Cannot watch embedded youtube videos on slashfilm

[birdie-github]

When this extension is enabled YouTube embeds at slashfilm.com don't work, e.g. https://www.slashfilm.com/burning-trailer-official-us/

I tried whitelisting slashfilm.com but that didn't help.

I tried adding

|\/embed\/

to the list of "Skip Links Matching" but the add-on apparently ignores this rule.

[birdie-github]

Log for slashfilm.com:

https://www.youtube.com/get_video_info?html5=1&video_id=MuehR3WuZ_E&cpn=Hro2QszWYvvi3jk8&eurl=https%3A%2F%2Fwww.youtube.com%2Fembed%2FMuehR3WuZ_E%3Ffeature%3Doembed&el=embedded&hl=en_US&sts=17798&lact=7&c=WEB_EMBEDDED_PLAYER&cver=20180921&cplayer=UNIPLAYER&cbr=Firefox&cbrver=62.0&cos=Windows&cosver=10.0&width=740&height=416&ei=Av6sW4i1LYXsgAfU-534CQ&iframe=1&embed_config=%7B%7D → https://www.youtube.com/embed/MuehR3WuZ_E?feature=oembed

https://www.youtube.com/get_video_info?html5=1&video_id=MuehR3WuZ_E&cpn=Hro2QszWYvvi3jk8&eurl=https%3A%2F%2Fwww.youtube.com%2Fembed%2FMuehR3WuZ_E%3Ffeature%3Doembed&el=embedded&hl=en_US&sts=17798&lact=7&c=WEB_EMBEDDED_PLAYER&cver=20180921&cplayer=UNIPLAYER&cbr=Firefox&cbrver=62.0&cos=Windows&cosver=10.0&width=740&height=416&ei=Av6sW4i1LYXsgAfU-534CQ&iframe=1&embed_config=%7B%7D&playerretry=1 → https://www.youtube.com/embed/MuehR3WuZ_E?feature=oembed

https://www.youtube.com/get_video_info?html5=1&video_id=MuehR3WuZ_E&cpn=Hro2QszWYvvi3jk8&eurl=https%3A%2F%2Fwww.youtube.com%2Fembed%2FMuehR3WuZ_E%3Ffeature%3Doembed&el=embedded&hl=en_US&sts=17798&lact=7&c=WEB_EMBEDDED_PLAYER&cver=20180921&cplayer=UNIPLAYER&cbr=Firefox&cbrver=62.0&cos=Windows&cosver=10.0&width=740&height=416&ei=Av6sW4i1LYXsgAfU-534CQ&iframe=1&embed_config=%7B%7D&playerretry=2 → https://www.youtube.com/embed/MuehR3WuZ_E?feature=oembed

[birdie-github]

Skip matching:

youtube\.com\/get_video_info

fixes the issue, but along with #42, are you sure we don't have too many exceptions?

Maybe "Clean Outgoing HTTP requests" by default is an overkill?

[Cimbali]

Whitelisting www.youtube.com works, but I suppose that's not satisfactory?

[birdie-github]

It's the opposite to what you add-on does because in this case external links at youtube will all go through youtube.com/redirect

[Cimbali]

I'm not sure what you mean, but whitelisting targets the origin domain. Now they are both the same in your example, so that's not a really useful distinction.

That is, https://domain.com/get_video_info?[...]&eurl=https%3A%2F%2Fexample.com matches on domain.com, not on example.com.

[birdie-github]

You're right but again I guess we'll have to have too many exceptions and update them all the time.

Disabling "Clean Outgoing HTTP requests" by default looks like a much better solution.

Most people care about actual redirects because it's a privacy issue but when the page fetches external resources in order to work - there's not a lot of privacy going on.

[Cimbali]

There's almost always something fishy privacy-wise when your current URL is passed to another request. To be honest I'm not even sure what youtube needs it for in this case, especially since they already pass the video_id separately.

I mean look at that URL, it's passing more of your information to youtube than it's asking about video information to youtube:

html5         1
video_id      MuehR3WuZ_E
cpn           Hro2QszWYvvi3jk8
eurl          https://www.youtube.com/embed/MuehR3WuZ_E?feature=oembed
el            embedded
hl            en_US
sts           17798
lact          7
c             WEB_EMBEDDED_PLAYER
cver          20180921
cplayer       UNIPLAYER
cbr           Firefox
cbrver        62.0
cos           Windows
cosver        10.0
width         740
height        416
ei            Av6sW4i1LYXsgAfU-534CQ
iframe        1
embed_config  {}
playerretry   2

So we can see you're on Windows 10.0 with Firefox 62.0, with a 740x416 viewport (for the iframe probably), and so on (I'm guessing CBR must be Client BRowser, COS must be Client OS. CPN some Client Product Number maybe?)

I think what we should strive towards is to strip the useless info instead, but I don't have a lot of time these days to implement finer-grain URL cleaning. For example, it would be interesting to see if just passing video_id on the URL is enough for the video to work?