- remove special handling of '\n' characters (used to be treated as comments

for pretty printer purposes - now properly ignored as white space since we
have line/col information)
- changed sample use in comment to an actually compiled function to make sure
sample is actually working
- added extra tests (checking line and column values, and the tokenize function)

R=rsc
DELTA=253  (61 added, 67 deleted, 125 changed)
OCL=26143
CL=26181

diff --git a/src/lib/go/scanner.go b/src/lib/go/scanner.go
index ccac8e111285115e95ed8a34c3fb35420ec0ce72..f665f10babc25b724e0cce4be1dfa90b95374695 100644 (file)
--- a/src/lib/go/scanner.go
+++ b/src/lib/go/scanner.go
@@ -4,23 +4,7 @@
 
 // A scanner for Go source text. Takes a []byte as source which can
 // then be tokenized through repeated calls to the Scan function.
-//
-// Sample use:
-//
-//     import "token"
-//     import "scanner"
-//
-//     func tokenize(src []byte) {
-//             var s scanner.Scanner;
-//             s.Init(src, nil /* no error handler */, false /* ignore comments */);
-//             for {
-//                     pos, tok, lit := s.Scan();
-//                     if tok == Scanner.EOF {
-//                             return;
-//                     }
-//                     println(pos, token.TokenString(tok), string(lit));
-//             }
-//     }
+// For a sample use of a scanner, see the implementation of Tokenize.
 //
 package scanner
 
@@ -62,7 +46,7 @@ type Scanner struct {
        scan_comments bool;  // if set, comments are reported as tokens
 
        // scanning state
-       loc Location;  // location of ch
+       loc Location;  // location before ch (src[loc.Pos] == ch)
        pos int;  // current reading position (position after ch)
        ch int;  // one char look-ahead
 }
@@ -78,7 +62,7 @@ func (S *Scanner) next() {
                switch {
                case r == '\n':
                        S.loc.Line++;
-                       S.loc.Col = 1;
+                       S.loc.Col = 0;
                case r >= 0x80:
                        // not ASCII
                        r, w = utf8.DecodeRune(S.src[S.pos : len(S.src)]);
@@ -94,9 +78,9 @@ func (S *Scanner) next() {
 
 // Init prepares the scanner S to tokenize the text src. Calls to Scan
 // will use the error handler err if they encounter a syntax error. The boolean
-// scan_comments specifies whether newline characters and comments should be
-// recognized and returned by Scan as token.COMMENT. If scan_comments is false,
-// they are treated as white space and ignored.
+// scan_comments specifies whether comments should be recognized and returned
+// by Scan as token.COMMENT. If scan_comments is false, they are treated as
+// white space and ignored.
 //
 func (S *Scanner) Init(src []byte, err ErrorHandler, scan_comments bool) {
        S.src = src;
@@ -137,24 +121,6 @@ func (S *Scanner) expect(ch int) {
 }
 
 
-func (S *Scanner) skipWhitespace() {
-       for {
-               switch S.ch {
-               case '\t', '\r', ' ':
-                       // nothing to do
-               case '\n':
-                       if S.scan_comments {
-                               return;
-                       }
-               default:
-                       return;
-               }
-               S.next();
-       }
-       panic("UNREACHABLE");
-}
-
-
 func (S *Scanner) scanComment(loc Location) {
        // first '/' already consumed
 
@@ -163,9 +129,7 @@ func (S *Scanner) scanComment(loc Location) {
                for S.ch >= 0 {
                        S.next();
                        if S.ch == '\n' {
-                               // '\n' terminates comment but we do not include
-                               // it in the comment (otherwise we don't see the
-                               // start of a newline in skipWhitespace()).
+                               S.next();  // '\n' belongs to the comment
                                return;
                        }
                }
@@ -412,14 +376,19 @@ func (S *Scanner) switch4(tok0, tok1, ch2, tok2, tok3 int) int {
 
 // Scan scans the next token and returns the token location loc,
 // the token tok, and the literal text lit corresponding to the
-// token.
+// token. The source end is indicated by token.EOF.
 //
 func (S *Scanner) Scan() (loc Location, tok int, lit []byte) {
 scan_again:
-       S.skipWhitespace();
+       // skip white space
+       for S.ch == ' ' || S.ch == '\t' || S.ch == '\n' || S.ch == '\r' {
+               S.next();
+       }
 
+       // current token start
        loc, tok = S.loc, token.ILLEGAL;
 
+       // determine token value
        switch ch := S.ch; {
        case isLetter(ch):
                tok = S.scanIdentifier();
@@ -429,7 +398,6 @@ scan_again:
                S.next();  // always make progress
                switch ch {
                case -1  : tok = token.EOF;
-               case '\n': tok = token.COMMENT;
                case '"' : tok = token.STRING; S.scanString(loc);
                case '\'': tok = token.CHAR; S.scanChar();
                case '`' : tok = token.STRING; S.scanRawString(loc);
@@ -487,3 +455,17 @@ scan_again:
 
        return loc, tok, S.src[loc.Pos : S.loc.Pos];
 }
+
+
+// Tokenize calls a function f with the token location, token value, and token
+// text for each token in the source src. The other parameters have the same
+// meaning as for the Init function. Tokenize keeps scanning until f returns
+// false (usually when the token value is token.EOF).
+//
+func Tokenize(src []byte, err ErrorHandler, scan_comments bool, f func (loc Location, tok int, lit []byte) bool) {
+       var s Scanner;
+       s.Init(src, err, scan_comments);
+       for f(s.Scan()) {
+               // action happens in f
+       }
+}

diff --git a/src/lib/go/scanner_test.go b/src/lib/go/scanner_test.go
index 94c2e51d536278bee8a4a23e50394a98cfb4a778..221f01256e4704e0c0c3a33aee0f8e721d0945c5 100644 (file)
--- a/src/lib/go/scanner_test.go
+++ b/src/lib/go/scanner_test.go
@@ -31,7 +31,6 @@ func tokenclass(tok int) int {
 
 
 type elt struct {
-       pos int;
        tok int;
        lit string;
        class int;
@@ -40,130 +39,120 @@ type elt struct {
 
 var tokens = [...]elt{
        // Special tokens
-       elt{ 0, token.COMMENT, "/* a comment */", special },
-       elt{ 0, token.COMMENT, "\n", special },
+       elt{ token.COMMENT, "/* a comment */", special },
+       elt{ token.COMMENT, "// a comment \n", special },
 
        // Identifiers and basic type literals
-       elt{ 0, token.IDENT, "foobar", literal },
-       elt{ 0, token.IDENT, "a۰۱۸", literal },
-       elt{ 0, token.IDENT, "foo६४", literal },
-       elt{ 0, token.IDENT, "bar９８７６", literal },
-       elt{ 0, token.INT, "0", literal },
-       elt{ 0, token.INT, "01234567", literal },
-       elt{ 0, token.INT, "0xcafebabe", literal },
-       elt{ 0, token.FLOAT, "0.", literal },
-       elt{ 0, token.FLOAT, ".0", literal },
-       elt{ 0, token.FLOAT, "3.14159265", literal },
-       elt{ 0, token.FLOAT, "1e0", literal },
-       elt{ 0, token.FLOAT, "1e+100", literal },
-       elt{ 0, token.FLOAT, "1e-100", literal },
-       elt{ 0, token.FLOAT, "2.71828e-1000", literal },
-       elt{ 0, token.CHAR, "'a'", literal },
-       elt{ 0, token.CHAR, "'\\000'", literal },
-       elt{ 0, token.CHAR, "'\\xFF'", literal },
-       elt{ 0, token.CHAR, "'\\uff16'", literal },
-       elt{ 0, token.CHAR, "'\\U0000ff16'", literal },
-       elt{ 0, token.STRING, "`foobar`", literal },
+       elt{ token.IDENT, "foobar", literal },
+       elt{ token.IDENT, "a۰۱۸", literal },
+       elt{ token.IDENT, "foo६४", literal },
+       elt{ token.IDENT, "bar９８７６", literal },
+       elt{ token.INT, "0", literal },
+       elt{ token.INT, "01234567", literal },
+       elt{ token.INT, "0xcafebabe", literal },
+       elt{ token.FLOAT, "0.", literal },
+       elt{ token.FLOAT, ".0", literal },
+       elt{ token.FLOAT, "3.14159265", literal },
+       elt{ token.FLOAT, "1e0", literal },
+       elt{ token.FLOAT, "1e+100", literal },
+       elt{ token.FLOAT, "1e-100", literal },
+       elt{ token.FLOAT, "2.71828e-1000", literal },
+       elt{ token.CHAR, "'a'", literal },
+       elt{ token.CHAR, "'\\000'", literal },
+       elt{ token.CHAR, "'\\xFF'", literal },
+       elt{ token.CHAR, "'\\uff16'", literal },
+       elt{ token.CHAR, "'\\U0000ff16'", literal },
+       elt{ token.STRING, "`foobar`", literal },
 
        // Operators and delimitors
-       elt{ 0, token.ADD, "+", operator },
-       elt{ 0, token.SUB, "-", operator },
-       elt{ 0, token.MUL, "*", operator },
-       elt{ 0, token.QUO, "/", operator },
-       elt{ 0, token.REM, "%", operator },
-
-       elt{ 0, token.AND, "&", operator },
-       elt{ 0, token.OR, "|", operator },
-       elt{ 0, token.XOR, "^", operator },
-       elt{ 0, token.SHL, "<<", operator },
-       elt{ 0, token.SHR, ">>", operator },
-
-       elt{ 0, token.ADD_ASSIGN, "+=", operator },
-       elt{ 0, token.SUB_ASSIGN, "-=", operator },
-       elt{ 0, token.MUL_ASSIGN, "*=", operator },
-       elt{ 0, token.QUO_ASSIGN, "/=", operator },
-       elt{ 0, token.REM_ASSIGN, "%=", operator },
-
-       elt{ 0, token.AND_ASSIGN, "&=", operator },
-       elt{ 0, token.OR_ASSIGN, "|=", operator },
-       elt{ 0, token.XOR_ASSIGN, "^=", operator },
-       elt{ 0, token.SHL_ASSIGN, "<<=", operator },
-       elt{ 0, token.SHR_ASSIGN, ">>=", operator },
-
-       elt{ 0, token.LAND, "&&", operator },
-       elt{ 0, token.LOR, "||", operator },
-       elt{ 0, token.ARROW, "<-", operator },
-       elt{ 0, token.INC, "++", operator },
-       elt{ 0, token.DEC, "--", operator },
-
-       elt{ 0, token.EQL, "==", operator },
-       elt{ 0, token.LSS, "<", operator },
-       elt{ 0, token.GTR, ">", operator },
-       elt{ 0, token.ASSIGN, "=", operator },
-       elt{ 0, token.NOT, "!", operator },
-
-       elt{ 0, token.NEQ, "!=", operator },
-       elt{ 0, token.LEQ, "<=", operator },
-       elt{ 0, token.GEQ, ">=", operator },
-       elt{ 0, token.DEFINE, ":=", operator },
-       elt{ 0, token.ELLIPSIS, "...", operator },
-
-       elt{ 0, token.LPAREN, "(", operator },
-       elt{ 0, token.LBRACK, "[", operator },
-       elt{ 0, token.LBRACE, "{", operator },
-       elt{ 0, token.COMMA, ",", operator },
-       elt{ 0, token.PERIOD, ".", operator },
-
-       elt{ 0, token.RPAREN, ")", operator },
-       elt{ 0, token.RBRACK, "]", operator },
-       elt{ 0, token.RBRACE, "}", operator },
-       elt{ 0, token.SEMICOLON, ";", operator },
-       elt{ 0, token.COLON, ":", operator },
+       elt{ token.ADD, "+", operator },
+       elt{ token.SUB, "-", operator },
+       elt{ token.MUL, "*", operator },
+       elt{ token.QUO, "/", operator },
+       elt{ token.REM, "%", operator },
+
+       elt{ token.AND, "&", operator },
+       elt{ token.OR, "|", operator },
+       elt{ token.XOR, "^", operator },
+       elt{ token.SHL, "<<", operator },
+       elt{ token.SHR, ">>", operator },
+
+       elt{ token.ADD_ASSIGN, "+=", operator },
+       elt{ token.SUB_ASSIGN, "-=", operator },
+       elt{ token.MUL_ASSIGN, "*=", operator },
+       elt{ token.QUO_ASSIGN, "/=", operator },
+       elt{ token.REM_ASSIGN, "%=", operator },
+
+       elt{ token.AND_ASSIGN, "&=", operator },
+       elt{ token.OR_ASSIGN, "|=", operator },
+       elt{ token.XOR_ASSIGN, "^=", operator },
+       elt{ token.SHL_ASSIGN, "<<=", operator },
+       elt{ token.SHR_ASSIGN, ">>=", operator },
+
+       elt{ token.LAND, "&&", operator },
+       elt{ token.LOR, "||", operator },
+       elt{ token.ARROW, "<-", operator },
+       elt{ token.INC, "++", operator },
+       elt{ token.DEC, "--", operator },
+
+       elt{ token.EQL, "==", operator },
+       elt{ token.LSS, "<", operator },
+       elt{ token.GTR, ">", operator },
+       elt{ token.ASSIGN, "=", operator },
+       elt{ token.NOT, "!", operator },
+
+       elt{ token.NEQ, "!=", operator },
+       elt{ token.LEQ, "<=", operator },
+       elt{ token.GEQ, ">=", operator },
+       elt{ token.DEFINE, ":=", operator },
+       elt{ token.ELLIPSIS, "...", operator },
+
+       elt{ token.LPAREN, "(", operator },
+       elt{ token.LBRACK, "[", operator },
+       elt{ token.LBRACE, "{", operator },
+       elt{ token.COMMA, ",", operator },
+       elt{ token.PERIOD, ".", operator },
+
+       elt{ token.RPAREN, ")", operator },
+       elt{ token.RBRACK, "]", operator },
+       elt{ token.RBRACE, "}", operator },
+       elt{ token.SEMICOLON, ";", operator },
+       elt{ token.COLON, ":", operator },
 
        // Keywords
-       elt{ 0, token.BREAK, "break", keyword },
-       elt{ 0, token.CASE, "case", keyword },
-       elt{ 0, token.CHAN, "chan", keyword },
-       elt{ 0, token.CONST, "const", keyword },
-       elt{ 0, token.CONTINUE, "continue", keyword },
-
-       elt{ 0, token.DEFAULT, "default", keyword },
-       elt{ 0, token.DEFER, "defer", keyword },
-       elt{ 0, token.ELSE, "else", keyword },
-       elt{ 0, token.FALLTHROUGH, "fallthrough", keyword },
-       elt{ 0, token.FOR, "for", keyword },
-
-       elt{ 0, token.FUNC, "func", keyword },
-       elt{ 0, token.GO, "go", keyword },
-       elt{ 0, token.GOTO, "goto", keyword },
-       elt{ 0, token.IF, "if", keyword },
-       elt{ 0, token.IMPORT, "import", keyword },
-
-       elt{ 0, token.INTERFACE, "interface", keyword },
-       elt{ 0, token.MAP, "map", keyword },
-       elt{ 0, token.PACKAGE, "package", keyword },
-       elt{ 0, token.RANGE, "range", keyword },
-       elt{ 0, token.RETURN, "return", keyword },
-
-       elt{ 0, token.SELECT, "select", keyword },
-       elt{ 0, token.STRUCT, "struct", keyword },
-       elt{ 0, token.SWITCH, "switch", keyword },
-       elt{ 0, token.TYPE, "type", keyword },
-       elt{ 0, token.VAR, "var", keyword },
+       elt{ token.BREAK, "break", keyword },
+       elt{ token.CASE, "case", keyword },
+       elt{ token.CHAN, "chan", keyword },
+       elt{ token.CONST, "const", keyword },
+       elt{ token.CONTINUE, "continue", keyword },
+
+       elt{ token.DEFAULT, "default", keyword },
+       elt{ token.DEFER, "defer", keyword },
+       elt{ token.ELSE, "else", keyword },
+       elt{ token.FALLTHROUGH, "fallthrough", keyword },
+       elt{ token.FOR, "for", keyword },
+
+       elt{ token.FUNC, "func", keyword },
+       elt{ token.GO, "go", keyword },
+       elt{ token.GOTO, "goto", keyword },
+       elt{ token.IF, "if", keyword },
+       elt{ token.IMPORT, "import", keyword },
+
+       elt{ token.INTERFACE, "interface", keyword },
+       elt{ token.MAP, "map", keyword },
+       elt{ token.PACKAGE, "package", keyword },
+       elt{ token.RANGE, "range", keyword },
+       elt{ token.RETURN, "return", keyword },
+
+       elt{ token.SELECT, "select", keyword },
+       elt{ token.STRUCT, "struct", keyword },
+       elt{ token.SWITCH, "switch", keyword },
+       elt{ token.TYPE, "type", keyword },
+       elt{ token.VAR, "var", keyword },
 }
 
 
-const whitespace = "  \t  ";  // to separate tokens
-
-func init() {
-       // set pos fields
-       pos := 0;
-       for i := 0; i < len(tokens); i++ {
-               tokens[i].pos = pos;
-               pos += len(tokens[i].lit) + len(whitespace);
-       }
-}
-
+const whitespace = "  \t  \n\n\n";  // to separate tokens
 
 type TestErrorHandler struct {
        t *testing.T
@@ -174,38 +163,61 @@ func (h *TestErrorHandler) Error(loc scanner.Location, msg string) {
 }
 
 
+func NewlineCount(s string) int {
+       n := 0;
+       for i := 0; i < len(s); i++ {
+               if s[i] == '\n' {
+                       n++;
+               }
+       }
+       return n;
+}
+
+
 func Test(t *testing.T) {
        // make source
        var src string;
        for i, e := range tokens {
                src += e.lit + whitespace;
        }
-
-       // set up scanner
-       var s scanner.Scanner;
-       s.Init(io.StringBytes(src), &TestErrorHandler{t}, true);
+       whitespace_linecount := NewlineCount(whitespace);
 
        // verify scan
-       for i, e := range tokens {
-               loc, tok, lit := s.Scan();
-               if loc.Pos != e.pos {
-                       t.Errorf("bad position for %s: got %d, expected %d", e.lit, loc.Pos, e.pos);
-               }
-               if tok != e.tok {
-                       t.Errorf("bad token for %s: got %s, expected %s", e.lit, token.TokenString(tok), token.TokenString(e.tok));
-               }
-               if token.IsLiteral(e.tok) && string(lit) != e.lit {
-                       t.Errorf("bad literal for %s: got %s, expected %s", e.lit, string(lit), e.lit);
+       index := 0;
+       eloc := scanner.Location{0, 1, 1};
+       scanner.Tokenize(io.StringBytes(src), &TestErrorHandler{t}, true,
+               func (loc Location, tok int, litb []byte) bool {
+                       e := elt{token.EOF, "", special};
+                       if index < len(tokens) {
+                               e = tokens[index];
+                       }
+                       lit := string(litb);
+                       if tok == token.EOF {
+                               lit = "<EOF>";
+                               eloc.Col = 0;
+                       }
+                       if loc.Pos != eloc.Pos {
+                               t.Errorf("bad position for %s: got %d, expected %d", lit, loc.Pos, eloc.Pos);
+                       }
+                       if loc.Line != eloc.Line {
+                               t.Errorf("bad line for %s: got %d, expected %d", lit, loc.Line, eloc.Line);
+                       }
+                       if loc.Col != eloc.Col {
+                               t.Errorf("bad column for %s: got %d, expected %d", lit, loc.Col, eloc.Col);
+                       }
+                       if tok != e.tok {
+                               t.Errorf("bad token for %s: got %s, expected %s", lit, token.TokenString(tok), token.TokenString(e.tok));
+                       }
+                       if token.IsLiteral(e.tok) && lit != e.lit {
+                               t.Errorf("bad literal for %s: got %s, expected %s", lit, lit, e.lit);
+                       }
+                       if tokenclass(tok) != e.class {
+                               t.Errorf("bad class for %s: got %d, expected %d", lit, tokenclass(tok), e.class);
+                       }
+                       eloc.Pos += len(lit) + len(whitespace);
+                       eloc.Line += NewlineCount(lit) + whitespace_linecount;
+                       index++;
+                       return tok != token.EOF;
                }
-               if tokenclass(tok) != e.class {
-                       t.Errorf("bad class for %s: got %d, expected %d", e.lit, tokenclass(tok), e.class);
-               }
-       }
-       loc, tok, lit := s.Scan();
-       if tok != token.EOF {
-               t.Errorf("bad token at eof: got %s, expected EOF", token.TokenString(tok));
-       }
-       if tokenclass(tok) != special {
-               t.Errorf("bad class at eof: got %d, expected %d", tokenclass(tok), special);
-       }
+       );
 }