feat: support multi-line, line-scope comments

errata-ai · May 19, 2024 · d7b0fe1 · d7b0fe1
1 parent cafee06
commit d7b0fe1
Show file tree

Hide file tree

Showing 9 changed files with 105 additions and 167 deletions.
diff --git a/internal/lint/code/comments.go b/internal/lint/code/comments.go
@@ -1,8 +1,10 @@
 package code
 
 import (
+	"bytes"
 	"context"
 	"sort"
+	"strings"
 
 	sitter "github.com/smacker/go-tree-sitter"
 )
@@ -16,6 +18,67 @@ type Comment struct {
 	Scope  string
 }
 
+// doneMerging determines when we should *stop* concatenating line-scoped
+// comments.
+func doneMerging(curr, prev Comment) bool {
+	if prev.Line != curr.Line-1 {
+		// If the comments aren't on consecutive lines, don't merge them.
+		return true
+	} else if prev.Offset != curr.Offset {
+		// If the comments aren't at the same offset, don't merge them.
+		return true
+	}
+	return false
+}
+
+func coalesce(comments []Comment) []Comment {
+	var joined []Comment
+
+	tBuf := bytes.Buffer{}
+	sBuf := bytes.Buffer{}
+
+	for i, comment := range comments {
+		if comment.Scope == "text.comment.block" { //nolint:gocritic
+			joined = append(joined, comment)
+		} else if i == 0 || doneMerging(comment, comments[i-1]) {
+			if tBuf.Len() > 0 {
+				// We have comments to merge ...
+				last := joined[len(joined)-1]
+
+				last.Text += tBuf.String()
+				last.Source += ("\n" + sBuf.String())
+
+				joined[len(joined)-1] = last
+
+				tBuf.Reset()
+				sBuf.Reset()
+			}
+			joined = append(joined, comment)
+		} else {
+			tBuf.WriteString(comment.Text)
+			sBuf.WriteString(comment.Source + "\n")
+		}
+	}
+
+	if tBuf.Len() > 0 {
+		last := joined[len(joined)-1]
+
+		last.Text += tBuf.String()
+		last.Source += ("\n" + sBuf.String())
+
+		joined[len(joined)-1] = last
+
+		tBuf.Reset()
+		sBuf.Reset()
+	}
+
+	for i, comment := range joined {
+		joined[i].Text = strings.TrimLeft(comment.Text, " ")
+	}
+
+	return joined
+}
+
 // GetComments returns all comments in the given source code.
 func GetComments(source []byte, lang *Language) ([]Comment, error) {
 	var comments []Comment
@@ -43,5 +106,5 @@ func GetComments(source []byte, lang *Language) ([]Comment, error) {
 		})
 	}
 
-	return comments, nil
+	return coalesce(comments), nil
 }
diff --git a/internal/lint/code/go.go b/internal/lint/code/go.go
@@ -8,7 +8,7 @@ import (
 
 func Go() *Language {
 	return &Language{
-		Delims:  regexp.MustCompile(`// ?|/\* ?| ?\*/`),
+		Delims:  regexp.MustCompile(`//|/\*|\*/`),
 		Parser:  golang.GetLanguage(),
 		Queries: []string{`(comment)+ @comment`},
 		Padding: cStyle,

diff --git a/internal/lint/code/py.go b/internal/lint/code/py.go
@@ -8,7 +8,7 @@ import (
 
 func Python() *Language {
 	return &Language{
-		Delims: regexp.MustCompile(`#\s?|\s*"""\s*|\s*'''\s*`),
+		Delims: regexp.MustCompile(`#|"""|'''`),
 		Parser: python.GetLanguage(),
 		Queries: []string{
 			`(comment)+ @comment`,
@@ -18,11 +18,11 @@ func Python() *Language {
  (#offset! @docstring 0 3 0 -3))`,
 			// Class docstring
 			`((class_definition
-  body: (block . (expression_statement (string) @rst)))
- (#offset! @rst 0 3 0 -3))`,
+  body: (block . (expression_statement (string) @docstring)))
+ (#offset! @docstring 0 3 0 -3))`,
 			// Module docstring
-			`((module . (expression_statement (string) @rst))
- (#offset! @rst 0 3 0 -3))`,
+			`((module . (expression_statement (string) @docstring))
+ (#offset! @docstring 0 3 0 -3))`,
 		},
 	}
 }
diff --git a/internal/lint/code/query.go b/internal/lint/code/query.go
@@ -1,6 +1,7 @@
 package code
 
 import (
+	"bytes"
 	"strings"
 
 	sitter "github.com/smacker/go-tree-sitter"
@@ -38,6 +39,14 @@ func (qe *QueryEngine) run(q *sitter.Query, source []byte) []Comment {
 			scope := "text.comment.line"
 			if strings.Count(cText, "\n") > 1 {
 				scope = "text.comment.block"
+
+				buf := bytes.Buffer{}
+				for _, line := range strings.Split(cText, "\n") {
+					buf.WriteString(strings.TrimLeft(line, " "))
+					buf.WriteString("\n")
+				}
+
+				cText = buf.String()
 			}
 
 			comments = append(comments, Comment{

diff --git a/internal/lint/code/rs.go b/internal/lint/code/rs.go
@@ -8,7 +8,7 @@ import (
 
 func Rust() *Language {
 	return &Language{
-		Delims:  regexp.MustCompile(`/{2,3}\s?`),
+		Delims:  regexp.MustCompile(`/{2,3}`),
 		Parser:  rust.GetLanguage(),
 		Queries: []string{`(line_comment)+ @comment`},
 	}

diff --git a/testdata/comments/in/2.py b/testdata/comments/in/2.py
@@ -3,7 +3,13 @@
 
 def FIXME():
     """
-    FIXME:
+    FIXME: this is *mardown*.
+
+    ```python
+    print("FIXME: this is *python*.")
+    ```
+
+    New line.
     """
     print("""
 FIXME: This should *not* be linted.

diff --git a/testdata/comments/out/0.json b/testdata/comments/out/0.json
@@ -1,53 +1,18 @@
 [
     {
-        "Text": "\nPackage lint implments Vale's syntax-aware linting functionality.\n\nThe package is split into core linting logic (this file), source code\n(code.go), and markup (markup.go).\n",
+        "Text": "\nPackage lint implments Vale's syntax-aware linting functionality.\n\nThe package is split into core linting logic (this file), source code\n(code.go), and markup (markup.go).\n\n",
         "Source": "/*\nPackage lint implments Vale's syntax-aware linting functionality.\n\nThe package is split into core linting logic (this file), source code\n(code.go), and markup (markup.go).\n*/",
         "Line": 1,
         "Offset": 0,
         "Scope": "text.comment.block"
     },
     {
-        "Text": "Println formats using the default formats for its oprands and writes to",
-        "Source": "// Println formats using the default formats for its oprands and writes to",
+        "Text": "Println formats using the default formats for its oprands and writes to standard output. Spaces are always added between operands and a newline is appended. It returns the number of bytes written and any write error encountered.",
+        "Source": "// Println formats using the default formats for its oprands and writes to\n// standard output.\n//\n// Spaces are always added between operands and a newline is appended.\n//\n// It returns the number of bytes written and any write error encountered.\n",
         "Line": 11,
         "Offset": 0,
         "Scope": "text.comment.line"
     },
-    {
-        "Text": "standard output.",
-        "Source": "// standard output.",
-        "Line": 12,
-        "Offset": 0,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "//",
-        "Line": 13,
-        "Offset": 0,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "Spaces are always added between operands and a newline is appended.",
-        "Source": "// Spaces are always added between operands and a newline is appended.",
-        "Line": 14,
-        "Offset": 0,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "//",
-        "Line": 15,
-        "Offset": 0,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "It returns the number of bytes written and any write error encountered.",
-        "Source": "// It returns the number of bytes written and any write error encountered.",
-        "Line": 16,
-        "Offset": 0,
-        "Scope": "text.comment.line"
-    },
     {
         "Text": "foo bar",
         "Source": "// foo bar",

diff --git a/testdata/comments/out/1.json b/testdata/comments/out/1.json
@@ -14,122 +14,17 @@
         "Scope": "text.comment.line"
     },
     {
-        "Text": "Returns a person with the name given them\n",
-        "Source": "/// Returns a person with the name given them\n",
+        "Text": "Returns a person with the name given them\n\n # Arguments\n\n * `foof` - A string slice that holds the name of the person\n\n # Examples\n\n ```rust\n  You can have rust code between fences inside the comments\n  If you pass --test to `rustdoc`, it will even test it for you!\n use doc::Person;\n let person = Person::new(\"name\");\n ```\n",
+        "Source": "/// Returns a person with the name given them\n\n///\n\n/// # Arguments\n\n///\n\n/// * `foof` - A string slice that holds the name of the person\n\n///\n\n/// # Examples\n\n///\n\n/// ```rust\n\n/// // You can have rust code between fences inside the comments\n\n/// // If you pass --test to `rustdoc`, it will even test it for you!\n\n/// use doc::Person;\n\n/// let person = Person::new(\"name\");\n\n/// ```\n\n",
         "Line": 10,
         "Offset": 4,
         "Scope": "text.comment.line"
     },
     {
-        "Text": "",
-        "Source": "///\n",
-        "Line": 11,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "# Arguments\n",
-        "Source": "/// # Arguments\n",
-        "Line": 12,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "///\n",
-        "Line": 13,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "* `foof` - A string slice that holds the name of the person\n",
-        "Source": "/// * `foof` - A string slice that holds the name of the person\n",
-        "Line": 14,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "///\n",
-        "Line": 15,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "# Examples\n",
-        "Source": "/// # Examples\n",
-        "Line": 16,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "///\n",
-        "Line": 17,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "```rust\n",
-        "Source": "/// ```rust\n",
-        "Line": 18,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "You can have rust code between fences inside the comments\n",
-        "Source": "/// // You can have rust code between fences inside the comments\n",
-        "Line": 19,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "If you pass --test to `rustdoc`, it will even test it for you!\n",
-        "Source": "/// // If you pass --test to `rustdoc`, it will even test it for you!\n",
-        "Line": 20,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "use doc::Person;\n",
-        "Source": "/// use doc::Person;\n",
-        "Line": 21,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "let person = Person::new(\"name\");\n",
-        "Source": "/// let person = Person::new(\"name\");\n",
-        "Line": 22,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "```\n",
-        "Source": "/// ```\n",
-        "Line": 23,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "Gives a friendly hello!\n",
-        "Source": "/// Gives a friendly hello!\n",
+        "Text": "Gives a friendly hello!\n\n Says \"Hello, [name]\" to the `Person` it is called on.\n",
+        "Source": "/// Gives a friendly hello!\n\n///\n\n/// Says \"Hello, [name]\" to the `Person` it is called on.\n\n",
         "Line": 30,
         "Offset": 4,
         "Scope": "text.comment.line"
-    },
-    {
-        "Text": "",
-        "Source": "///\n",
-        "Line": 31,
-        "Offset": 4,
-        "Scope": "text.comment.line"
-    },
-    {
-        "Text": "Says \"Hello, [name]\" to the `Person` it is called on.\n",
-        "Source": "/// Says \"Hello, [name]\" to the `Person` it is called on.\n",
-        "Line": 32,
-        "Offset": 4,
-        "Scope": "text.comment.line"
     }
 ]