func getInclude(z *html.Tokenizer, attrs []html.Attribute) (startMarker, endMarker string, error error) {
var srcString string
if url, hasUrl := getAttr(attrs, "src"); !hasUrl {
return "", "", fmt.Errorf("include definition without src %s", z.Raw())
} else {
srcString = strings.TrimSpace(url.Val)
if strings.HasPrefix(srcString, "#") {
srcString = srcString[1:]
}
}
required := false
if r, hasRequired := getAttr(attrs, "required"); hasRequired {
if requiredBool, err := strconv.ParseBool(r.Val); err != nil {
return "", "", fmt.Errorf("error parsing bool in %s: %s", z.Raw(), err.Error())
} else {
required = requiredBool
}
}
if required {
return fmt.Sprintf("§[> %s]§", srcString), "", nil
} else {
return fmt.Sprintf("§[#> %s]§", srcString), fmt.Sprintf("§[/%s]§", srcString), nil
}
}
func buildHTML(tokenizer *html.Tokenizer) (s string, err error) {
buf := new(bytes.Buffer)
bp := 0
if tag, _ := tokenizer.TagName(); string(tag) == "div" {
div := tokenizer.Raw()
buf.Write(div)
bp = len(div)
err = nextToken(tokenizer)
}
ep := bp
for err != io.EOF {
if err != nil && err != io.EOF {
return
}
ep = buf.Len()
b := tokenizer.Raw()
if _, err := buf.Write(b); err != nil {
return "", err
}
err = nextToken(tokenizer)
}
b := buf.Bytes()
if bp > 0 {
b = b[bp:ep]
}
return string(b), nil
}
func parseFragment(z *html.Tokenizer) (f Fragment, dependencies []*FetchDefinition, err error) {
attrs := make([]html.Attribute, 0, 10)
dependencies = make([]*FetchDefinition, 0, 0)
buff := bytes.NewBuffer(nil)
forloop:
for {
tt := z.Next()
tag, _ := z.TagName()
raw := byteCopy(z.Raw()) // create a copy here, because readAttributes modifies z.Raw, if attributes contain an &
attrs = readAttributes(z, attrs)
switch {
case tt == html.ErrorToken:
if z.Err() != io.EOF {
return nil, nil, z.Err()
}
break forloop
case tt == html.StartTagToken || tt == html.SelfClosingTagToken:
if string(tag) == UicInclude {
if replaceTextStart, replaceTextEnd, err := getInclude(z, attrs); err != nil {
return nil, nil, err
} else {
fmt.Fprintf(buff, replaceTextStart)
// Enhancement: WriteOut sub tree, to allow alternative content
// for optional includes.
fmt.Fprintf(buff, replaceTextEnd)
continue
}
}
if skipSubtreeIfUicRemove(z, tt, string(tag), attrs) {
continue
}
case tt == html.EndTagToken:
if string(tag) == UicFragment || string(tag) == UicTail {
break forloop
}
}
buff.Write(raw)
}
return StringFragment(buff.String()), dependencies, nil
}
func (parser *HtmlContentParser) parseHead(z *html.Tokenizer, c *MemoryContent) error {
attrs := make([]html.Attribute, 0, 10)
headBuff := bytes.NewBuffer(nil)
forloop:
for {
tt := z.Next()
tag, _ := z.TagName()
raw := byteCopy(z.Raw()) // create a copy here, because readAttributes modifies z.Raw, if attributes contain an &
attrs = readAttributes(z, attrs)
switch {
case tt == html.ErrorToken:
if z.Err() != io.EOF {
return z.Err()
}
break forloop
case tt == html.StartTagToken || tt == html.SelfClosingTagToken:
if skipSubtreeIfUicRemove(z, tt, string(tag), attrs) {
continue
}
if string(tag) == "script" && attrHasValue(attrs, "type", ScriptTypeMeta) {
if err := parseMetaJson(z, c); err != nil {
return err
}
continue
}
case tt == html.EndTagToken:
if string(tag) == "head" {
break forloop
}
}
headBuff.Write(raw)
}
s := headBuff.String()
st := strings.Trim(s, " \n")
if len(st) > 0 {
c.head = StringFragment(st)
}
return nil
}
func parseMetaJson(z *html.Tokenizer, c *MemoryContent) error {
tt := z.Next()
if tt != html.TextToken {
return fmt.Errorf("expected text node for meta json, but found %v, (%s)", tt.String(), z.Raw())
}
bytes := z.Text()
err := json.Unmarshal(bytes, &c.meta)
if err != nil {
return fmt.Errorf("error while parsing json from meta json element: %v", err.Error())
}
tt = z.Next()
tag, _ := z.TagName()
if tt != html.EndTagToken || string(tag) != "script" {
return fmt.Errorf("Tag not properly ended. Expected </script>, but found %s", z.Raw())
}
return nil
}
func parseToken(tokenizer *html.Tokenizer, htmlDoc *htmlDocument, parent *tagElement) (bool, bool, string) {
tokenType := tokenizer.Next()
switch tokenType {
case html.ErrorToken:
return true, false, ""
case html.TextToken:
text := string(tokenizer.Text())
if strings.TrimSpace(text) == "" {
break
}
textElement := &textElement{text: text}
appendElement(htmlDoc, parent, textElement)
case html.StartTagToken:
tagElement := &tagElement{tagName: getTagName(tokenizer), startTagRaw: string(tokenizer.Raw())}
appendElement(htmlDoc, parent, tagElement)
for {
errorToken, parentEnded, unsetEndTag := parseToken(tokenizer, htmlDoc, tagElement)
if errorToken {
return true, false, ""
}
if parentEnded {
if unsetEndTag != "" {
return false, false, unsetEndTag
}
break
}
if unsetEndTag != "" {
return false, false, setEndTagRaw(tokenizer, tagElement, unsetEndTag)
}
}
case html.EndTagToken:
return false, true, setEndTagRaw(tokenizer, parent, getTagName(tokenizer))
case html.DoctypeToken, html.SelfClosingTagToken, html.CommentToken:
tagElement := &tagElement{tagName: getTagName(tokenizer), startTagRaw: string(tokenizer.Raw())}
appendElement(htmlDoc, parent, tagElement)
}
return false, false, ""
}
func getFetch(z *html.Tokenizer, attrs []html.Attribute) (*FetchDefinition, error) {
fd := &FetchDefinition{}
url, hasUrl := getAttr(attrs, "src")
if !hasUrl {
return nil, fmt.Errorf("include definition without src %s", z.Raw())
}
fd.URL = strings.TrimSpace(url.Val)
if name, hasName := getAttr(attrs, "name"); hasName {
fd.Name = name.Val
} else {
fd.Name = urlToName(fd.URL)
}
if timeout, hasTimeout := getAttr(attrs, "timeout"); hasTimeout {
if timeoutInt, err := strconv.Atoi(timeout.Val); err != nil {
return nil, fmt.Errorf("error parsing timeout in %s: %s", z.Raw(), err.Error())
} else {
fd.Timeout = time.Millisecond * time.Duration(timeoutInt)
}
}
if required, hasRequired := getAttr(attrs, "required"); hasRequired {
if requiredBool, err := strconv.ParseBool(required.Val); err != nil {
return nil, fmt.Errorf("error parsing bool in %s: %s", z.Raw(), err.Error())
} else {
fd.Required = requiredBool
}
}
attr, found := getAttr(attrs, "discoveredby")
if found {
fd.DiscoveredBy(attr.Val)
}
return fd, nil
}
func (parser *HtmlContentParser) parseBody(z *html.Tokenizer, c *MemoryContent) error {
attrs := make([]html.Attribute, 0, 10)
bodyBuff := bytes.NewBuffer(nil)
attrs = readAttributes(z, attrs)
if len(attrs) > 0 {
c.bodyAttributes = StringFragment(joinAttrs(attrs))
}
forloop:
for {
tt := z.Next()
tag, _ := z.TagName()
raw := byteCopy(z.Raw()) // create a copy here, because readAttributes modifies z.Raw, if attributes contain an &
attrs = readAttributes(z, attrs)
switch {
case tt == html.ErrorToken:
if z.Err() != io.EOF {
return z.Err()
}
break forloop
case tt == html.StartTagToken || tt == html.SelfClosingTagToken:
if skipSubtreeIfUicRemove(z, tt, string(tag), attrs) {
continue
}
if string(tag) == UicFragment {
if f, deps, err := parseFragment(z); err != nil {
return err
} else {
c.body[getFragmentName(attrs)] = f
for _, dep := range deps {
c.requiredContent[dep.URL] = dep
}
}
continue
}
if string(tag) == UicTail {
if f, deps, err := parseFragment(z); err != nil {
return err
} else {
c.tail = f
for _, dep := range deps {
c.requiredContent[dep.URL] = dep
}
}
continue
}
if string(tag) == UicFetch {
if fd, err := getFetch(z, attrs); err != nil {
return err
} else {
c.requiredContent[fd.URL] = fd
continue
}
}
if string(tag) == UicInclude {
if replaceTextStart, replaceTextEnd, err := getInclude(z, attrs); err != nil {
return err
} else {
bodyBuff.WriteString(replaceTextStart)
// Enhancement: WriteOut sub tree, to allow alternative content
// for optional includes.
bodyBuff.WriteString(replaceTextEnd)
continue
}
}
case tt == html.EndTagToken:
if string(tag) == "body" {
break forloop
}
}
bodyBuff.Write(raw)
}
s := bodyBuff.String()
if _, defaultFragmentExists := c.body[""]; !defaultFragmentExists {
if st := strings.Trim(s, " \n"); len(st) > 0 {
c.body[""] = StringFragment(st)
}
}
return nil
}
// setEndTagRaw sets an endTagRaw to the parent.
func setEndTagRaw(tokenizer *html.Tokenizer, parent *tagElement, tagName string) string {
if parent != nil && parent.tagName == tagName {
parent.endTagRaw = string(tokenizer.Raw())
return ""
}
return tagName
}
