fix(MD013): preserve list marker spacing to prevent corruption of nested code blocks

src/rules/md013_line_length/helpers.rs

@@ -123,16 +123,25 @@ pub(crate) fn extract_list_marker_and_content(line: &str) -> (String, String) {
     for bullet in ["- ", "* ", "+ "] {
         if let Some(rest) = trimmed.strip_prefix(bullet) {
             let marker_prefix = &bullet[..bullet.len() - 1]; // "-", "*", or "+"
+
+            // Hoist space extraction and exclude tabs
+            let spaces_len = rest.len() - rest.trim_start_matches(' ').len();
+            let extra_spaces = &rest[..spaces_len];
+            let content = &rest[spaces_len..];
+
             // Include GFM task list checkboxes in the non-wrappable marker prefix
             for checkbox in ["[ ] ", "[x] ", "[X] "] {
-                if let Some(content) = rest.strip_prefix(checkbox) {
+                if let Some(actual_content) = content.strip_prefix(checkbox) {
                     return (
-                        format!("{indent}{marker_prefix} {checkbox}"),
-                        trim_preserving_hard_break(content),
+                        format!("{indent}{marker_prefix} {extra_spaces}{checkbox}"),
+                        trim_preserving_hard_break(actual_content),
                     );
                 }
             }
-            return (format!("{indent}{bullet}"), trim_preserving_hard_break(rest));
+            return (
+                format!("{indent}{marker_prefix} {extra_spaces}"),
+                trim_preserving_hard_break(content),
+            );
         }
     }
 
@@ -149,17 +158,25 @@ pub(crate) fn extract_list_marker_and_content(line: &str) -> (String, String) {
             {
                 marker_content.push(next);
                 let rest = chars.as_str();
+
+                // Hoist space extraction and exclude tabs
+                let spaces_len = rest.len() - rest.trim_start_matches(' ').len();
+                let extra_spaces = &rest[..spaces_len];
+                let content = &rest[spaces_len..];
+
                 // Check for GFM task list checkboxes
                 for checkbox in ["[ ] ", "[x] ", "[X] "] {
-                    if let Some(content) = rest.strip_prefix(checkbox) {
+                    if let Some(actual_content) = content.strip_prefix(checkbox) {
                         return (
-                            format!("{indent}{marker_content}{checkbox}"),
-                            trim_preserving_hard_break(content),
+                            format!("{indent}{marker_content}{extra_spaces}{checkbox}"),
+                            trim_preserving_hard_break(actual_content),
                         );
                     }
                 }
-                let content = trim_preserving_hard_break(rest);
-                return (format!("{indent}{marker_content}"), content);
+                return (
+                    format!("{indent}{marker_content}{extra_spaces}"),
+                    trim_preserving_hard_break(content),
+                );
             }
             break;
         }
@@ -538,6 +555,34 @@ mod tests {
             extract_list_marker_and_content("99. [x] multi-digit ordered"),
             ("99. [x] ".to_string(), "multi-digit ordered".to_string())
         );
+
+        // Extra spaces preservation
+        assert_eq!(
+            extract_list_marker_and_content("-   item"),
+            ("-   ".to_string(), "item".to_string())
+        );
+        assert_eq!(
+            extract_list_marker_and_content("1.   item"),
+            ("1.   ".to_string(), "item".to_string())
+        );
+        assert_eq!(
+            extract_list_marker_and_content("-   [ ] item"),
+            ("-   [ ] ".to_string(), "item".to_string())
+        );
+        assert_eq!(
+            extract_list_marker_and_content("1.   [ ] item"),
+            ("1.   [ ] ".to_string(), "item".to_string())
+        );
+
+        // Tabs should not be preserved in marker
+        assert_eq!(
+            extract_list_marker_and_content("- \titem"),
+            ("- ".to_string(), "\titem".to_string())
+        );
+        assert_eq!(
+            extract_list_marker_and_content("1. \titem"),
+            ("1. ".to_string(), "\titem".to_string())
+        );
     }
 
     #[test]

src/rules/md013_line_length/tests.rs

@@ -1043,6 +1043,47 @@ And a bullet list:
     }
 }
 
+#[test]
+fn test_text_reflow_preserves_nested_code_blocks_in_lists() {
+    let config = MD013Config {
+        line_length: crate::types::LineLength::from_const(40),
+        reflow: true,
+        ..Default::default()
+    };
+    let mut rule = MD013LineLength::from_config_struct(config);
+    rule.list_spacing = MD030Config {
+        ul_single: crate::types::PositiveUsize::from_const(3),
+        ul_multi: crate::types::PositiveUsize::from_const(3),
+        ..Default::default()
+    };
+
+    let content = indoc! {"
+        -   This is a very long list item text that will definitely exceed the forty character limit and force a reflow.
+
+            ```
+            code block line 1
+            code block line 2
+            ```
+    "};
+
+    let ctx = LintContext::new(content, crate::config::MarkdownFlavor::Standard, None);
+    let fixed = rule.fix(&ctx).unwrap();
+
+    let expected = indoc! {"
+        -   This is a very long list item text
+            that will definitely exceed the
+            forty character limit and force a
+            reflow.
+
+            ```
+            code block line 1
+            code block line 2
+            ```
+    "};
+
+    assert_eq!(fixed, expected);
+}
+
 #[test]
 fn test_issue_83_numbered_list_with_backticks() {
     // Test for issue #83: enable_reflow was incorrectly handling numbered lists

tests/formats/md013_mkdocs_reflow_test.rs

@@ -1332,7 +1332,7 @@ fn test_issue_509_exact_repro_4space_list_admonition_code_block() {
     let expected = "\
 # Test
 
-1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+1.  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note
@@ -1372,7 +1372,7 @@ fn test_issue_509_unordered_list_variant() {
     let fixed = rule.fix(&ctx).unwrap();
 
     let expected = "\
-- Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+-   Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note
@@ -1413,7 +1413,7 @@ fn test_issue_509_admonition_code_block_long_text_after_4space() {
     let fixed = rule.fix(&ctx).unwrap();
 
     let expected = "\
-1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+1.  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note
@@ -1496,7 +1496,7 @@ fn test_issue_509_code_block_with_blank_lines_and_deep_indent() {
     // MD030 normalizes "1.  " to "1. " — only the overlong line and
     // marker spacing should change, everything else verbatim.
     let expected = "\
-1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+1.  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note
@@ -1546,7 +1546,7 @@ fn test_issue_509_mixed_fence_types_not_confused() {
     let fixed = rule.fix(&ctx).unwrap();
 
     let expected = "\
-1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+1.  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note
@@ -1595,7 +1595,7 @@ fn test_issue_509_info_string_not_treated_as_closing_fence() {
     let fixed = rule.fix(&ctx).unwrap();
 
     let expected = "\
-1. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
+1.  Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod
     tempor incididunt ut labore et dolore magna aliqua.
 
     !!! note