1
0
Fork 0
mirror of https://github.com/gocsaf/csaf.git synced 2025-12-22 05:40:11 +01:00

Implemented links on directory listings checking

This commit is contained in:
Sascha L. Teichmann 2022-05-17 01:22:30 +02:00
parent 32be3602b6
commit 3bbd37c441
4 changed files with 92 additions and 9 deletions

View file

@ -10,11 +10,49 @@ package main
import (
"io"
"net/http"
"net/url"
"github.com/PuerkitoBio/goquery"
)
func linksOnPage(r io.Reader) ([]string, error) {
func (p *processor) linksOnPageURL(baseDir string) ([]string, error) {
base, err := url.Parse(baseDir)
if err != nil {
return nil, err
}
client := p.httpClient()
p.checkTLS(baseDir)
res, err := client.Get(baseDir)
p.badDirListings.use()
if err != nil {
p.badDirListings.add("Fetching %s failed: %v", base, err)
return nil, errContinue
}
if res.StatusCode != http.StatusOK {
p.badDirListings.add("Fetching %s failed. Status code %d (%s)",
base, res.StatusCode, res.Status)
return nil, errContinue
}
defer res.Body.Close()
// Links may be relative
return linksOnPage(res.Body, func(link string) (string, error) {
u, err := url.Parse(link)
if err != nil {
return "", err
}
return base.ResolveReference(u).String(), nil
})
}
func linksOnPage(r io.Reader, resolve func(string) (string, error)) ([]string, error) {
doc, err := goquery.NewDocumentFromReader(r)
if err != nil {
@ -23,11 +61,16 @@ func linksOnPage(r io.Reader) ([]string, error) {
var links []string
doc.Find("a").Each(func(i int, s *goquery.Selection) {
doc.Find("a").Each(func(_ int, s *goquery.Selection) {
if err != nil {
return
}
if link, ok := s.Attr("href"); ok {
links = append(links, link)
if link, err = resolve(link); err == nil {
links = append(links, link)
}
}
})
return links, nil
return links, err
}