WIP: improvments to linting

Author: lolbinarycat

src/librustdoc/passes/lint/html_tags.rs

@@ -103,12 +103,7 @@ pub(crate) fn visit_item(cx: &DocContext<'_>, item: &Item, hir_id: HirId, dox: &
         });
     };
 
-    let mut tagp = TagParser{
-        tags: Vec::new(),
-        tag_name: String::with_capacity(8),
-        is_closing: false,
-        tag_start_pos: 0,
-    };
+    let mut tagp = TagParser::new();
     let mut is_in_comment = None;
     let mut in_code_block = false;
 
@@ -257,6 +252,15 @@ struct TagParser {
 }
 
 impl TagParser {
+    fn new() -> Self {
+        Self{
+            tags: Vec::new(),
+            tag_name: String::with_capacity(8),
+            is_closing: false,
+            tag_start_pos: 0,
+        }
+    }
+
     fn drop_tag(
         &mut self,
         range: Range<usize>,
@@ -293,6 +297,19 @@ impl TagParser {
         }
     }
 
+    /// Handle a `<` that appeared while parsing a tag.
+    fn handle_lt_in_tag(&mut self, range: Range<usize>, lt_pos: usize,         f: &impl Fn(String, &Range<usize>, bool),) {
+        let global_pos = range.start + lt_pos;
+        // is this check needed?
+        if global_pos == self.tag_start_pos {
+            // `<` is in the tag because it is the start.
+            return;
+        }
+        // tried to start a new tag while in a tag
+        f(format!("incomplete HTML tag `{}`", &self.tag_name), &(self.tag_start_pos..global_pos), false);
+        self.tag_parsed();
+    }
+
 
     fn extract_html_tag(
         &mut self,
@@ -341,9 +358,7 @@ impl TagParser {
                                         r.end = range.start + new_pos + 1;
                                         found = true;
                                     } else if c == '<' {
-                                        // tried to start a new tag while in a tag
-                                        f(format!("incomplete HTML tag `{}`", &self.tag_name), &(self.tag_start_pos..range.start + start_pos + new_pos), false);
-                                        self.tag_parsed();
+                                        self.handle_lt_in_tag(range.clone(), pos + new_pos, f);
 
                                     }
                                     break;
@@ -371,9 +386,7 @@ impl TagParser {
                                     } else if c == '>' {
                                         break;
                                     } else if c == '<' {
-                                        // tried to start a new tag while in a tag
-                                        f(format!("incomplete HTML tag `{}`", &self.tag_name), &(self.tag_start_pos..range.start + start_pos + i), false);
-                                        self.tag_parsed();
+                                        self.handle_lt_in_tag(range.clone(), pos + i, f);
                                     } else if c == '/' && !after_eq {
                                         is_self_closing = true;
                                     } else {
@@ -454,12 +467,15 @@ impl TagParser {
                         end: range.start + start_pos + 3,
                     });
                 } else {
-                    if !self.tag_name.is_empty() {
+                    if self.tag_name.is_empty() {
+                        self.tag_start_pos = range.start + start_pos;
+                    }
+                    /*if !self.tag_name.is_empty() {
+                        
                         f(format!("incomplete HTML tag `{}`", &self.tag_name), &(self.tag_start_pos..range.start + start_pos), false);
                         self.tag_parsed();
 
-                    }
-                    self.tag_start_pos = range.start + start_pos;
+                    }*/
                     self.extract_html_tag(text, &range, dox, start_pos, &mut iter, f);
                 }
             } else if !self.tag_name.is_empty() {
@@ -470,3 +486,19 @@ impl TagParser {
     }
 
 }
+
+#[test]
+fn test_extract_tags_nested_unclosed() {
+    use std::rc::Rc;
+    use std::cell::RefCell;
+    use std::ops::Deref;
+
+    let mut tagp = TagParser::new();
+    let mut diags = RefCell::new(Vec::new());
+    let dox = "<div>\n<br<div>";
+    tagp.extract_tags(dox, 0..dox.len(), dox, &mut None, &|s, r, b| {
+        diags.borrow_mut().push((s, r.clone(), b));
+    });
+    assert_eq!(diags.borrow().len(), 1, "did not get expected diagnostics: {diags:?}");
+    assert_eq!(diags.borrow()[0].1, 6..9)
+}

tests/rustdoc-ui/lints/invalid-html-tags.rs

@@ -165,6 +165,7 @@ pub fn r() {}
 //~^ ERROR incomplete HTML tag `img`
 pub fn s() {}
 
+/// <br>
 /// <br<br>
 //~^ ERROR incomplete HTML tag `br`
 pub fn t() {}

tests/rustdoc-ui/lints/invalid-html-tags.stderr

@@ -116,7 +116,7 @@ error: incomplete HTML tag `img`
   --> $DIR/invalid-html-tags.rs:158:8
    |
 LL | /// <p><img</p>
-   |        ^
+   |        ^^^^
 
 error: unclosed HTML tag ``
   --> $DIR/invalid-html-tags.rs:158:8
@@ -125,10 +125,22 @@ LL | /// <p><img</p>
    |        ^^^^
 
 error: incomplete HTML tag `br`
-  --> $DIR/invalid-html-tags.rs:173:5
+  --> $DIR/invalid-html-tags.rs:169:5
+   |
+LL | /// <br<br>
+   |     ^^^
+
+error: unclosed HTML tag ``
+  --> $DIR/invalid-html-tags.rs:169:5
+   |
+LL | /// <br<br>
+   |     ^^^
+
+error: incomplete HTML tag `br`
+  --> $DIR/invalid-html-tags.rs:174:5
    |
 LL | /// <br
    |     ^^^
 
-error: aborting due to 21 previous errors
+error: aborting due to 23 previous errors