rumdl_lib/rules/
md033_no_inline_html.rs

1//!
2//! Rule MD033: No HTML tags
3//!
4//! See [docs/md033.md](../../docs/md033.md) for full documentation, configuration, and examples.
5
6use crate::rule::{Fix, LintError, LintResult, LintWarning, Rule, RuleCategory, Severity};
7use crate::utils::kramdown_utils::{is_kramdown_block_attribute, is_kramdown_extension};
8use crate::utils::range_utils::calculate_html_tag_range;
9use crate::utils::regex_cache::*;
10use std::collections::HashSet;
11
12mod md033_config;
13use md033_config::MD033Config;
14
15#[derive(Clone)]
16pub struct MD033NoInlineHtml {
17    config: MD033Config,
18    allowed: HashSet<String>,
19}
20
21impl Default for MD033NoInlineHtml {
22    fn default() -> Self {
23        let config = MD033Config::default();
24        let allowed = config.allowed_set();
25        Self { config, allowed }
26    }
27}
28
29impl MD033NoInlineHtml {
30    pub fn new() -> Self {
31        Self::default()
32    }
33
34    pub fn with_allowed(allowed_vec: Vec<String>) -> Self {
35        let config = MD033Config {
36            allowed: allowed_vec.clone(),
37        };
38        let allowed = config.allowed_set();
39        Self { config, allowed }
40    }
41
42    pub fn from_config_struct(config: MD033Config) -> Self {
43        let allowed = config.allowed_set();
44        Self { config, allowed }
45    }
46
47    // Efficient check for allowed tags using HashSet (case-insensitive)
48    #[inline]
49    fn is_tag_allowed(&self, tag: &str) -> bool {
50        if self.allowed.is_empty() {
51            return false;
52        }
53        // Remove angle brackets and slashes, then split by whitespace or '>'
54        let tag = tag.trim_start_matches('<').trim_start_matches('/');
55        let tag_name = tag
56            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
57            .next()
58            .unwrap_or("");
59        self.allowed.contains(&tag_name.to_lowercase())
60    }
61
62    // Check if a tag is an HTML comment
63    #[inline]
64    fn is_html_comment(&self, tag: &str) -> bool {
65        tag.starts_with("<!--") && tag.ends_with("-->")
66    }
67
68    // Check if a tag is likely a programming type annotation rather than HTML
69    #[inline]
70    fn is_likely_type_annotation(&self, tag: &str) -> bool {
71        // Common programming type names that are often used in generics
72        const COMMON_TYPES: &[&str] = &[
73            "string",
74            "number",
75            "any",
76            "void",
77            "null",
78            "undefined",
79            "array",
80            "promise",
81            "function",
82            "error",
83            "date",
84            "regexp",
85            "symbol",
86            "bigint",
87            "map",
88            "set",
89            "weakmap",
90            "weakset",
91            "iterator",
92            "generator",
93            "t",
94            "u",
95            "v",
96            "k",
97            "e", // Common single-letter type parameters
98            "userdata",
99            "apiresponse",
100            "config",
101            "options",
102            "params",
103            "result",
104            "response",
105            "request",
106            "data",
107            "item",
108            "element",
109            "node",
110        ];
111
112        let tag_content = tag
113            .trim_start_matches('<')
114            .trim_end_matches('>')
115            .trim_start_matches('/');
116        let tag_name = tag_content
117            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
118            .next()
119            .unwrap_or("");
120
121        // Check if it's a simple tag (no attributes) with a common type name
122        if !tag_content.contains(' ') && !tag_content.contains('=') {
123            COMMON_TYPES.contains(&tag_name.to_ascii_lowercase().as_str())
124        } else {
125            false
126        }
127    }
128
129    // Check if a tag is actually an email address in angle brackets
130    #[inline]
131    fn is_email_address(&self, tag: &str) -> bool {
132        let content = tag.trim_start_matches('<').trim_end_matches('>');
133        // Simple email pattern: contains @ and has reasonable structure
134        content.contains('@')
135            && content.chars().all(|c| c.is_alphanumeric() || "@.-_+".contains(c))
136            && content.split('@').count() == 2
137            && content.split('@').all(|part| !part.is_empty())
138    }
139
140    // Check if a tag has the markdown attribute (MkDocs/Material for MkDocs)
141    #[inline]
142    fn has_markdown_attribute(&self, tag: &str) -> bool {
143        // Check for various forms of markdown attribute
144        // Examples: <div markdown>, <div markdown="1">, <div class="result" markdown>
145        tag.contains(" markdown>") || tag.contains(" markdown=") || tag.contains(" markdown ")
146    }
147
148    // Check if a tag is actually a URL in angle brackets
149    #[inline]
150    fn is_url_in_angle_brackets(&self, tag: &str) -> bool {
151        let content = tag.trim_start_matches('<').trim_end_matches('>');
152        // Check for common URL schemes
153        content.starts_with("http://")
154            || content.starts_with("https://")
155            || content.starts_with("ftp://")
156            || content.starts_with("ftps://")
157            || content.starts_with("mailto:")
158    }
159
160    /// Calculate fix to remove HTML tags while keeping content
161    ///
162    /// For self-closing tags like `<br/>`, returns a single fix to remove the tag.
163    /// For paired tags like `<span>text</span>`, returns the replacement text (just the content).
164    ///
165    /// Returns (range, replacement_text) where range is the bytes to replace
166    /// and replacement_text is what to put there (content without tags, or empty for self-closing).
167    fn calculate_fix(
168        &self,
169        content: &str,
170        opening_tag: &str,
171        tag_byte_start: usize,
172    ) -> Option<(std::ops::Range<usize>, String)> {
173        // Check if it's a self-closing tag (ends with />)
174        if opening_tag.ends_with("/>") {
175            return Some((tag_byte_start..tag_byte_start + opening_tag.len(), String::new()));
176        }
177
178        // Extract tag name from opening tag (e.g., "<div>" -> "div", "<span class='x'>" -> "span")
179        let tag_name = opening_tag
180            .trim_start_matches('<')
181            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
182            .next()?
183            .to_lowercase();
184
185        // Build the closing tag pattern
186        let closing_tag = format!("</{tag_name}>");
187
188        // Search for the closing tag after the opening tag
189        let search_start = tag_byte_start + opening_tag.len();
190        if let Some(closing_pos) = content[search_start..].find(&closing_tag) {
191            let closing_byte_start = search_start + closing_pos;
192            let closing_byte_end = closing_byte_start + closing_tag.len();
193
194            // Extract the content between tags
195            let inner_content = &content[search_start..closing_byte_start];
196
197            return Some((tag_byte_start..closing_byte_end, inner_content.to_string()));
198        }
199
200        // If no closing tag found, just remove the opening tag
201        Some((tag_byte_start..tag_byte_start + opening_tag.len(), String::new()))
202    }
203
204    /// Find HTML tags that span multiple lines
205    fn find_multiline_html_tags(
206        &self,
207        ctx: &crate::lint_context::LintContext,
208        content: &str,
209        nomarkdown_ranges: &[(usize, usize)],
210        warnings: &mut Vec<LintWarning>,
211    ) {
212        // Early return: if content has no incomplete tags at line ends, skip processing
213        if !content.contains('<') || !content.lines().any(|line| line.trim_end().ends_with('<')) {
214            return;
215        }
216
217        // Simple approach: use regex to find patterns like <tagname and then look for closing >
218        lazy_static::lazy_static! {
219            static ref INCOMPLETE_TAG_START: regex::Regex = regex::Regex::new(r"(?i)<[a-zA-Z][^>]*$").unwrap();
220        }
221
222        let lines: Vec<&str> = content.lines().collect();
223
224        for (i, line) in lines.iter().enumerate() {
225            let line_num = i + 1;
226
227            // Skip code blocks and empty lines
228            if line.trim().is_empty() || ctx.line_info(line_num).is_some_and(|info| info.in_code_block) {
229                continue;
230            }
231
232            // Skip lines inside nomarkdown blocks
233            if nomarkdown_ranges
234                .iter()
235                .any(|(start, end)| line_num >= *start && line_num <= *end)
236            {
237                continue;
238            }
239
240            // Early return: skip lines that don't end with incomplete tags
241            if !line.contains('<') {
242                continue;
243            }
244
245            // Look for incomplete HTML tags at the end of the line
246            if let Some(incomplete_match) = INCOMPLETE_TAG_START.find(line) {
247                let start_column = incomplete_match.start() + 1; // 1-indexed
248
249                // Build the complete tag by looking at subsequent lines
250                let mut complete_tag = incomplete_match.as_str().to_string();
251                let mut found_end = false;
252
253                // Look for the closing > in subsequent lines (limit search to 10 lines)
254                for (j, next_line) in lines.iter().enumerate().skip(i + 1).take(10) {
255                    let next_line_num = j + 1;
256
257                    // Stop if we hit a code block
258                    if ctx.line_info(next_line_num).is_some_and(|info| info.in_code_block) {
259                        break;
260                    }
261
262                    complete_tag.push(' '); // Add space to normalize whitespace
263                    complete_tag.push_str(next_line.trim());
264
265                    if next_line.contains('>') {
266                        found_end = true;
267                        break;
268                    }
269                }
270
271                if found_end {
272                    // Extract just the tag part (up to the first >)
273                    if let Some(end_pos) = complete_tag.find('>') {
274                        let final_tag = &complete_tag[0..=end_pos];
275
276                        // Apply the same filters as single-line tags
277                        let skip_mkdocs_markdown = ctx.flavor == crate::config::MarkdownFlavor::MkDocs
278                            && self.has_markdown_attribute(final_tag);
279
280                        if !self.is_html_comment(final_tag)
281                            && !self.is_likely_type_annotation(final_tag)
282                            && !self.is_email_address(final_tag)
283                            && !self.is_url_in_angle_brackets(final_tag)
284                            && !self.is_tag_allowed(final_tag)
285                            && !skip_mkdocs_markdown
286                            && HTML_OPENING_TAG_FINDER.is_match(final_tag)
287                        {
288                            // Check for duplicates (avoid flagging the same position twice)
289                            let already_warned =
290                                warnings.iter().any(|w| w.line == line_num && w.column == start_column);
291
292                            if !already_warned {
293                                let (start_line, start_col, end_line, end_col) = calculate_html_tag_range(
294                                    line_num,
295                                    line,
296                                    incomplete_match.start(),
297                                    incomplete_match.len(),
298                                );
299                                warnings.push(LintWarning {
300                                    rule_name: Some(self.name().to_string()),
301                                    line: start_line,
302                                    column: start_col,
303                                    end_line,
304                                    end_column: end_col,
305                                    message: format!("HTML tag found: {final_tag}"),
306                                    severity: Severity::Warning,
307                                    fix: None,
308                                });
309                            }
310                        }
311                    }
312                }
313            }
314        }
315    }
316}
317
318impl Rule for MD033NoInlineHtml {
319    fn name(&self) -> &'static str {
320        "MD033"
321    }
322
323    fn description(&self) -> &'static str {
324        "Inline HTML is not allowed"
325    }
326
327    fn check(&self, ctx: &crate::lint_context::LintContext) -> LintResult {
328        let content = ctx.content;
329
330        // Early return: if no HTML tags at all, skip processing
331        if content.is_empty() || !ctx.likely_has_html() {
332            return Ok(Vec::new());
333        }
334
335        // Quick check for HTML tag pattern before expensive processing
336        if !HTML_TAG_QUICK_CHECK.is_match(content) {
337            return Ok(Vec::new());
338        }
339
340        let mut warnings = Vec::new();
341        let lines: Vec<&str> = content.lines().collect();
342
343        // Track nomarkdown and comment blocks
344        let mut in_nomarkdown = false;
345        let mut in_comment = false;
346        let mut nomarkdown_ranges: Vec<(usize, usize)> = Vec::new();
347        let mut nomarkdown_start = 0;
348        let mut comment_start = 0;
349
350        // First pass: identify nomarkdown and comment blocks
351        for (i, line) in lines.iter().enumerate() {
352            let line_num = i + 1;
353
354            // Check for nomarkdown start
355            if line.trim() == "{::nomarkdown}" {
356                in_nomarkdown = true;
357                nomarkdown_start = line_num;
358            } else if line.trim() == "{:/nomarkdown}" && in_nomarkdown {
359                in_nomarkdown = false;
360                nomarkdown_ranges.push((nomarkdown_start, line_num));
361            }
362
363            // Check for comment blocks
364            if line.trim() == "{::comment}" {
365                in_comment = true;
366                comment_start = line_num;
367            } else if line.trim() == "{:/comment}" && in_comment {
368                in_comment = false;
369                nomarkdown_ranges.push((comment_start, line_num));
370            }
371        }
372
373        // Second pass: find single-line HTML tags
374        // To match markdownlint behavior, report one warning per HTML tag
375        for (i, line) in lines.iter().enumerate() {
376            let line_num = i + 1;
377
378            if line.trim().is_empty() {
379                continue;
380            }
381            if ctx.line_info(line_num).is_some_and(|info| info.in_code_block) {
382                continue;
383            }
384            // Skip lines that are indented code blocks (4+ spaces or tab) per CommonMark spec
385            // Even if they're not in the structure's code blocks (e.g., HTML blocks)
386            if line.starts_with("    ") || line.starts_with('\t') {
387                continue;
388            }
389
390            // Skip lines inside nomarkdown blocks
391            if nomarkdown_ranges
392                .iter()
393                .any(|(start, end)| line_num >= *start && line_num <= *end)
394            {
395                continue;
396            }
397
398            // Skip Kramdown extensions and block attributes
399            if is_kramdown_extension(line) || is_kramdown_block_attribute(line) {
400                continue;
401            }
402
403            // Find all HTML opening tags in the line using regex
404            for tag_match in HTML_OPENING_TAG_FINDER.find_iter(line) {
405                let tag = tag_match.as_str();
406
407                // Skip HTML comments
408                if self.is_html_comment(tag) {
409                    continue;
410                }
411
412                // Skip JSX components in MDX files (e.g., <Chart />, <MyComponent>)
413                // JSX components start with uppercase letter
414                if ctx.flavor.supports_jsx() {
415                    // Extract tag name (remove angle brackets, slashes, and attributes)
416                    let tag_clean = tag.trim_start_matches('<').trim_start_matches('/');
417                    let tag_name = tag_clean
418                        .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
419                        .next()
420                        .unwrap_or("");
421
422                    if tag_name.chars().next().is_some_and(|c| c.is_uppercase()) {
423                        continue;
424                    }
425                }
426
427                // Skip likely programming type annotations
428                if self.is_likely_type_annotation(tag) {
429                    continue;
430                }
431
432                // Skip email addresses in angle brackets
433                if self.is_email_address(tag) {
434                    continue;
435                }
436
437                // Skip URLs in angle brackets
438                if self.is_url_in_angle_brackets(tag) {
439                    continue;
440                }
441
442                // Skip tags inside code spans
443                let tag_start_col = tag_match.start() + 1; // 1-indexed
444                if ctx.is_in_code_span(line_num, tag_start_col) {
445                    continue;
446                }
447
448                // Skip allowed tags
449                if self.is_tag_allowed(tag) {
450                    continue;
451                }
452
453                // Skip tags with markdown attribute in MkDocs mode
454                if ctx.flavor == crate::config::MarkdownFlavor::MkDocs && self.has_markdown_attribute(tag) {
455                    continue;
456                }
457
458                // Report each HTML tag individually (true markdownlint compatibility)
459                let (start_line, start_col, end_line, end_col) =
460                    calculate_html_tag_range(line_num, line, tag_match.start(), tag_match.len());
461
462                // Calculate byte offset for the tag in the entire content
463                let line_byte_offset: usize = content
464                    .lines()
465                    .take(line_num - 1)
466                    .map(|l| l.len() + 1) // +1 for newline
467                    .sum();
468                let tag_byte_start = line_byte_offset + tag_match.start();
469
470                // Calculate fix to remove HTML tags but keep content
471                let fix = self
472                    .calculate_fix(content, tag, tag_byte_start)
473                    .map(|(range, replacement)| Fix { range, replacement });
474
475                warnings.push(LintWarning {
476                    rule_name: Some(self.name().to_string()),
477                    line: start_line,
478                    column: start_col,
479                    end_line,
480                    end_column: end_col,
481                    message: format!("Inline HTML found: {tag}"),
482                    severity: Severity::Warning,
483                    fix,
484                });
485            }
486        }
487
488        // Third pass: find multi-line HTML tags
489        self.find_multiline_html_tags(ctx, ctx.content, &nomarkdown_ranges, &mut warnings);
490
491        Ok(warnings)
492    }
493
494    fn fix(&self, ctx: &crate::lint_context::LintContext) -> Result<String, LintError> {
495        // No fix for MD033: do not remove or alter HTML, just return the input unchanged
496        Ok(ctx.content.to_string())
497    }
498
499    fn fix_capability(&self) -> crate::rule::FixCapability {
500        crate::rule::FixCapability::Unfixable
501    }
502
503    /// Get the category of this rule for selective processing
504    fn category(&self) -> RuleCategory {
505        RuleCategory::Html
506    }
507
508    /// Check if this rule should be skipped
509    fn should_skip(&self, ctx: &crate::lint_context::LintContext) -> bool {
510        ctx.content.is_empty() || !ctx.likely_has_html()
511    }
512
513    fn as_any(&self) -> &dyn std::any::Any {
514        self
515    }
516
517    fn default_config_section(&self) -> Option<(String, toml::Value)> {
518        let json_value = serde_json::to_value(&self.config).ok()?;
519        Some((
520            self.name().to_string(),
521            crate::rule_config_serde::json_to_toml_value(&json_value)?,
522        ))
523    }
524
525    fn from_config(config: &crate::config::Config) -> Box<dyn Rule>
526    where
527        Self: Sized,
528    {
529        let rule_config = crate::rule_config_serde::load_rule_config::<MD033Config>(config);
530        Box::new(Self::from_config_struct(rule_config))
531    }
532}
533
534#[cfg(test)]
535mod tests {
536    use super::*;
537    use crate::lint_context::LintContext;
538    use crate::rule::Rule;
539
540    #[test]
541    fn test_md033_basic_html() {
542        let rule = MD033NoInlineHtml::default();
543        let content = "<div>Some content</div>";
544        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
545        let result = rule.check(&ctx).unwrap();
546        // Only reports opening tags, not closing tags
547        assert_eq!(result.len(), 1); // Only <div>, not </div>
548        assert!(result[0].message.starts_with("Inline HTML found: <div>"));
549    }
550
551    #[test]
552    fn test_md033_case_insensitive() {
553        let rule = MD033NoInlineHtml::default();
554        let content = "<DiV>Some <B>content</B></dIv>";
555        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
556        let result = rule.check(&ctx).unwrap();
557        // Only reports opening tags, not closing tags
558        assert_eq!(result.len(), 2); // <DiV>, <B> (not </B>, </dIv>)
559        assert_eq!(result[0].message, "Inline HTML found: <DiV>");
560        assert_eq!(result[1].message, "Inline HTML found: <B>");
561    }
562
563    #[test]
564    fn test_md033_allowed_tags() {
565        let rule = MD033NoInlineHtml::with_allowed(vec!["div".to_string(), "br".to_string()]);
566        let content = "<div>Allowed</div><p>Not allowed</p><br/>";
567        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
568        let result = rule.check(&ctx).unwrap();
569        // Only warnings for non-allowed opening tags (<p> only, div and br are allowed)
570        assert_eq!(result.len(), 1);
571        assert_eq!(result[0].message, "Inline HTML found: <p>");
572
573        // Test case-insensitivity of allowed tags
574        let content2 = "<DIV>Allowed</DIV><P>Not allowed</P><BR/>";
575        let ctx2 = LintContext::new(content2, crate::config::MarkdownFlavor::Standard);
576        let result2 = rule.check(&ctx2).unwrap();
577        assert_eq!(result2.len(), 1); // Only <P> flagged
578        assert_eq!(result2[0].message, "Inline HTML found: <P>");
579    }
580
581    #[test]
582    fn test_md033_html_comments() {
583        let rule = MD033NoInlineHtml::default();
584        let content = "<!-- This is a comment --> <p>Not a comment</p>";
585        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
586        let result = rule.check(&ctx).unwrap();
587        // Should detect warnings for HTML opening tags (comments are skipped, closing tags not reported)
588        assert_eq!(result.len(), 1); // Only <p>
589        assert_eq!(result[0].message, "Inline HTML found: <p>");
590    }
591
592    #[test]
593    fn test_md033_tags_in_links() {
594        let rule = MD033NoInlineHtml::default();
595        let content = "[Link](http://example.com/<div>)";
596        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
597        let result = rule.check(&ctx).unwrap();
598        // The <div> in the URL should be detected as HTML (not skipped)
599        assert_eq!(result.len(), 1);
600        assert_eq!(result[0].message, "Inline HTML found: <div>");
601
602        let content2 = "[Link <a>text</a>](url)";
603        let ctx2 = LintContext::new(content2, crate::config::MarkdownFlavor::Standard);
604        let result2 = rule.check(&ctx2).unwrap();
605        // Only reports opening tags
606        assert_eq!(result2.len(), 1); // Only <a>
607        assert_eq!(result2[0].message, "Inline HTML found: <a>");
608    }
609
610    #[test]
611    fn test_md033_fix_escaping() {
612        let rule = MD033NoInlineHtml::default();
613        let content = "Text with <div> and <br/> tags.";
614        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
615        let fixed_content = rule.fix(&ctx).unwrap();
616        // No fix for HTML tags; output should be unchanged
617        assert_eq!(fixed_content, content);
618    }
619
620    #[test]
621    fn test_md033_in_code_blocks() {
622        let rule = MD033NoInlineHtml::default();
623        let content = "```html\n<div>Code</div>\n```\n<div>Not code</div>";
624        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
625        let result = rule.check(&ctx).unwrap();
626        // Only reports opening tags outside code block
627        assert_eq!(result.len(), 1); // Only <div> outside code block
628        assert_eq!(result[0].message, "Inline HTML found: <div>");
629    }
630
631    #[test]
632    fn test_md033_in_code_spans() {
633        let rule = MD033NoInlineHtml::default();
634        let content = "Text with `<p>in code</p>` span. <br/> Not in span.";
635        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
636        let result = rule.check(&ctx).unwrap();
637        // Should detect <br/> outside code span, but not tags inside code span
638        assert_eq!(result.len(), 1);
639        assert_eq!(result[0].message, "Inline HTML found: <br/>");
640    }
641
642    #[test]
643    fn test_md033_issue_90_code_span_with_diff_block() {
644        // Test for issue #90: inline code span followed by diff code block
645        let rule = MD033NoInlineHtml::default();
646        let content = r#"# Heading
647
648`<env>`
649
650```diff
651- this
652+ that
653```"#;
654        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
655        let result = rule.check(&ctx).unwrap();
656        // Should NOT detect <env> as HTML since it's inside backticks
657        assert_eq!(result.len(), 0, "Should not report HTML tags inside code spans");
658    }
659
660    #[test]
661    fn test_md033_multiple_code_spans_with_angle_brackets() {
662        // Test multiple code spans on same line
663        let rule = MD033NoInlineHtml::default();
664        let content = "`<one>` and `<two>` and `<three>` are all code spans";
665        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
666        let result = rule.check(&ctx).unwrap();
667        assert_eq!(result.len(), 0, "Should not report HTML tags inside any code spans");
668    }
669
670    #[test]
671    fn test_md033_nested_angle_brackets_in_code_span() {
672        // Test nested angle brackets
673        let rule = MD033NoInlineHtml::default();
674        let content = "Text with `<<nested>>` brackets";
675        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
676        let result = rule.check(&ctx).unwrap();
677        assert_eq!(result.len(), 0, "Should handle nested angle brackets in code spans");
678    }
679
680    #[test]
681    fn test_md033_code_span_at_end_before_code_block() {
682        // Test code span at end of line before code block
683        let rule = MD033NoInlineHtml::default();
684        let content = "Testing `<test>`\n```\ncode here\n```";
685        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
686        let result = rule.check(&ctx).unwrap();
687        assert_eq!(result.len(), 0, "Should handle code span before code block");
688    }
689
690    #[test]
691    fn test_md033_quick_fix_inline_tag() {
692        // Test Quick Fix for inline HTML tags - keeps content, removes tags
693        let rule = MD033NoInlineHtml::default();
694        let content = "This has <span>inline text</span> that should keep content.";
695        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
696        let result = rule.check(&ctx).unwrap();
697
698        assert_eq!(result.len(), 1, "Should find one HTML tag");
699        assert!(result[0].fix.is_some(), "Should have a fix");
700
701        let fix = result[0].fix.as_ref().unwrap();
702        assert_eq!(&content[fix.range.clone()], "<span>inline text</span>");
703        assert_eq!(fix.replacement, "inline text");
704    }
705
706    #[test]
707    fn test_md033_quick_fix_multiline_tag() {
708        // Test Quick Fix for multiline HTML tags - keeps content
709        let rule = MD033NoInlineHtml::default();
710        let content = "<div>\nBlock content\n</div>";
711        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
712        let result = rule.check(&ctx).unwrap();
713
714        assert_eq!(result.len(), 1, "Should find one HTML tag");
715        assert!(result[0].fix.is_some(), "Should have a fix");
716
717        let fix = result[0].fix.as_ref().unwrap();
718        assert_eq!(&content[fix.range.clone()], "<div>\nBlock content\n</div>");
719        assert_eq!(fix.replacement, "\nBlock content\n");
720    }
721
722    #[test]
723    fn test_md033_quick_fix_self_closing_tag() {
724        // Test Quick Fix for self-closing tags - removes tag (no content)
725        let rule = MD033NoInlineHtml::default();
726        let content = "Self-closing: <br/>";
727        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
728        let result = rule.check(&ctx).unwrap();
729
730        assert_eq!(result.len(), 1, "Should find one HTML tag");
731        assert!(result[0].fix.is_some(), "Should have a fix");
732
733        let fix = result[0].fix.as_ref().unwrap();
734        assert_eq!(&content[fix.range.clone()], "<br/>");
735        assert_eq!(fix.replacement, "");
736    }
737
738    #[test]
739    fn test_md033_quick_fix_multiple_tags() {
740        // Test Quick Fix with multiple HTML tags - keeps content for both
741        let rule = MD033NoInlineHtml::default();
742        let content = "<span>first</span> and <strong>second</strong>";
743        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
744        let result = rule.check(&ctx).unwrap();
745
746        assert_eq!(result.len(), 2, "Should find two HTML tags");
747        assert!(result[0].fix.is_some(), "First tag should have a fix");
748        assert!(result[1].fix.is_some(), "Second tag should have a fix");
749
750        let fix1 = result[0].fix.as_ref().unwrap();
751        assert_eq!(&content[fix1.range.clone()], "<span>first</span>");
752        assert_eq!(fix1.replacement, "first");
753
754        let fix2 = result[1].fix.as_ref().unwrap();
755        assert_eq!(&content[fix2.range.clone()], "<strong>second</strong>");
756        assert_eq!(fix2.replacement, "second");
757    }
758}
rumdl_lib/rules/md033_no_inline_html.rs

rumdl_lib/rules/
md033_no_inline_html.rs
