cmd/compile: rationalize (lex)?lineno handling

Previously, many error messages inconsistantly used either lexlineno
and lineno.  In general this works out okay because they're almost
always the same.  The only exceptional case is after lexing a
multi-line raw string literal, where lineno will be the line number of
the opening quote and lexlineno is the line number of the closing
quote.

This CL makes the compiler's error message more consistent:

- Lexer error messages related to invalid byte sequences (i.e., NUL
bytes, bad UTF-8 sequences, and non-initial BOMs) are emitted at
lexlineno (i.e., the source line that contains the invalid byte
sequence).

- All other error messages (notably the parser's "syntax errors") now
use lineno.  The minor change from this is that bogus input like:

    package `
    bogus`

will emit "syntax error: unexpected string literal, expecting name"
error at line 1, instead of line 2.

- Instead of maintaining prevlineno all the time, just record it
when/where actually needed and not already available elsewhere (which
turns out to be just one function).

- Lastly, we remove the legacy "syntax error near ..." fallback in
Yerror, now that the parser always emits more detailed syntax error
messages.

Change-Id: Iaf5f784223d0385fa3a5b09ef2b2ad447feab02f
Reviewed-on: https://go-review.googlesource.com/19925Reviewed-by: Robert Griesemer <gri@golang.org>

src/cmd/compile/internal/gc/go.go

@@ -425,12 +425,13 @@ var sizeof_String int // runtime sizeof(String)
 
 var dotlist [10]Dlist // size is max depth of embeddeds
 
+// lexlineno is the line number _after_ the most recently read rune.
+// In particular, it's advanced (or rewound) as newlines are read (or unread).
 var lexlineno int32
 
+// lineno is the line number at the start of the most recently lexed token.
 var lineno int32
 
-var prevlineno int32
-
 var pragcgobuf string
 
 var infile string

src/cmd/compile/internal/gc/lex.go

@@ -923,8 +923,6 @@ const (
 )
 
 func (l *lexer) next() {
-	prevlineno = lineno
-
 	nlsemi := l.nlsemi
 	l.nlsemi = false
 
@@ -933,12 +931,12 @@ l0:
 	c := l.getr()
 	for isSpace(c) {
 		if c == '\n' && nlsemi {
-			// TODO(gri) we may be able avoid the ungetr and simply use lexlineno-1 below
-			l.ungetr(c) // for correct line number
 			if Debug['x'] != 0 {
 				fmt.Printf("lex: implicit semi\n")
 			}
-			lineno = lexlineno
+			// Insert implicit semicolon on previous line,
+			// before the newline character.
+			lineno = lexlineno - 1
 			l.tok = ';'
 			return
 		}
@@ -1231,9 +1229,9 @@ l0:
 lx:
 	if Debug['x'] != 0 {
 		if c > 0xff {
-			fmt.Printf("%v lex: TOKEN %s\n", Ctxt.Line(int(lexlineno)), lexname(c))
+			fmt.Printf("%v lex: TOKEN %s\n", Ctxt.Line(int(lineno)), lexname(c))
 		} else {
-			fmt.Printf("%v lex: TOKEN '%c'\n", Ctxt.Line(int(lexlineno)), c)
+			fmt.Printf("%v lex: TOKEN '%c'\n", Ctxt.Line(int(lineno)), c)
 		}
 	}
 
@@ -1850,8 +1848,7 @@ redo:
 	c := obj.Bgetc(l.bin)
 	if c < utf8.RuneSelf {
 		if c == 0 {
-			// TODO(gri) do we need lineno = lexlineno here? Why not?
-			Yyerror("illegal NUL byte")
+			yyerrorl(int(lexlineno), "illegal NUL byte")
 			return 0
 		}
 		if c == '\n' && importpkg == nil {
@@ -1872,15 +1869,13 @@ redo:
 
 	r, w := utf8.DecodeRune(buf[:i])
 	if r == utf8.RuneError && w == 1 {
-		lineno = lexlineno
 		// The string conversion here makes a copy for passing
 		// to fmt.Printf, so that buf itself does not escape and
 		// can be allocated on the stack.
-		Yyerror("illegal UTF-8 sequence % x", string(buf[:i]))
+		yyerrorl(int(lexlineno), "illegal UTF-8 sequence % x", string(buf[:i]))
 	}
 
 	if r == BOM {
-		// TODO(gri) can we use Yyerror here? Why not?
 		yyerrorl(int(lexlineno), "Unicode (UTF-8) BOM in middle of file")
 		goto redo
 	}

src/cmd/compile/internal/gc/parser.go

@@ -110,11 +110,11 @@ func (p *parser) syntax_error(msg string) {
 }
 
 // Like syntax_error, but reports error at given line rather than current lexer line.
-func (p *parser) syntax_error_at(lineno int32, msg string) {
-	defer func(lineno int32) {
-		lexlineno = lineno
-	}(lexlineno)
-	lexlineno = lineno
+func (p *parser) syntax_error_at(lno int32, msg string) {
+	defer func(lno int32) {
+		lineno = lno
+	}(lineno)
+	lineno = lno
 	p.syntax_error(msg)
 }
 
@@ -687,7 +687,7 @@ func (p *parser) labeled_stmt(label *Node) *Node {
 		ls = p.stmt()
 		if ls == missing_stmt {
 			// report error at line of ':' token
-			p.syntax_error_at(prevlineno, "missing statement after label")
+			p.syntax_error_at(label.Lineno, "missing statement after label")
 			// we are already at the end of the labeled statement - no need to advance
 			return missing_stmt
 		}
@@ -1609,13 +1609,15 @@ func (p *parser) new_name(sym *Sym) *Node {
 	return nil
 }
 
-func (p *parser) dcl_name(sym *Sym) *Node {
+func (p *parser) dcl_name() *Node {
 	if trace && Debug['x'] != 0 {
 		defer p.trace("dcl_name")()
 	}
 
+	symlineno := lineno
+	sym := p.sym()
 	if sym == nil {
-		yyerrorl(int(prevlineno), "invalid declaration")
+		yyerrorl(int(symlineno), "invalid declaration")
 		return nil
 	}
 	return dclname(sym)
@@ -2637,9 +2639,9 @@ func (p *parser) dcl_name_list() *NodeList {
 		defer p.trace("dcl_name_list")()
 	}
 
-	l := list1(p.dcl_name(p.sym()))
+	l := list1(p.dcl_name())
 	for p.got(',') {
-		l = list(l, p.dcl_name(p.sym()))
+		l = list(l, p.dcl_name())
 	}
 	return l
 }

src/cmd/compile/internal/gc/subr.go

@@ -109,7 +109,7 @@ func yyerrorl(line int, format string, args ...interface{}) {
 	}
 }
 
-var yyerror_lastsyntax int
+var yyerror_lastsyntax int32
 
 func Yyerror(format string, args ...interface{}) {
 	msg := fmt.Sprintf(format, args...)
@@ -117,18 +117,12 @@ func Yyerror(format string, args ...interface{}) {
 		nsyntaxerrors++
 
 		// only one syntax error per line
-		if int32(yyerror_lastsyntax) == lexlineno {
+		if yyerror_lastsyntax == lineno {
 			return
 		}
-		yyerror_lastsyntax = int(lexlineno)
+		yyerror_lastsyntax = lineno
 
-		// plain "syntax error" gets "near foo" added
-		if msg == "syntax error" {
-			yyerrorl(int(lexlineno), "syntax error near %s", lexbuf.String())
-			return
-		}
-
-		yyerrorl(int(lexlineno), "%s", msg)
+		yyerrorl(int(lineno), "%s", msg)
 		return
 	}