rumdl/
lib.rs

1pub mod config;
2pub mod exit_codes;
3pub mod inline_config;
4pub mod lint_context;
5pub mod lsp;
6pub mod markdownlint_config;
7pub mod output;
8pub mod parallel;
9pub mod performance;
10pub mod profiling;
11pub mod rule;
12pub mod vscode;
13#[macro_use]
14pub mod rule_config;
15#[macro_use]
16pub mod rule_config_serde;
17pub mod rules;
18pub mod utils;
19
20#[cfg(feature = "python")]
21pub mod python;
22
23pub use rules::heading_utils::{Heading, HeadingStyle};
24pub use rules::*;
25
26pub use crate::lint_context::{LineInfo, LintContext, ListItemInfo};
27use crate::rule::{LintResult, Rule, RuleCategory};
28use crate::utils::document_structure::DocumentStructure;
29use std::time::Instant;
30
31/// Content characteristics for efficient rule filtering
32#[derive(Debug, Default)]
33struct ContentCharacteristics {
34    has_headings: bool,    // # or setext headings
35    has_lists: bool,       // *, -, +, 1. etc
36    has_links: bool,       // [text](url) or [text][ref]
37    has_code: bool,        // ``` or ~~~ or indented code
38    has_emphasis: bool,    // * or _ for emphasis
39    has_html: bool,        // < > tags
40    has_tables: bool,      // | pipes
41    has_blockquotes: bool, // > markers
42    has_images: bool,      // ![alt](url)
43}
44
45impl ContentCharacteristics {
46    fn analyze(content: &str) -> Self {
47        let mut chars = Self { ..Default::default() };
48
49        // Quick single-pass analysis
50        let mut has_atx_heading = false;
51        let mut has_setext_heading = false;
52
53        for line in content.lines() {
54            let trimmed = line.trim();
55
56            // Headings: ATX (#) or Setext (underlines)
57            if !has_atx_heading && trimmed.starts_with('#') {
58                has_atx_heading = true;
59            }
60            if !has_setext_heading && (trimmed.chars().all(|c| c == '=' || c == '-') && trimmed.len() > 1) {
61                has_setext_heading = true;
62            }
63
64            // Quick character-based detection (more efficient than regex)
65            if !chars.has_lists && (line.contains("* ") || line.contains("- ") || line.contains("+ ")) {
66                chars.has_lists = true;
67            }
68            if !chars.has_lists && line.chars().next().is_some_and(|c| c.is_ascii_digit()) && line.contains(". ") {
69                chars.has_lists = true;
70            }
71            if !chars.has_links
72                && (line.contains('[')
73                    || line.contains("http://")
74                    || line.contains("https://")
75                    || line.contains("ftp://"))
76            {
77                chars.has_links = true;
78            }
79            if !chars.has_images && line.contains("![") {
80                chars.has_images = true;
81            }
82            if !chars.has_code && (line.contains('`') || line.contains("~~~")) {
83                chars.has_code = true;
84            }
85            if !chars.has_emphasis && (line.contains('*') || line.contains('_')) {
86                chars.has_emphasis = true;
87            }
88            if !chars.has_html && line.contains('<') {
89                chars.has_html = true;
90            }
91            if !chars.has_tables && line.contains('|') {
92                chars.has_tables = true;
93            }
94            if !chars.has_blockquotes && line.starts_with('>') {
95                chars.has_blockquotes = true;
96            }
97        }
98
99        chars.has_headings = has_atx_heading || has_setext_heading;
100        chars
101    }
102
103    /// Check if a rule should be skipped based on content characteristics
104    fn should_skip_rule(&self, rule: &dyn Rule) -> bool {
105        match rule.category() {
106            RuleCategory::Heading => !self.has_headings,
107            RuleCategory::List => !self.has_lists,
108            RuleCategory::Link => !self.has_links && !self.has_images,
109            RuleCategory::Image => !self.has_images,
110            RuleCategory::CodeBlock => !self.has_code,
111            RuleCategory::Html => !self.has_html,
112            RuleCategory::Emphasis => !self.has_emphasis,
113            RuleCategory::Blockquote => !self.has_blockquotes,
114            RuleCategory::Table => !self.has_tables,
115            // Always check these categories as they apply to all content
116            RuleCategory::Whitespace | RuleCategory::FrontMatter | RuleCategory::Other => false,
117        }
118    }
119}
120
121/// Lint a file against the given rules with intelligent rule filtering
122/// Assumes the provided `rules` vector contains the final,
123/// configured, and filtered set of rules to be executed.
124pub fn lint(content: &str, rules: &[Box<dyn Rule>], _verbose: bool) -> LintResult {
125    let mut warnings = Vec::new();
126    let _overall_start = Instant::now();
127
128    // Early return for empty content
129    if content.is_empty() {
130        return Ok(warnings);
131    }
132
133    // Parse inline configuration comments once
134    let inline_config = crate::inline_config::InlineConfig::from_content(content);
135
136    // Analyze content characteristics for rule filtering
137    let characteristics = ContentCharacteristics::analyze(content);
138
139    // Filter rules based on content characteristics
140    let applicable_rules: Vec<_> = rules
141        .iter()
142        .filter(|rule| !characteristics.should_skip_rule(rule.as_ref()))
143        .collect();
144
145    // Calculate skipped rules count before consuming applicable_rules
146    let _total_rules = rules.len();
147    let _applicable_count = applicable_rules.len();
148
149    // Parse DocumentStructure once
150    let structure = DocumentStructure::new(content);
151
152    // Parse AST once for rules that can benefit from it
153    let ast_rules_count = applicable_rules.iter().filter(|rule| rule.uses_ast()).count();
154    let ast = if ast_rules_count > 0 {
155        Some(crate::utils::ast_utils::get_cached_ast(content))
156    } else {
157        None
158    };
159
160    // Parse LintContext once (migration step)
161    let lint_ctx = crate::lint_context::LintContext::new(content);
162
163    for rule in applicable_rules {
164        let _rule_start = Instant::now();
165
166        // Try optimized paths in order of preference
167        let result = if rule.uses_ast() {
168            if let Some(ref ast_ref) = ast {
169                // 1. AST-based path
170                rule.as_maybe_ast()
171                    .and_then(|ext| ext.check_with_ast_opt(&lint_ctx, ast_ref))
172                    .unwrap_or_else(|| rule.check_with_ast(&lint_ctx, ast_ref))
173            } else {
174                // Fallback to regular check if no AST
175                rule.as_maybe_document_structure()
176                    .and_then(|ext| ext.check_with_structure_opt(&lint_ctx, &structure))
177                    .unwrap_or_else(|| rule.check(&lint_ctx))
178            }
179        } else {
180            // 2. Document structure path
181            rule.as_maybe_document_structure()
182                .and_then(|ext| ext.check_with_structure_opt(&lint_ctx, &structure))
183                .unwrap_or_else(|| rule.check(&lint_ctx))
184        };
185
186        match result {
187            Ok(rule_warnings) => {
188                // Filter out warnings for rules disabled via inline comments
189                let filtered_warnings: Vec<_> = rule_warnings
190                    .into_iter()
191                    .filter(|warning| {
192                        // Use the warning's rule_name if available, otherwise use the rule's name
193                        let rule_name_to_check = warning.rule_name.unwrap_or(rule.name());
194
195                        // Extract the base rule name for sub-rules like "MD029-style" -> "MD029"
196                        let base_rule_name = if let Some(dash_pos) = rule_name_to_check.find('-') {
197                            &rule_name_to_check[..dash_pos]
198                        } else {
199                            rule_name_to_check
200                        };
201
202                        !inline_config.is_rule_disabled(
203                            base_rule_name,
204                            warning.line, // Already 1-indexed
205                        )
206                    })
207                    .collect();
208                warnings.extend(filtered_warnings);
209            }
210            Err(e) => {
211                log::error!("Error checking rule {}: {}", rule.name(), e);
212                return Err(e);
213            }
214        }
215
216        #[cfg(not(test))]
217        if _verbose {
218            let rule_duration = _rule_start.elapsed();
219            if rule_duration.as_millis() > 500 {
220                log::debug!("Rule {} took {:?}", rule.name(), rule_duration);
221            }
222        }
223    }
224
225    #[cfg(not(test))]
226    if _verbose {
227        let skipped_rules = _total_rules - _applicable_count;
228        if skipped_rules > 0 {
229            log::debug!("Skipped {skipped_rules} of {_total_rules} rules based on content analysis");
230        }
231        if ast.is_some() {
232            log::debug!("Used shared AST for {ast_rules_count} rules");
233        }
234    }
235
236    Ok(warnings)
237}
238
239/// Get the profiling report
240pub fn get_profiling_report() -> String {
241    profiling::get_report()
242}
243
244/// Reset the profiling data
245pub fn reset_profiling() {
246    profiling::reset()
247}
248
249/// Get regex cache statistics for performance monitoring
250pub fn get_regex_cache_stats() -> std::collections::HashMap<String, u64> {
251    crate::utils::regex_cache::get_cache_stats()
252}
253
254/// Get AST cache statistics for performance monitoring
255pub fn get_ast_cache_stats() -> std::collections::HashMap<u64, u64> {
256    crate::utils::ast_utils::get_ast_cache_stats()
257}
258
259/// Clear all caches (useful for testing and memory management)
260pub fn clear_all_caches() {
261    crate::utils::ast_utils::clear_ast_cache();
262    // Note: Regex cache is intentionally not cleared as it's global and shared
263}
264
265/// Get comprehensive cache performance report
266pub fn get_cache_performance_report() -> String {
267    let regex_stats = get_regex_cache_stats();
268    let ast_stats = get_ast_cache_stats();
269
270    let mut report = String::new();
271
272    report.push_str("=== Cache Performance Report ===\n\n");
273
274    // Regex cache statistics
275    report.push_str("Regex Cache:\n");
276    if regex_stats.is_empty() {
277        report.push_str("  No regex patterns cached\n");
278    } else {
279        let total_usage: u64 = regex_stats.values().sum();
280        report.push_str(&format!("  Total patterns: {}\n", regex_stats.len()));
281        report.push_str(&format!("  Total usage: {total_usage}\n"));
282
283        // Show top 5 most used patterns
284        let mut sorted_patterns: Vec<_> = regex_stats.iter().collect();
285        sorted_patterns.sort_by(|a, b| b.1.cmp(a.1));
286
287        report.push_str("  Top patterns by usage:\n");
288        for (pattern, count) in sorted_patterns.iter().take(5) {
289            let truncated_pattern = if pattern.len() > 50 {
290                format!("{}...", &pattern[..47])
291            } else {
292                pattern.to_string()
293            };
294            report.push_str(&format!(
295                "    {} ({}x): {}\n",
296                count,
297                pattern.len().min(50),
298                truncated_pattern
299            ));
300        }
301    }
302
303    report.push('\n');
304
305    // AST cache statistics
306    report.push_str("AST Cache:\n");
307    if ast_stats.is_empty() {
308        report.push_str("  No AST nodes cached\n");
309    } else {
310        let total_usage: u64 = ast_stats.values().sum();
311        report.push_str(&format!("  Total ASTs: {}\n", ast_stats.len()));
312        report.push_str(&format!("  Total usage: {total_usage}\n"));
313
314        if total_usage > ast_stats.len() as u64 {
315            let cache_hit_rate = ((total_usage - ast_stats.len() as u64) as f64 / total_usage as f64) * 100.0;
316            report.push_str(&format!("  Cache hit rate: {cache_hit_rate:.1}%\n"));
317        }
318    }
319
320    report
321}
322
323#[cfg(test)]
324mod tests {
325    use super::*;
326    use crate::rule::Rule;
327    use crate::rules::{MD001HeadingIncrement, MD009TrailingSpaces, MD012NoMultipleBlanks};
328
329    #[test]
330    fn test_content_characteristics_analyze() {
331        // Test empty content
332        let chars = ContentCharacteristics::analyze("");
333        assert!(!chars.has_headings);
334        assert!(!chars.has_lists);
335        assert!(!chars.has_links);
336        assert!(!chars.has_code);
337        assert!(!chars.has_emphasis);
338        assert!(!chars.has_html);
339        assert!(!chars.has_tables);
340        assert!(!chars.has_blockquotes);
341        assert!(!chars.has_images);
342
343        // Test content with headings
344        let chars = ContentCharacteristics::analyze("# Heading");
345        assert!(chars.has_headings);
346
347        // Test setext headings
348        let chars = ContentCharacteristics::analyze("Heading\n=======");
349        assert!(chars.has_headings);
350
351        // Test lists
352        let chars = ContentCharacteristics::analyze("* Item\n- Item 2\n+ Item 3");
353        assert!(chars.has_lists);
354
355        // Test ordered lists
356        let chars = ContentCharacteristics::analyze("1. First\n2. Second");
357        assert!(chars.has_lists);
358
359        // Test links
360        let chars = ContentCharacteristics::analyze("[link](url)");
361        assert!(chars.has_links);
362
363        // Test URLs
364        let chars = ContentCharacteristics::analyze("Visit https://example.com");
365        assert!(chars.has_links);
366
367        // Test images
368        let chars = ContentCharacteristics::analyze("![alt text](image.png)");
369        assert!(chars.has_images);
370
371        // Test code
372        let chars = ContentCharacteristics::analyze("`inline code`");
373        assert!(chars.has_code);
374
375        let chars = ContentCharacteristics::analyze("~~~\ncode block\n~~~");
376        assert!(chars.has_code);
377
378        // Test emphasis
379        let chars = ContentCharacteristics::analyze("*emphasis* and _more_");
380        assert!(chars.has_emphasis);
381
382        // Test HTML
383        let chars = ContentCharacteristics::analyze("<div>HTML content</div>");
384        assert!(chars.has_html);
385
386        // Test tables
387        let chars = ContentCharacteristics::analyze("| Header | Header |\n|--------|--------|");
388        assert!(chars.has_tables);
389
390        // Test blockquotes
391        let chars = ContentCharacteristics::analyze("> Quote");
392        assert!(chars.has_blockquotes);
393
394        // Test mixed content
395        let content = "# Heading\n* List item\n[link](url)\n`code`\n*emphasis*\n<p>html</p>\n| table |\n> quote\n![image](img.png)";
396        let chars = ContentCharacteristics::analyze(content);
397        assert!(chars.has_headings);
398        assert!(chars.has_lists);
399        assert!(chars.has_links);
400        assert!(chars.has_code);
401        assert!(chars.has_emphasis);
402        assert!(chars.has_html);
403        assert!(chars.has_tables);
404        assert!(chars.has_blockquotes);
405        assert!(chars.has_images);
406    }
407
408    #[test]
409    fn test_content_characteristics_should_skip_rule() {
410        let chars = ContentCharacteristics {
411            has_headings: true,
412            has_lists: false,
413            has_links: true,
414            has_code: false,
415            has_emphasis: true,
416            has_html: false,
417            has_tables: true,
418            has_blockquotes: false,
419            has_images: false,
420        };
421
422        // Create test rules for different categories
423        let heading_rule = MD001HeadingIncrement;
424        assert!(!chars.should_skip_rule(&heading_rule));
425
426        let trailing_spaces_rule = MD009TrailingSpaces::new(2, false);
427        assert!(!chars.should_skip_rule(&trailing_spaces_rule)); // Whitespace rules always run
428
429        // Test skipping based on content
430        let chars_no_headings = ContentCharacteristics {
431            has_headings: false,
432            ..Default::default()
433        };
434        assert!(chars_no_headings.should_skip_rule(&heading_rule));
435    }
436
437    #[test]
438    fn test_lint_empty_content() {
439        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
440
441        let result = lint("", &rules, false);
442        assert!(result.is_ok());
443        assert!(result.unwrap().is_empty());
444    }
445
446    #[test]
447    fn test_lint_with_violations() {
448        let content = "## Level 2\n#### Level 4"; // Skips level 3
449        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
450
451        let result = lint(content, &rules, false);
452        assert!(result.is_ok());
453        let warnings = result.unwrap();
454        assert!(!warnings.is_empty());
455        // Check the rule field of LintWarning struct
456        assert_eq!(warnings[0].rule_name, Some("MD001"));
457    }
458
459    #[test]
460    fn test_lint_with_inline_disable() {
461        let content = "<!-- rumdl-disable MD001 -->\n## Level 2\n#### Level 4";
462        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD001HeadingIncrement)];
463
464        let result = lint(content, &rules, false);
465        assert!(result.is_ok());
466        let warnings = result.unwrap();
467        assert!(warnings.is_empty()); // Should be disabled by inline comment
468    }
469
470    #[test]
471    fn test_lint_rule_filtering() {
472        // Content with no lists
473        let content = "# Heading\nJust text";
474        let rules: Vec<Box<dyn Rule>> = vec![
475            Box::new(MD001HeadingIncrement),
476            // A list-related rule would be skipped
477        ];
478
479        let result = lint(content, &rules, false);
480        assert!(result.is_ok());
481    }
482
483    #[test]
484    fn test_get_profiling_report() {
485        // Just test that it returns a string without panicking
486        let report = get_profiling_report();
487        assert!(!report.is_empty());
488        assert!(report.contains("Profiling"));
489    }
490
491    #[test]
492    fn test_reset_profiling() {
493        // Test that reset_profiling doesn't panic
494        reset_profiling();
495
496        // After reset, report should indicate no measurements or profiling disabled
497        let report = get_profiling_report();
498        assert!(report.contains("disabled") || report.contains("no measurements"));
499    }
500
501    #[test]
502    fn test_get_regex_cache_stats() {
503        let stats = get_regex_cache_stats();
504        // Stats should be a valid HashMap (might be empty)
505        assert!(stats.is_empty() || !stats.is_empty());
506
507        // If not empty, all values should be positive
508        for count in stats.values() {
509            assert!(*count > 0);
510        }
511    }
512
513    #[test]
514    fn test_get_ast_cache_stats() {
515        let stats = get_ast_cache_stats();
516        // Stats should be a valid HashMap (might be empty)
517        assert!(stats.is_empty() || !stats.is_empty());
518
519        // If not empty, all values should be positive
520        for count in stats.values() {
521            assert!(*count > 0);
522        }
523    }
524
525    #[test]
526    fn test_clear_all_caches() {
527        // Test that clear_all_caches doesn't panic
528        clear_all_caches();
529
530        // After clearing, AST cache should be empty
531        let ast_stats = get_ast_cache_stats();
532        assert!(ast_stats.is_empty());
533    }
534
535    #[test]
536    fn test_get_cache_performance_report() {
537        let report = get_cache_performance_report();
538
539        // Report should contain expected sections
540        assert!(report.contains("Cache Performance Report"));
541        assert!(report.contains("Regex Cache:"));
542        assert!(report.contains("AST Cache:"));
543
544        // Test with empty caches
545        clear_all_caches();
546        let report_empty = get_cache_performance_report();
547        assert!(report_empty.contains("No AST nodes cached"));
548    }
549
550    #[test]
551    fn test_lint_with_ast_rules() {
552        // Create content that would benefit from AST parsing
553        let content = "# Heading\n\nParagraph with **bold** text.";
554        let rules: Vec<Box<dyn Rule>> = vec![Box::new(MD012NoMultipleBlanks::new(1))];
555
556        let result = lint(content, &rules, false);
557        assert!(result.is_ok());
558    }
559
560    #[test]
561    fn test_content_characteristics_edge_cases() {
562        // Test setext heading edge case
563        let chars = ContentCharacteristics::analyze("-"); // Single dash, not a heading
564        assert!(!chars.has_headings);
565
566        let chars = ContentCharacteristics::analyze("--"); // Two dashes, valid setext
567        assert!(chars.has_headings);
568
569        // Test list detection edge cases
570        let chars = ContentCharacteristics::analyze("*emphasis*"); // Not a list
571        assert!(!chars.has_lists);
572
573        let chars = ContentCharacteristics::analyze("1.Item"); // No space after period
574        assert!(!chars.has_lists);
575
576        // Test blockquote must be at start of line
577        let chars = ContentCharacteristics::analyze("text > not a quote");
578        assert!(!chars.has_blockquotes);
579    }
580
581    #[test]
582    fn test_cache_performance_report_formatting() {
583        // Add some data to caches to test formatting
584        // (Would require actual usage of the caches, which happens during linting)
585
586        let report = get_cache_performance_report();
587
588        // Test truncation of long patterns
589        // Since we can't easily add a long pattern to the cache in this test,
590        // we'll just verify the report structure is correct
591        assert!(!report.is_empty());
592        assert!(report.lines().count() > 3); // Should have multiple lines
593    }
594}
rumdl/lib.rs

rumdl/
lib.rs
