())
115
+ }