bytes, strings: allow -1 in Map to mean "drop this character".

xml: drop invalid characters in attribute names
    when constructing struct field names.

R=rsc
CC=r
https://golang.org/cl/157104

src/pkg/bytes/bytes.go

@@ -207,7 +207,8 @@ func HasSuffix(s, suffix []byte) bool {
 }
 
 // Map returns a copy of the byte array s with all its characters modified
-// according to the mapping function.
+// according to the mapping function. If mapping returns a negative value, the character is
+// dropped from the string with no replacement.
 func Map(mapping func(rune int) int, s []byte) []byte {
 	// In the worst case, the array can grow when mapped, making
 	// things unpleasant.  But it's so rare we barge in assuming it's
@@ -222,16 +223,18 @@ func Map(mapping func(rune int) int, s []byte) []byte {
 			rune, wid = utf8.DecodeRune(s[i:])
 		}
 		rune = mapping(rune);
-		if nbytes+utf8.RuneLen(rune) > maxbytes {
-			// Grow the buffer.
-			maxbytes = maxbytes*2 + utf8.UTFMax;
-			nb := make([]byte, maxbytes);
-			for i, c := range b[0:nbytes] {
-				nb[i] = c
+		if rune >= 0 {
+			if nbytes+utf8.RuneLen(rune) > maxbytes {
+				// Grow the buffer.
+				maxbytes = maxbytes*2 + utf8.UTFMax;
+				nb := make([]byte, maxbytes);
+				for i, c := range b[0:nbytes] {
+					nb[i] = c
+				}
+				b = nb;
 			}
-			b = nb;
+			nbytes += utf8.EncodeRune(rune, b[nbytes:maxbytes]);
 		}
-		nbytes += utf8.EncodeRune(rune, b[nbytes:maxbytes]);
 		i += wid;
 	}
 	return b[0:nbytes];

src/pkg/bytes/bytes_test.go

@@ -365,6 +365,19 @@ func TestMap(t *testing.T) {
 	if string(m) != expect {
 		t.Errorf("rot13: expected %q got %q", expect, m)
 	}
+
+	// 5. Drop
+	dropNotLatin := func(rune int) int {
+		if unicode.Is(unicode.Latin, rune) {
+			return rune
+		}
+		return -1;
+	};
+	m = Map(dropNotLatin, Bytes("Hello, 세계"));
+	expect = "Hello";
+	if string(m) != expect {
+		t.Errorf("drop: expected %q got %q", expect, m)
+	}
 }
 
 func TestToUpper(t *testing.T)	{ runStringTests(t, ToUpper, "ToUpper", upperTests) }

src/pkg/strings/strings.go

@@ -178,7 +178,8 @@ func HasSuffix(s, suffix string) bool {
 }
 
 // Map returns a copy of the string s with all its characters modified
-// according to the mapping function.
+// according to the mapping function. If mapping returns a negative value, the character is
+// dropped from the string with no replacement.
 func Map(mapping func(rune int) int, s string) string {
 	// In the worst case, the string can grow when mapped, making
 	// things unpleasant.  But it's so rare we barge in assuming it's
@@ -188,20 +189,22 @@ func Map(mapping func(rune int) int, s string) string {
 	b := make([]byte, maxbytes);
 	for _, c := range s {
 		rune := mapping(c);
-		wid := 1;
-		if rune >= utf8.RuneSelf {
-			wid = utf8.RuneLen(rune)
-		}
-		if nbytes+wid > maxbytes {
-			// Grow the buffer.
-			maxbytes = maxbytes*2 + utf8.UTFMax;
-			nb := make([]byte, maxbytes);
-			for i, c := range b[0:nbytes] {
-				nb[i] = c
+		if rune >= 0 {
+			wid := 1;
+			if rune >= utf8.RuneSelf {
+				wid = utf8.RuneLen(rune)
+			}
+			if nbytes+wid > maxbytes {
+				// Grow the buffer.
+				maxbytes = maxbytes*2 + utf8.UTFMax;
+				nb := make([]byte, maxbytes);
+				for i, c := range b[0:nbytes] {
+					nb[i] = c
+				}
+				b = nb;
 			}
-			b = nb;
+			nbytes += utf8.EncodeRune(rune, b[nbytes:maxbytes]);
 		}
-		nbytes += utf8.EncodeRune(rune, b[nbytes:maxbytes]);
 	}
 	return string(b[0:nbytes]);
 }

src/pkg/strings/strings_test.go

@@ -279,6 +279,19 @@ func TestMap(t *testing.T) {
 	if m != expect {
 		t.Errorf("rot13: expected %q got %q", expect, m)
 	}
+
+	// 5. Drop
+	dropNotLatin := func(rune int) int {
+		if unicode.Is(unicode.Latin, rune) {
+			return rune
+		}
+		return -1;
+	};
+	m = Map(dropNotLatin, "Hello, 세계");
+	expect = "Hello";
+	if m != expect {
+		t.Errorf("drop: expected %q got %q", expect, m)
+	}
 }
 
 func TestToUpper(t *testing.T)	{ runStringTests(t, ToUpper, "ToUpper", upperTests) }

src/pkg/xml/read.go

@@ -10,6 +10,7 @@ import (
 	"os";
 	"reflect";
 	"strings";
+	"unicode";
 )
 
 // BUG(rsc): Mapping between XML elements and data structures is inherently flawed:
@@ -144,6 +145,20 @@ func (p *Parser) Unmarshal(val interface{}, start *StartElement) os.Error {
 	return p.unmarshal(v.Elem(), start);
 }
 
+// fieldName strips invalid characters from an XML name
+// to create a valid Go struct name.  It also converts the
+// name to lower case letters.
+func fieldName(original string) string {
+	return strings.Map(
+		func(x int) int {
+			if unicode.IsDigit(x) || unicode.IsLetter(x) {
+				return unicode.ToLower(x)
+			}
+			return -1;
+		},
+		original)
+}
+
 // Unmarshal a single XML element into val.
 func (p *Parser) unmarshal(val reflect.Value, start *StartElement) os.Error {
 	// Find start element if we need it.
@@ -269,7 +284,7 @@ func (p *Parser) unmarshal(val reflect.Value, start *StartElement) os.Error {
 				val := "";
 				k := strings.ToLower(f.Name);
 				for _, a := range start.Attr {
-					if strings.ToLower(a.Name.Local) == k {
+					if fieldName(a.Name.Local) == k {
 						val = a.Value;
 						break;
 					}
@@ -303,7 +318,7 @@ Loop:
 			// Look up by tag name.
 			// If that fails, fall back to mop-up field named "Any".
 			if sv != nil {
-				k := strings.ToLower(t.Name.Local);
+				k := fieldName(t.Name.Local);
 				any := -1;
 				for i, n := 0, styp.NumField(); i < n; i++ {
 					f := styp.Field(i);

src/pkg/xml/read_test.go

@@ -24,8 +24,8 @@ func TestUnmarshalFeed(t *testing.T) {
 // hget http://codereview.appspot.com/rss/mine/rsc
 const rssFeedString = `
 <?xml version="1.0" encoding="utf-8"?>
-<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en-us"><title>Code Review - My issues</title><link href="http://codereview.appspot.com/" rel="alternate"></link><link href="http://codereview.appspot.com/rss/mine/rsc" rel="self"></link><id>http://codereview.appspot.com/</id><updated>2009-10-04T01:35:58+00:00</updated><author><name>rietveld</name></author><entry><title>rietveld: an attempt at pubsubhubbub
-</title><link href="http://codereview.appspot.com/126085" rel="alternate"></link><updated>2009-10-04T01:35:58+00:00</updated><author><name>email-address-removed</name></author><id>urn:md5:134d9179c41f806be79b3a5f7877d19a</id><summary type="html">
+<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en-us"><title>Code Review - My issues</title><link href="http://codereview.appspot.com/" rel="alternate"></link><li-nk href="http://codereview.appspot.com/rss/mine/rsc" rel="self"></li-nk><id>http://codereview.appspot.com/</id><updated>2009-10-04T01:35:58+00:00</updated><author><name>rietveld</name></author><entry><title>rietveld: an attempt at pubsubhubbub
+</title><link hre-f="http://codereview.appspot.com/126085" rel="alternate"></link><updated>2009-10-04T01:35:58+00:00</updated><author><name>email-address-removed</name></author><id>urn:md5:134d9179c41f806be79b3a5f7877d19a</id><summary type="html">
   An attempt at adding pubsubhubbub support to Rietveld.
 http://code.google.com/p/pubsubhubbub
 http://code.google.com/p/rietveld/issues/detail?id=155
@@ -208,3 +208,21 @@ not being used from outside intra_region_diff.py.
 		},
 	},
 }
+
+type FieldNameTest struct {
+	in, out string;
+}
+
+var FieldNameTests = []FieldNameTest{
+	FieldNameTest{"Profile-Image", "profileimage"},
+	FieldNameTest{"_score", "score"},
+}
+
+func TestFieldName(t *testing.T) {
+	for _, tt := range FieldNameTests {
+		a := fieldName(tt.in);
+		if a != tt.out {
+			t.Fatalf("have %#v\nwant %#v\n\n", a, tt.out)
+		}
+	}
+}