Fix Extra Whitespace When Comments Are Within List Initializers

Treat single-line comments that include wrapping newlines as multiline
comments.

Remove the space added around list initialization elements when injected
comments contain newlines.

formatter/format_test.go

@@ -1,6 +1,7 @@
 package formatter
 
 import (
+	"fmt"
 	"strconv"
 	"testing"
 
@@ -22,7 +23,6 @@ func (e *testErrorListener) SyntaxError(_ antlr.Recognizer, _ interface{}, line,
 func TestStatement(t *testing.T) {
 	if testing.Verbose() {
 		log.SetLevel(log.DebugLevel)
-
 	}
 	tests :=
 		[]struct {
@@ -427,7 +427,9 @@ func TestMemberDeclaration(t *testing.T) {
 func TestCompilationUnit(t *testing.T) {
 	if testing.Verbose() {
 		log.SetLevel(log.DebugLevel)
-
+		log.SetFormatter(&log.TextFormatter{
+			DisableQuote: true,
+		})
 	}
 	tests :=
 		[]struct {
@@ -561,27 +563,50 @@ public class Top {
 	/* comment
 	 */
 	public String it;
+}`,
+			},
+			{
+				`class TestClass {
+   List<String> vals = new List<String>{ 
+   // test comment 1
+   'val1', 'val2', 
+   // test comment 2
+   'val3'
+   // test comment 3
+    };
+}`,
+				`class TestClass {
+	List<String> vals = new List<String>{
+	// test comment 1
+	'val1', 'val2',
+	// test comment 2
+	'val3'
+	// test comment 3
+	};
 }`,
 			},
 		}
-	for _, tt := range tests {
-		input := antlr.NewInputStream(tt.input)
-		lexer := parser.NewApexLexer(input)
-		stream := antlr.NewCommonTokenStream(lexer, antlr.TokenDefaultChannel)
+	for i, tt := range tests {
+		t.Run(fmt.Sprintf("%d", i), func(t *testing.T) {
 
-		p := parser.NewApexParser(stream)
-		p.RemoveErrorListeners()
-		p.AddErrorListener(&testErrorListener{t: t})
+			input := antlr.NewInputStream(tt.input)
+			lexer := parser.NewApexLexer(input)
+			stream := antlr.NewCommonTokenStream(lexer, antlr.TokenDefaultChannel)
 
-		v := NewFormatVisitor(stream)
-		out, ok := v.visitRule(p.CompilationUnit()).(string)
-		if !ok {
-			t.Errorf("Unexpected result parsing apex")
-		}
-		out = removeExtraCommentIndentation(out)
-		if out != tt.output {
-			t.Errorf("unexpected format.  expected:\n%q\ngot:\n%q\n", tt.output, out)
-		}
+			p := parser.NewApexParser(stream)
+			p.RemoveErrorListeners()
+			p.AddErrorListener(&testErrorListener{t: t})
+
+			v := NewFormatVisitor(stream)
+			out, ok := v.visitRule(p.CompilationUnit()).(string)
+			if !ok {
+				t.Errorf("Unexpected result parsing apex")
+			}
+			out = removeExtraCommentIndentation(out)
+			if out != tt.output {
+				t.Errorf("unexpected format.  expected:\n%q\ngot:\n%q\n", tt.output, out)
+			}
+		})
 	}
 }
 

formatter/formatter.go

@@ -144,9 +144,13 @@ func removeIndentationFromComment(comment string) string {
 // comment.
 func removeExtraCommentIndentation(input string) string {
 	// Remove extra grammar-specific newlines added unaware of newline-preserving comments injected
-	newlinePrefixedMultilineComment := regexp.MustCompile("\n(\t*\uFFFA)")
+	newlinePrefixedMultilineComment := regexp.MustCompile("[\n ]*(\t*\uFFFA)")
 	input = newlinePrefixedMultilineComment.ReplaceAllString(input, "$1")
 
+	// Remove extra grammar-specific space added unaware of newline-preserving comments injected
+	spacePaddedMultilineComment := regexp.MustCompile(`(` + "\uFFFB\n*\t*" + `) +`)
+	input = spacePaddedMultilineComment.ReplaceAllString(input, "$1")
+
 	newlinePrefixedInlineComment := regexp.MustCompile("\n\t*\uFFF9\n")
 	input = newlinePrefixedInlineComment.ReplaceAllString(input, "\uFFF9\n")
 

formatter/visitor.go

@@ -338,8 +338,8 @@ func appendHiddenTokens(v *FormatVisitor, result interface{}, tokens []antlr.Tok
 					trailing = " "
 				}
 				// Strip leading whitespace so the comment can be indented to the right location
-				text = strings.TrimSpace(text)
 				containsNewline := strings.Contains(text, "\n")
+				text = strings.TrimSpace(text)
 
 				text = fmt.Sprintf("%s%s%s", leading, text, trailing)
 				log.Debug(fmt.Sprintf("NORMALIZED COMMENT: %q\n", text))
@@ -353,7 +353,7 @@ func appendHiddenTokens(v *FormatVisitor, result interface{}, tokens []antlr.Tok
 				text = "\n" // Replace multiple blank lines with a single blank line
 			} else {
 				// whitespace to ignore
-				text = ""
+				continue
 			}
 
 			tokenLines = append(tokenLines, text)