ebnflint command

- basic verification of EBNF grammars
- tested with (and has testcase for) go_spec.html

R=rsc
DELTA=150  (148 added, 0 deleted, 2 changed)
OCL=31481
CL=31517

src/cmd/clean.bash

@@ -3,7 +3,7 @@
 # Use of this source code is governed by a BSD-style
 # license that can be found in the LICENSE file.
 
-for i in cc 6l 6a 6c 8l 8a 8c 8g 5l 5a 5c 5g gc 6g ar db nm acid cov gobuild godefs godoc gofmt prof gotest
+for i in cc 6l 6a 6c 8l 8a 8c 8g 5l 5a 5c 5g gc 6g ar db nm acid cov ebnflint gobuild godefs godoc gofmt prof gotest
 do
 	cd $i
 	make clean

src/cmd/ebnflint/Makefile

+# Copyright 2009 The Go Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style
+# license that can be found in the LICENSE file.
+
+include $(GOROOT)/src/Make.$(GOARCH)
+
+TARG=ebnflint
+OFILES=\
+	ebnflint.$O\
+
+$(TARG): $(OFILES)
+	$(LD) -o $(TARG) $(OFILES)
+
+test: $(TARG)
+	$(TARG) -start="SourceFile" $(GOROOT)/doc/go_spec.html
+
+clean:
+	rm -f $(OFILES) $(TARG)
+
+install: $(TARG)
+	cp $(TARG) $(HOME)/bin/$(TARG)
+
+%.$O:	%.go
+	$(GC) $<

src/cmd/ebnflint/ebnflint.go

+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package main
+
+import (
+	"bytes";
+	"ebnf";
+	"flag";
+	"fmt";
+	"io";
+	"os";
+	"path";
+	"sort";
+	"strings";
+)
+
+
+var start = flag.String("start", "Start", "name of start production");
+
+
+func usage() {
+	fmt.Fprintf(os.Stderr, "usage: ebnflint [flags] [filename]\n");
+	flag.PrintDefaults();
+	os.Exit(1);
+}
+
+
+// Markers around EBNF sections in .html files
+var (
+	open = strings.Bytes(`<pre class="ebnf">`);
+	close = strings.Bytes(`</pre>`);
+)
+
+
+func extractEBNF(src []byte) []byte {
+	var buf bytes.Buffer;
+
+	for i, j, n := 0, 0, len(src); ; {
+		// i = beginning of EBNF section
+		i = bytes.Index(src[j : n], open);
+		if i < 0 {
+			break;
+		}
+		i += j+len(open);
+
+		// write as many newlines as found in the excluded text
+		// to maintain correct line numbers in error messages 
+		for _, ch := range src[j : i] {
+			if ch == '\n' {
+				buf.WriteByte('\n');
+			}
+		}
+
+		// j = end of EBNF section
+		j = bytes.Index(src[i : n], close);
+		if j < 0 {
+			// missing closing
+			// TODO(gri) should this be an error?
+			j = n-i;
+		}
+		j += i;
+
+		// copy EBNF section
+		buf.Write(src[i : j]);
+	}
+
+	return buf.Data();
+}
+
+
+// TODO(gri) This is the same code for reportError as in gofmt.
+//           Should factor this out as part of some parsing framework
+//           that could also deal with reading various input sources.
+
+func reportError(filename string, err os.Error) {
+	if errors, ok := err.(ebnf.ErrorList); ok {
+		sort.Sort(errors);
+		for _, e := range errors {
+			fmt.Fprintf(os.Stderr, "%s:%v\n", filename, e);
+		}
+	} else {
+		fmt.Fprintf(os.Stderr, "%s: %v\n", filename, err);
+	}
+	os.Exit(1);
+}
+
+
+func main() {
+	flag.Parse();
+
+	var filename string;
+	switch flag.NArg() {
+	case 0:
+		filename = "/dev/stdin";
+	case 1:
+		filename = flag.Arg(0);
+	default:
+		usage();
+	}
+
+	src, err := io.ReadFile(filename);
+	if err != nil {
+		reportError(filename, err);
+	}
+
+	if path.Ext(filename) == ".html" {
+		src = extractEBNF(src);
+	}
+
+	grammar, err := ebnf.Parse(src);
+	if err != nil {
+		reportError(filename, err);
+	}
+
+	if err = ebnf.Verify(grammar, *start); err != nil {
+		reportError(filename, err);
+	}
+}

src/make.bash

@@ -18,7 +18,7 @@ rm -f $HOME/bin/quietgcc
 cp quietgcc.bash $HOME/bin/quietgcc
 chmod +x $HOME/bin/quietgcc
 
-for i in lib9 libbio libmach_amd64 libregexp cmd pkg cmd/gobuild cmd/godoc cmd/gofmt
+for i in lib9 libbio libmach_amd64 libregexp cmd pkg cmd/ebnflint cmd/gobuild cmd/godoc cmd/gofmt
 do
 	# The ( ) here are to preserve the current directory
 	# for the next round despite the cd $i below.

src/run.bash

@@ -42,6 +42,12 @@ time make
 time make smoketest
 ) || exit $?
 
+(xcd cmd/ebnflint
+make clean
+time make
+time make test
+) || exit $?
+
 (xcd ../doc/progs
 time ./run
 ) || exit $?