Add the YouTube submission path

Step 5. A URL goes through the same pipeline as an upload — it just gains a
download step and a source_url.

- The host allowlist is checked on the parsed hostname before yt-dlp is
  invoked, so lookalikes and userinfo tricks are refused too
- yt-dlp -J reads metadata synchronously with a 15s timeout; a timeout leaves
  the fields blank rather than failing the submission
- Over-long tracks are refused from that metadata, before a byte is downloaded
- Failed URL submissions offer Yritä uudelleen with the typed text intact;
  uploads cannot retry, so they offer re-upload

Prefill takes track then title, and artist then creator then uploader, and
leaves a field blank rather than inventing one. testdata/ytdlp-noose.json is a
real dump of an ordinary upload, which has none of the music fields.

Also fixes a URL-only submit being blocked by the file input's required
attribute — HTML cannot express "one of these two", so the server says it.

The image now takes yt-dlp from Alpine 3.24 instead of pip, which drops
python3 and pip entirely; see decision 19.
This commit is contained in:
Esa Kataja
2026-07-31 21:54:15 +03:00
parent 80d3e36679
commit 91e136055c
12 changed files with 307 additions and 16 deletions
+81
View File
@@ -0,0 +1,81 @@
package main
import (
"os"
"testing"
"time"
)
func TestAllowedYouTubeURL(t *testing.T) {
for _, ok := range []string{
"https://www.youtube.com/watch?v=XnfMBo4IQ-g",
"https://youtu.be/XnfMBo4IQ-g",
"https://music.youtube.com/watch?v=XnfMBo4IQ-g",
"http://m.youtube.com/watch?v=XnfMBo4IQ-g",
} {
if _, allowed := allowedYouTubeURL(ok); !allowed {
t.Errorf("%q was rejected", ok)
}
}
for _, bad := range []string{
"",
"not a url",
"file:///etc/passwd",
"https://evil.example.com/watch?v=x",
// The allowlist is on the host, so a lookalike path or userinfo must not pass.
"https://evil.example.com/www.youtube.com/watch?v=x",
"https://youtube.com.evil.example.com/watch?v=x",
"https://[email protected]/",
"-oExecuteMe",
} {
if _, allowed := allowedYouTubeURL(bad); allowed {
t.Errorf("%q was allowed", bad)
}
}
}
// A real dump of an ordinary upload: no track, no artist, no creator, no album — just a title with
// double spaces and an uploader. This is why prefill must never invent an "Unknown".
func TestYouTubeMetaFromOrdinaryUpload(t *testing.T) {
raw, err := os.ReadFile("testdata/ytdlp-noose.json")
if err != nil {
t.Skipf("fixture missing: %v", err)
}
meta, err := parseYouTubeMeta(raw)
if err != nil {
t.Fatal(err)
}
if meta.Title != "Sentenced Noose" {
t.Errorf("title = %q, want the cleaned video title", meta.Title)
}
if meta.Artist != "Heikki Rokkonen" {
t.Errorf("artist = %q, want the uploader as the last fallback", meta.Artist)
}
if meta.Duration != 245*time.Second {
t.Errorf("duration = %v, want 4m5s", meta.Duration)
}
}
func TestYouTubeMetaPrefersMusicFields(t *testing.T) {
meta, err := parseYouTubeMeta([]byte(`{
"title": "Sentenced - Noose (Official Video)",
"track": "Noose", "artist": "Sentenced", "creator": "ignored",
"uploader": "SentencedVEVO", "duration": 245.0}`))
if err != nil {
t.Fatal(err)
}
if meta.Title != "Noose" || meta.Artist != "Sentenced" {
t.Errorf("got %q by %q, want the track/artist fields to win", meta.Title, meta.Artist)
}
}
// Empty stays empty: a blank field prompts the submitter, a plausible "Unknown" does not.
func TestYouTubeMetaLeavesBlanksBlank(t *testing.T) {
meta, err := parseYouTubeMeta([]byte(`{"duration": 10.0}`))
if err != nil {
t.Fatal(err)
}
if meta.Title != "" || meta.Artist != "" {
t.Errorf("got %q by %q, want both empty", meta.Title, meta.Artist)
}
}