opt

html/entity.go

@@ -17,14 +17,13 @@ const longestEntityWithoutSemicolon = 6
 // https://www.w3.org/TR/html4/sgml/entities.html
 var entity map[string]entityVal
 
-type entityVal struct {
-	// 8 bytes total, for good alignment
-	Len int16
-	Val [6]byte
-}
+type entityVal [8]byte
 
 func mkEntityVal(str string) (ret entityVal) {
-	ret.Len = int16(copy(ret.Val[:], str))
+	ret[0] = byte(len(str))
+	if copy(ret[1:], str) < len(str) {
+		panic("entity value does not fit in buffer: "+str)
+	}
 	return ret
 }
 

html/entity_test.go

@@ -17,7 +17,13 @@ func TestEntityLength(t *testing.T) {
 		t.Fatal("maps not loaded")
 	}
 
-	for k := range entity {
+	// We verify that the length of UTF-8 encoding of each value
+	// is no more than 1 + len("&"+key), which is an assuption
+	// made in unescapeEntity.
+	for k, v := range entity {
+		if 2+len(k) < int(v[0]) {
+			t.Error("escaped entity &" + k + " is more than 1 byte shorter than its UTF-8 encoding " + string(v[1:1+v[0]]))
+		}
 		if len(k) > longestEntityWithoutSemicolon && k[len(k)-1] != ';' {
 			t.Errorf("entity name %s is %d characters, but longestEntityWithoutSemicolon=%d", k, len(k), longestEntityWithoutSemicolon)
 		}

html/escape.go

@@ -157,28 +157,28 @@ func unescapeEntity[S ~[]byte | string](dst []byte, src S, dstPos, srcPos int, a
 		// No-op.
 	} else if attribute && entityName[len(entityName)-1] != ';' && len(s) > i && s[i] == '=' {
 		// No-op.
-	} else if entityVal := entity[string(entityName)]; entityVal.Len != 0 {
-		if entityVal.Len > int16(i) {
+	} else if x := entity[string(entityName)]; x[0] != 0 {
+		if int(x[0]) > i {
 			// This assumes that it only ever has to grow by 1 byte per entity.
 			if dstPos == srcPos && dstIsSrc {
 				// make a copy + grow
 				dst = append(dst[:len(dst):len(dst)], 0)
-			} else if dstPos+int(entityVal.Len) >= len(dst) {
+			} else if dstPos+int(x[0]) >= len(dst) {
 				// grow, but don't necessarily make a copy
 				dst = append(dst, 0)
 			}
 		}
-		return dst, dstPos + copy(dst[dstPos:], entityVal.Val[:entityVal.Len]), srcPos + i
+		return dst, dstPos + copy(dst[dstPos:], x[1:1+x[0]]), srcPos + i
 	} else if !attribute {
 		maxLen := len(entityName) - 1
 		if maxLen > longestEntityWithoutSemicolon {
 			maxLen = longestEntityWithoutSemicolon
 		}
 		for j := maxLen; j > 1; j-- {
-			if entityVal := entity[string(entityName[:j])]; entityVal.Len != 0 {
+			if x := entity[string(entityName[:j])]; x[0] != 0 {
 				// This assumes that no entity without a semicolon
 				// has a value that is wider than its name.
-				return dst, dstPos + copy(dst[dstPos:], entityVal.Val[:entityVal.Len]), srcPos + j + 1
+				return dst, dstPos + copy(dst[dstPos:], x[1:1+x[0]]), srcPos + j + 1
 			}
 		}
 	}