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;
11use std::sync::LazyLock;
12
13mod md033_config;
14use md033_config::MD033Config;
15
16#[derive(Clone)]
17pub struct MD033NoInlineHtml {
18    config: MD033Config,
19    allowed: HashSet<String>,
20}
21
22impl Default for MD033NoInlineHtml {
23    fn default() -> Self {
24        let config = MD033Config::default();
25        let allowed = config.allowed_set();
26        Self { config, allowed }
27    }
28}
29
30impl MD033NoInlineHtml {
31    pub fn new() -> Self {
32        Self::default()
33    }
34
35    pub fn with_allowed(allowed_vec: Vec<String>) -> Self {
36        let config = MD033Config {
37            allowed: allowed_vec.clone(),
38        };
39        let allowed = config.allowed_set();
40        Self { config, allowed }
41    }
42
43    pub fn from_config_struct(config: MD033Config) -> Self {
44        let allowed = config.allowed_set();
45        Self { config, allowed }
46    }
47
48    // Efficient check for allowed tags using HashSet (case-insensitive)
49    #[inline]
50    fn is_tag_allowed(&self, tag: &str) -> bool {
51        if self.allowed.is_empty() {
52            return false;
53        }
54        // Remove angle brackets and slashes, then split by whitespace or '>'
55        let tag = tag.trim_start_matches('<').trim_start_matches('/');
56        let tag_name = tag
57            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
58            .next()
59            .unwrap_or("");
60        self.allowed.contains(&tag_name.to_lowercase())
61    }
62
63    // Check if a tag is an HTML comment
64    #[inline]
65    fn is_html_comment(&self, tag: &str) -> bool {
66        tag.starts_with("<!--") && tag.ends_with("-->")
67    }
68
69    // Check if a tag is likely a programming type annotation rather than HTML
70    #[inline]
71    fn is_likely_type_annotation(&self, tag: &str) -> bool {
72        // Common programming type names that are often used in generics
73        const COMMON_TYPES: &[&str] = &[
74            "string",
75            "number",
76            "any",
77            "void",
78            "null",
79            "undefined",
80            "array",
81            "promise",
82            "function",
83            "error",
84            "date",
85            "regexp",
86            "symbol",
87            "bigint",
88            "map",
89            "set",
90            "weakmap",
91            "weakset",
92            "iterator",
93            "generator",
94            "t",
95            "u",
96            "v",
97            "k",
98            "e", // Common single-letter type parameters
99            "userdata",
100            "apiresponse",
101            "config",
102            "options",
103            "params",
104            "result",
105            "response",
106            "request",
107            "data",
108            "item",
109            "element",
110            "node",
111        ];
112
113        let tag_content = tag
114            .trim_start_matches('<')
115            .trim_end_matches('>')
116            .trim_start_matches('/');
117        let tag_name = tag_content
118            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
119            .next()
120            .unwrap_or("");
121
122        // Check if it's a simple tag (no attributes) with a common type name
123        if !tag_content.contains(' ') && !tag_content.contains('=') {
124            COMMON_TYPES.contains(&tag_name.to_ascii_lowercase().as_str())
125        } else {
126            false
127        }
128    }
129
130    // Check if a tag is actually an email address in angle brackets
131    #[inline]
132    fn is_email_address(&self, tag: &str) -> bool {
133        let content = tag.trim_start_matches('<').trim_end_matches('>');
134        // Simple email pattern: contains @ and has reasonable structure
135        content.contains('@')
136            && content.chars().all(|c| c.is_alphanumeric() || "@.-_+".contains(c))
137            && content.split('@').count() == 2
138            && content.split('@').all(|part| !part.is_empty())
139    }
140
141    // Check if a tag has the markdown attribute (MkDocs/Material for MkDocs)
142    #[inline]
143    fn has_markdown_attribute(&self, tag: &str) -> bool {
144        // Check for various forms of markdown attribute
145        // Examples: <div markdown>, <div markdown="1">, <div class="result" markdown>
146        tag.contains(" markdown>") || tag.contains(" markdown=") || tag.contains(" markdown ")
147    }
148
149    // Check if a tag is actually a URL in angle brackets
150    #[inline]
151    fn is_url_in_angle_brackets(&self, tag: &str) -> bool {
152        let content = tag.trim_start_matches('<').trim_end_matches('>');
153        // Check for common URL schemes
154        content.starts_with("http://")
155            || content.starts_with("https://")
156            || content.starts_with("ftp://")
157            || content.starts_with("ftps://")
158            || content.starts_with("mailto:")
159    }
160
161    /// Calculate fix to remove HTML tags while keeping content
162    ///
163    /// For self-closing tags like `<br/>`, returns a single fix to remove the tag.
164    /// For paired tags like `<span>text</span>`, returns the replacement text (just the content).
165    ///
166    /// Returns (range, replacement_text) where range is the bytes to replace
167    /// and replacement_text is what to put there (content without tags, or empty for self-closing).
168    fn calculate_fix(
169        &self,
170        content: &str,
171        opening_tag: &str,
172        tag_byte_start: usize,
173    ) -> Option<(std::ops::Range<usize>, String)> {
174        // Check if it's a self-closing tag (ends with />)
175        if opening_tag.ends_with("/>") {
176            return Some((tag_byte_start..tag_byte_start + opening_tag.len(), String::new()));
177        }
178
179        // Extract tag name from opening tag (e.g., "<div>" -> "div", "<span class='x'>" -> "span")
180        let tag_name = opening_tag
181            .trim_start_matches('<')
182            .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
183            .next()?
184            .to_lowercase();
185
186        // Build the closing tag pattern
187        let closing_tag = format!("</{tag_name}>");
188
189        // Search for the closing tag after the opening tag
190        let search_start = tag_byte_start + opening_tag.len();
191        if let Some(closing_pos) = content[search_start..].find(&closing_tag) {
192            let closing_byte_start = search_start + closing_pos;
193            let closing_byte_end = closing_byte_start + closing_tag.len();
194
195            // Extract the content between tags
196            let inner_content = &content[search_start..closing_byte_start];
197
198            return Some((tag_byte_start..closing_byte_end, inner_content.to_string()));
199        }
200
201        // If no closing tag found, just remove the opening tag
202        Some((tag_byte_start..tag_byte_start + opening_tag.len(), String::new()))
203    }
204
205    /// Find HTML tags that span multiple lines
206    fn find_multiline_html_tags(
207        &self,
208        ctx: &crate::lint_context::LintContext,
209        content: &str,
210        nomarkdown_ranges: &[(usize, usize)],
211        warnings: &mut Vec<LintWarning>,
212    ) {
213        // Early return: if content has no incomplete tags at line ends, skip processing
214        if !content.contains('<') || !content.lines().any(|line| line.trim_end().ends_with('<')) {
215            return;
216        }
217
218        // Simple approach: use regex to find patterns like <tagname and then look for closing >
219        static INCOMPLETE_TAG_START: LazyLock<regex::Regex> =
220            LazyLock::new(|| regex::Regex::new(r"(?i)<[a-zA-Z][^>]*$").unwrap());
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            // Calculate line byte offset once per line (not inside the loop)
404            let line_byte_offset: usize = ctx.line_index.get_line_start_byte(line_num).unwrap_or(0);
405
406            // Find all HTML opening tags in the line using regex
407            for tag_match in HTML_OPENING_TAG_FINDER.find_iter(line) {
408                let tag = tag_match.as_str();
409                let tag_byte_start = line_byte_offset + tag_match.start();
410
411                // Skip HTML tags inside HTML comments
412                if ctx.is_in_html_comment(tag_byte_start) {
413                    continue;
414                }
415
416                // Skip HTML comments themselves
417                if self.is_html_comment(tag) {
418                    continue;
419                }
420
421                // Skip JSX components in MDX files (e.g., <Chart />, <MyComponent>)
422                // JSX components start with uppercase letter
423                if ctx.flavor.supports_jsx() {
424                    // Extract tag name (remove angle brackets, slashes, and attributes)
425                    let tag_clean = tag.trim_start_matches('<').trim_start_matches('/');
426                    let tag_name = tag_clean
427                        .split(|c: char| c.is_whitespace() || c == '>' || c == '/')
428                        .next()
429                        .unwrap_or("");
430
431                    if tag_name.chars().next().is_some_and(|c| c.is_uppercase()) {
432                        continue;
433                    }
434                }
435
436                // Skip likely programming type annotations
437                if self.is_likely_type_annotation(tag) {
438                    continue;
439                }
440
441                // Skip email addresses in angle brackets
442                if self.is_email_address(tag) {
443                    continue;
444                }
445
446                // Skip URLs in angle brackets
447                if self.is_url_in_angle_brackets(tag) {
448                    continue;
449                }
450
451                // Skip tags inside code spans
452                let tag_start_col = tag_match.start() + 1; // 1-indexed
453                if ctx.is_in_code_span(line_num, tag_start_col) {
454                    continue;
455                }
456
457                // Skip allowed tags
458                if self.is_tag_allowed(tag) {
459                    continue;
460                }
461
462                // Skip tags with markdown attribute in MkDocs mode
463                if ctx.flavor == crate::config::MarkdownFlavor::MkDocs && self.has_markdown_attribute(tag) {
464                    continue;
465                }
466
467                // Report each HTML tag individually (true markdownlint compatibility)
468                let (start_line, start_col, end_line, end_col) =
469                    calculate_html_tag_range(line_num, line, tag_match.start(), tag_match.len());
470
471                // Calculate fix to remove HTML tags but keep content
472                let fix = self
473                    .calculate_fix(content, tag, tag_byte_start)
474                    .map(|(range, replacement)| Fix { range, replacement });
475
476                warnings.push(LintWarning {
477                    rule_name: Some(self.name().to_string()),
478                    line: start_line,
479                    column: start_col,
480                    end_line,
481                    end_column: end_col,
482                    message: format!("Inline HTML found: {tag}"),
483                    severity: Severity::Warning,
484                    fix,
485                });
486            }
487        }
488
489        // Third pass: find multi-line HTML tags
490        self.find_multiline_html_tags(ctx, ctx.content, &nomarkdown_ranges, &mut warnings);
491
492        Ok(warnings)
493    }
494
495    fn fix(&self, ctx: &crate::lint_context::LintContext) -> Result<String, LintError> {
496        // No fix for MD033: do not remove or alter HTML, just return the input unchanged
497        Ok(ctx.content.to_string())
498    }
499
500    fn fix_capability(&self) -> crate::rule::FixCapability {
501        crate::rule::FixCapability::Unfixable
502    }
503
504    /// Get the category of this rule for selective processing
505    fn category(&self) -> RuleCategory {
506        RuleCategory::Html
507    }
508
509    /// Check if this rule should be skipped
510    fn should_skip(&self, ctx: &crate::lint_context::LintContext) -> bool {
511        ctx.content.is_empty() || !ctx.likely_has_html()
512    }
513
514    fn as_any(&self) -> &dyn std::any::Any {
515        self
516    }
517
518    fn default_config_section(&self) -> Option<(String, toml::Value)> {
519        let json_value = serde_json::to_value(&self.config).ok()?;
520        Some((
521            self.name().to_string(),
522            crate::rule_config_serde::json_to_toml_value(&json_value)?,
523        ))
524    }
525
526    fn from_config(config: &crate::config::Config) -> Box<dyn Rule>
527    where
528        Self: Sized,
529    {
530        let rule_config = crate::rule_config_serde::load_rule_config::<MD033Config>(config);
531        Box::new(Self::from_config_struct(rule_config))
532    }
533}
534
535#[cfg(test)]
536mod tests {
537    use super::*;
538    use crate::lint_context::LintContext;
539    use crate::rule::Rule;
540
541    #[test]
542    fn test_md033_basic_html() {
543        let rule = MD033NoInlineHtml::default();
544        let content = "<div>Some content</div>";
545        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
546        let result = rule.check(&ctx).unwrap();
547        // Only reports opening tags, not closing tags
548        assert_eq!(result.len(), 1); // Only <div>, not </div>
549        assert!(result[0].message.starts_with("Inline HTML found: <div>"));
550    }
551
552    #[test]
553    fn test_md033_case_insensitive() {
554        let rule = MD033NoInlineHtml::default();
555        let content = "<DiV>Some <B>content</B></dIv>";
556        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
557        let result = rule.check(&ctx).unwrap();
558        // Only reports opening tags, not closing tags
559        assert_eq!(result.len(), 2); // <DiV>, <B> (not </B>, </dIv>)
560        assert_eq!(result[0].message, "Inline HTML found: <DiV>");
561        assert_eq!(result[1].message, "Inline HTML found: <B>");
562    }
563
564    #[test]
565    fn test_md033_allowed_tags() {
566        let rule = MD033NoInlineHtml::with_allowed(vec!["div".to_string(), "br".to_string()]);
567        let content = "<div>Allowed</div><p>Not allowed</p><br/>";
568        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
569        let result = rule.check(&ctx).unwrap();
570        // Only warnings for non-allowed opening tags (<p> only, div and br are allowed)
571        assert_eq!(result.len(), 1);
572        assert_eq!(result[0].message, "Inline HTML found: <p>");
573
574        // Test case-insensitivity of allowed tags
575        let content2 = "<DIV>Allowed</DIV><P>Not allowed</P><BR/>";
576        let ctx2 = LintContext::new(content2, crate::config::MarkdownFlavor::Standard);
577        let result2 = rule.check(&ctx2).unwrap();
578        assert_eq!(result2.len(), 1); // Only <P> flagged
579        assert_eq!(result2[0].message, "Inline HTML found: <P>");
580    }
581
582    #[test]
583    fn test_md033_html_comments() {
584        let rule = MD033NoInlineHtml::default();
585        let content = "<!-- This is a comment --> <p>Not a comment</p>";
586        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
587        let result = rule.check(&ctx).unwrap();
588        // Should detect warnings for HTML opening tags (comments are skipped, closing tags not reported)
589        assert_eq!(result.len(), 1); // Only <p>
590        assert_eq!(result[0].message, "Inline HTML found: <p>");
591    }
592
593    #[test]
594    fn test_md033_tags_in_links() {
595        let rule = MD033NoInlineHtml::default();
596        let content = "[Link](http://example.com/<div>)";
597        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
598        let result = rule.check(&ctx).unwrap();
599        // The <div> in the URL should be detected as HTML (not skipped)
600        assert_eq!(result.len(), 1);
601        assert_eq!(result[0].message, "Inline HTML found: <div>");
602
603        let content2 = "[Link <a>text</a>](url)";
604        let ctx2 = LintContext::new(content2, crate::config::MarkdownFlavor::Standard);
605        let result2 = rule.check(&ctx2).unwrap();
606        // Only reports opening tags
607        assert_eq!(result2.len(), 1); // Only <a>
608        assert_eq!(result2[0].message, "Inline HTML found: <a>");
609    }
610
611    #[test]
612    fn test_md033_fix_escaping() {
613        let rule = MD033NoInlineHtml::default();
614        let content = "Text with <div> and <br/> tags.";
615        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
616        let fixed_content = rule.fix(&ctx).unwrap();
617        // No fix for HTML tags; output should be unchanged
618        assert_eq!(fixed_content, content);
619    }
620
621    #[test]
622    fn test_md033_in_code_blocks() {
623        let rule = MD033NoInlineHtml::default();
624        let content = "```html\n<div>Code</div>\n```\n<div>Not code</div>";
625        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
626        let result = rule.check(&ctx).unwrap();
627        // Only reports opening tags outside code block
628        assert_eq!(result.len(), 1); // Only <div> outside code block
629        assert_eq!(result[0].message, "Inline HTML found: <div>");
630    }
631
632    #[test]
633    fn test_md033_in_code_spans() {
634        let rule = MD033NoInlineHtml::default();
635        let content = "Text with `<p>in code</p>` span. <br/> Not in span.";
636        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
637        let result = rule.check(&ctx).unwrap();
638        // Should detect <br/> outside code span, but not tags inside code span
639        assert_eq!(result.len(), 1);
640        assert_eq!(result[0].message, "Inline HTML found: <br/>");
641    }
642
643    #[test]
644    fn test_md033_issue_90_code_span_with_diff_block() {
645        // Test for issue #90: inline code span followed by diff code block
646        let rule = MD033NoInlineHtml::default();
647        let content = r#"# Heading
648
649`<env>`
650
651```diff
652- this
653+ that
654```"#;
655        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
656        let result = rule.check(&ctx).unwrap();
657        // Should NOT detect <env> as HTML since it's inside backticks
658        assert_eq!(result.len(), 0, "Should not report HTML tags inside code spans");
659    }
660
661    #[test]
662    fn test_md033_multiple_code_spans_with_angle_brackets() {
663        // Test multiple code spans on same line
664        let rule = MD033NoInlineHtml::default();
665        let content = "`<one>` and `<two>` and `<three>` are all code spans";
666        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
667        let result = rule.check(&ctx).unwrap();
668        assert_eq!(result.len(), 0, "Should not report HTML tags inside any code spans");
669    }
670
671    #[test]
672    fn test_md033_nested_angle_brackets_in_code_span() {
673        // Test nested angle brackets
674        let rule = MD033NoInlineHtml::default();
675        let content = "Text with `<<nested>>` brackets";
676        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
677        let result = rule.check(&ctx).unwrap();
678        assert_eq!(result.len(), 0, "Should handle nested angle brackets in code spans");
679    }
680
681    #[test]
682    fn test_md033_code_span_at_end_before_code_block() {
683        // Test code span at end of line before code block
684        let rule = MD033NoInlineHtml::default();
685        let content = "Testing `<test>`\n```\ncode here\n```";
686        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
687        let result = rule.check(&ctx).unwrap();
688        assert_eq!(result.len(), 0, "Should handle code span before code block");
689    }
690
691    #[test]
692    fn test_md033_quick_fix_inline_tag() {
693        // Test Quick Fix for inline HTML tags - keeps content, removes tags
694        let rule = MD033NoInlineHtml::default();
695        let content = "This has <span>inline text</span> that should keep content.";
696        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
697        let result = rule.check(&ctx).unwrap();
698
699        assert_eq!(result.len(), 1, "Should find one HTML tag");
700        assert!(result[0].fix.is_some(), "Should have a fix");
701
702        let fix = result[0].fix.as_ref().unwrap();
703        assert_eq!(&content[fix.range.clone()], "<span>inline text</span>");
704        assert_eq!(fix.replacement, "inline text");
705    }
706
707    #[test]
708    fn test_md033_quick_fix_multiline_tag() {
709        // Test Quick Fix for multiline HTML tags - keeps content
710        let rule = MD033NoInlineHtml::default();
711        let content = "<div>\nBlock content\n</div>";
712        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
713        let result = rule.check(&ctx).unwrap();
714
715        assert_eq!(result.len(), 1, "Should find one HTML tag");
716        assert!(result[0].fix.is_some(), "Should have a fix");
717
718        let fix = result[0].fix.as_ref().unwrap();
719        assert_eq!(&content[fix.range.clone()], "<div>\nBlock content\n</div>");
720        assert_eq!(fix.replacement, "\nBlock content\n");
721    }
722
723    #[test]
724    fn test_md033_quick_fix_self_closing_tag() {
725        // Test Quick Fix for self-closing tags - removes tag (no content)
726        let rule = MD033NoInlineHtml::default();
727        let content = "Self-closing: <br/>";
728        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
729        let result = rule.check(&ctx).unwrap();
730
731        assert_eq!(result.len(), 1, "Should find one HTML tag");
732        assert!(result[0].fix.is_some(), "Should have a fix");
733
734        let fix = result[0].fix.as_ref().unwrap();
735        assert_eq!(&content[fix.range.clone()], "<br/>");
736        assert_eq!(fix.replacement, "");
737    }
738
739    #[test]
740    fn test_md033_quick_fix_multiple_tags() {
741        // Test Quick Fix with multiple HTML tags - keeps content for both
742        let rule = MD033NoInlineHtml::default();
743        let content = "<span>first</span> and <strong>second</strong>";
744        let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard);
745        let result = rule.check(&ctx).unwrap();
746
747        assert_eq!(result.len(), 2, "Should find two HTML tags");
748        assert!(result[0].fix.is_some(), "First tag should have a fix");
749        assert!(result[1].fix.is_some(), "Second tag should have a fix");
750
751        let fix1 = result[0].fix.as_ref().unwrap();
752        assert_eq!(&content[fix1.range.clone()], "<span>first</span>");
753        assert_eq!(fix1.replacement, "first");
754
755        let fix2 = result[1].fix.as_ref().unwrap();
756        assert_eq!(&content[fix2.range.clone()], "<strong>second</strong>");
757        assert_eq!(fix2.replacement, "second");
758    }
759}
rumdl_lib/rules/md033_no_inline_html.rs

rumdl_lib/rules/
md033_no_inline_html.rs
