cmd/compile/internal/syntax: differentiate between ';' and '\n' in syntax errors

Towards better syntax error messages: With this change, the parser knows whether
a semicolon was an actual ';' in the source, or whether it was an automatically
inserted semicolon as result of a '\n' or EOF. Using this information in error
messages makes them more understandable.

For #17328.

Change-Id: I8cd9accee8681b62569d0ecef922d38682b401eb
Reviewed-on: https://go-review.googlesource.com/36636
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

3 files changed, 15 insertions, 3 deletions

diff --git a/src/cmd/compile/internal/syntax/parser.go b/src/cmd/compile/internal/syntax/parser.go
index 3ebc670950..f016a6adbe 100644
--- a/src/cmd/compile/internal/syntax/parser.go
+++ b/src/cmd/compile/internal/syntax/parser.go
@@ -144,7 +144,7 @@ func (p *parser) syntax_error_at(pos src.Pos, msg string) {
 	// determine token string
 	var tok string
 	switch p.tok {
-	case _Name:
+	case _Name, _Semi:
 		tok = p.lit
 	case _Literal:
 		tok = "literal " + p.lit
@@ -215,7 +215,7 @@ func tokstring(tok token) string {
 	case _Comma:
 		return "comma"
 	case _Semi:
-		return "semicolon or newline"
+		return "semicolon"
 	}
 	return tok.String()
 }
diff --git a/src/cmd/compile/internal/syntax/scanner.go b/src/cmd/compile/internal/syntax/scanner.go
index 149827b21c..ede3b00a34 100644
--- a/src/cmd/compile/internal/syntax/scanner.go
+++ b/src/cmd/compile/internal/syntax/scanner.go
@@ -27,7 +27,7 @@ type scanner struct {
 	// current token, valid after calling next()
 	line, col uint
 	tok       token
-	lit       string   // valid if tok is _Name or _Literal
+	lit       string   // valid if tok is _Name, _Literal, or _Semi ("semicolon", "newline", or "EOF")
 	kind      LitKind  // valid if tok is _Literal
 	op        Operator // valid if tok is _Operator, _AssignOp, or _IncOp
 	prec      int      // valid if tok is _Operator, _AssignOp, or _IncOp
@@ -73,12 +73,14 @@ redo:
 	switch c {
 	case -1:
 		if nlsemi {
+			s.lit = "EOF"
 			s.tok = _Semi
 			break
 		}
 		s.tok = _EOF
 
 	case '\n':
+		s.lit = "newline"
 		s.tok = _Semi
 
 	case '0', '1', '2', '3', '4', '5', '6', '7', '8', '9':
@@ -106,6 +108,7 @@ redo:
 		s.tok = _Comma
 
 	case ';':
+		s.lit = "semicolon"
 		s.tok = _Semi
 
 	case ')':
@@ -185,6 +188,7 @@ redo:
 			if s.source.line > s.line && nlsemi {
 				// A multi-line comment acts like a newline;
 				// it translates to a ';' if nlsemi is set.
+				s.lit = "newline"
 				s.tok = _Semi
 				break
 			}
diff --git a/src/cmd/compile/internal/syntax/scanner_test.go b/src/cmd/compile/internal/syntax/scanner_test.go
index fcdc3348e4..ed240df069 100644
--- a/src/cmd/compile/internal/syntax/scanner_test.go
+++ b/src/cmd/compile/internal/syntax/scanner_test.go
@@ -66,6 +66,11 @@ func TestTokens(t *testing.T) {
 		}
 
 		switch want.tok {
+		case _Semi:
+			if got.lit != "semicolon" {
+				t.Errorf("got %s; want semicolon", got.lit)
+			}
+
 		case _Name, _Literal:
 			if got.lit != want.src {
 				t.Errorf("got lit = %q; want %q", got.lit, want.src)
@@ -94,6 +99,9 @@ func TestTokens(t *testing.T) {
 				t.Errorf("got tok = %s; want ;", got.tok)
 				continue
 			}
+			if got.lit != "newline" {
+				t.Errorf("got %s; want newline", got.lit)
+			}
 		}
 
 		got.next()