Fix removal of query parameters on youtube embed links

This uses go's html and url parser plus regex, instead of
using only a single regex for simplicity sake. A single regex expression
might be error prone, for example, when trying to matching html entities.

Fixes #328
This commit is contained in:
Dami 2020-08-07 00:05:43 -06:00
parent 849e5b8503
commit 9b614bc922
1 changed files with 21 additions and 4 deletions

View File

@ -16,6 +16,7 @@ import (
"html"
"html/template"
"net/http"
"net/url"
"regexp"
"strings"
"unicode"
@ -73,6 +74,25 @@ func applyMarkdown(data []byte, baseURL string, cfg *config.Config) string {
return applyMarkdownSpecial(data, false, baseURL, cfg)
}
func disableYoutubeAutoplay(outHTML string) string {
for _, match := range youtubeReg.FindAllString(outHTML, -1) {
u, err := url.Parse(match)
if err != nil {
log.Error("Couldn't parse youtube url: %v", err)
}
u.RawQuery = html.UnescapeString(u.RawQuery)
q := u.Query()
// Set Youtube autoplay url parameter, if any, to 0
if len(q["autoplay"]) == 1 {
q.Set("autoplay", "0")
}
u.RawQuery = q.Encode()
clean_url := u.String()
outHTML = strings.Replace(outHTML, match, clean_url, 1)
}
return outHTML
}
func applyMarkdownSpecial(data []byte, skipNoFollow bool, baseURL string, cfg *config.Config) string {
mdExtensions := 0 |
blackfriday.EXTENSION_TABLES |
@ -108,10 +128,7 @@ func applyMarkdownSpecial(data []byte, skipNoFollow bool, baseURL string, cfg *c
// Strip newlines on certain block elements that render with them
outHTML = blockReg.ReplaceAllString(outHTML, "<$1>")
outHTML = endBlockReg.ReplaceAllString(outHTML, "</$1></$2>")
// Remove all query parameters on YouTube embed links
// TODO: make this more specific. Taking the nuclear approach here to strip ?autoplay=1
outHTML = youtubeReg.ReplaceAllString(outHTML, "$1")
outHTML = disableYoutubeAutoplay(outHTML)
return outHTML
}