func absoluteURLs(base *url.URL, s string) (result []*url.URL) {
matches := urlRegexp.FindAllString(s, -1)
for _, match := range matches {
u, err := urlx.ParseRef(base, match)
if err != nil {
continue
}
result = append(result, u)
}
return
}
func windowLocation(base *url.URL, s string) (result []*url.URL) {
matches := windowLocationRegexp.FindAllStringSubmatch(s, -1)
for _, match := range matches {
u, err := urlx.ParseRef(base, match[3])
if err != nil {
continue
}
result = append(result, u)
}
return
}
func (r *Response) scanLocation() {
var baseurl *url.URL
if baseurl = r.NewURL; baseurl == nil {
baseurl = r.URL
}
if loc := r.Header.Get("Content-Location"); loc != "" {
r.ContentLocation, _ = urlx.ParseRef(baseurl, loc)
}
if s := r.Header.Get("Refresh"); s != "" {
r.Refresh.Seconds, r.Refresh.URL = parseRefresh(s, baseurl)
}
}
func ExtractHref(base *url.URL, reader io.Reader, ch chan<- *url.URL) error {
z := html.NewTokenizer(reader)
f := func(z *html.Tokenizer, base *url.URL) *url.URL {
for {
key, val, more := z.TagAttr()
if bytes.Equal(key, []byte("href")) {
if u, err := urlx.ParseRef(base, string(val)); err == nil {
return u
}
break
}
if !more {
break
}
}
return nil
}
LOOP:
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
if err := z.Err(); err != io.EOF {
return err
}
break LOOP
case html.StartTagToken:
tn, hasAttr := z.TagName()
if hasAttr && len(tn) == 1 && tn[0] == 'a' {
if u := f(z, base); u != nil {
ch <- u
}
}
case html.SelfClosingTagToken:
tn, hasAttr := z.TagName()
if hasAttr && len(tn) == 4 && bytes.Equal(tn, []byte("base")) {
if u := f(z, base); u != nil {
base = u
}
}
}
}
return nil
}
func parseRefresh(s string, u *url.URL) (second int, uu *url.URL) {
const blank = " \t\n\f\r"
var i int
var err error
if i = strings.IndexAny(s, ";,"); i == -1 {
second, _ = strconv.Atoi(strings.TrimRight(s, blank))
return
}
if second, err = strconv.Atoi(strings.TrimRight(s[:i], blank)); err != nil {
return
}
s = strings.TrimLeft(s[i+1:], blank)
if i = strings.Index(s, "url"); i == -1 {
return
}
s = strings.TrimLeft(s[i+len("url"):], blank)
if !strings.HasPrefix(s, "=") {
return
}
s = strings.TrimLeft(s[1:], blank)
uu, _ = urlx.ParseRef(u, s)
return
}
func (e *Extractor) tokenLoop(
r *crawler.Response, body io.Reader, ch chan<- *url.URL, chErr chan<- error,
) {
defer close(chErr)
defer close(ch)
z := html.NewTokenizer(body)
base := *r.NewURL
normalize := e.Normalize
dest := e.Pos
if normalize == nil {
normalize = urlx.Normalize
}
if len(dest) == 0 {
dest = []struct{ Tag, Attr string }{{"a", "href"}}
}
var prev html.Token
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
if err := z.Err(); err != io.EOF {
chErr <- err
}
return
case html.StartTagToken, html.SelfClosingTagToken:
token := z.Token()
prev = token
if len(token.Attr) == 0 {
continue
}
var (
v string
u *url.URL
ok bool
err error
name = string(token.Data)
)
for _, d := range dest {
if name != d.Tag {
continue
} else if v, ok = get(&token, d.Attr); !ok || v == "" {
continue
} else if u, err = urlx.ParseRef(
&base, v,
); err != nil {
continue
}
if name == "base" {
base = *u
}
if err = normalize(u); err != nil {
continue
}
ch <- u
}
case html.TextToken:
token := z.Token()
var urls []*url.URL
switch {
case e.SniffFlags&SniffWindowLocation != 0:
if prev.Type == html.StartTagToken && prev.Data == "script" {
urls = windowLocation(&base, token.Data)
}
case e.SniffFlags&SniffAbsoluteURLs != 0:
urls = absoluteURLs(&base, token.Data)
}
for _, u := range urls {
if err := normalize(u); err != nil {
continue
}
ch <- u
}
prev = token
default:
prev = html.Token{}
}
}
}
