Stopped working on particular sites after Firefox update

[Name Redacted]

Hello,

Just yesterday I let Linux update Firefox to 112.0.2, and IMMEDIATELY I lost ability to download from

https://www.winktv.co.kr/live

https://www.pandalive.co.kr/live

"No media to process in the current tab" while obviously there is a video stream playing, same as any other day right before I ran the update.

While downloading from other sites like YouTube and some other sites still works. 

Linux Mint Cinnamon 19.3

Firefox 112.0.2

Video Downloadhelper 7.6.6

Companion app 1.6.3

PLEASE HELP!

[Name Redacted]

It wasn't just Firefox which got updated, unfortunately. A couple dozen of system updates. But I can tell that the pre-update Firefox was v111.0.1.

I tried reinstalling companion app, tried resetting all settings to default, and tried disabling and re-enabling the Extension.

[Wild Willy]

Patience. It takes a bit to write a proper response. I have a response. Stop posting
for a bit. I've got an answer for you.

[Wild Willy]

I'm using Windows 7 64-bit, Firefox 112.0.2 64-bit, licensed VDH 7.6.5.a3 beta (same as
7.6.6), CoApp 1.6.3.

I clicked the winktv link you gave, then just picked the first thing on that page. Once
there, I clicked the play button in the player. VDH indeed recognized nothing. So I
went to Plan B, which we have discussed so many times in this forum; Open the Network
Monitor. I assume F12 works the same on Linux as on Windows. Then I reloaded the page.
Then I sorted the Network Monitor display on the Type column, as shown in the attached
image #01. I scrolled through the Network Monitor & saw lots of occurrences of js, json,
images, html, css . . . all useless to us. I normally look for media (mp4 or webm or mkv
etc.) & especially master manifests (m3u8 for HLS or mpd for DASH). No luck here. So I
just picked the very first item in the list & popped up its context menu. I tried a
couple of other things that normally give me the contents of objects but they didn't
work. The one that worked was the indicated Save All As HAR, That gave a file selection
dialog. I saved the file & opened it in Notepad++ (a Notepad replacement that I favor).
I then did a search for the string m3u8. You should understand that I was guessing here.
I got lucky with this approach, which is typical. Every site is different & there's no
telling whether this will work on any other site or even on any other content on this
site.

There were several hits on m3u8 but only the very last 3 occurrences were of any use.
I'm showing them in attached image #02. Notepad++ has a handy feature whereby if you
highlight one string, Notepad++ automatically highlights all other occurrences of the
string in the file. As you can see, the 3 occurrences of m3u8 occur in a URL that is
repeated. Looks like 3 URLs, it's really only one. So I copied the one URL out to where
I could look at it by itself. I noticed something rather odd. Every / in the URL was
preceded by \\. I don't know why that is. You can see that the beginning & end of each
URL is \". This is typical. It's a convention I've seen in lots of other cases. The "
is being escaped by \. But this is the first time I've seen / being escaped. And I have
no idea why \ isn't enough to escape /. They are using \\ to escape each / in the URLs.
Whatever. You need to remove each occurrence of \\. Then you'll have a proper URL you
can give to ffprobe. I have attached the results of that as ffprobe.txt.

There is a tremendous load of lines in the ffprobe report showing it skipping over stuff.
This is typical of livestreams. You can scan through the skipped lines for educational
purposes. At the bottom of the report, you'll find the truly helpful information.
You'll note that the duration of this stream is given as N/A. That is typical of
livestreams & it makes sense. You don't know how long a livestream might last. You'll
also notice that the livestream started 47240.742556 in the past. I'm not sure if this
is seconds or milliseconds. But it seems like the stream had been live for a while when
I peeked in. I'll mention something about this below.

This stream is being offered in 4 different resolutions. This is represented by Program
0, Program 1, Program 2, & Program 3. Each Program consists of 3 Streams: an audio, a
video, & a timed_ID3 data. Timed_ID3 is something we have encountered multiple times in
this forum. You can do a search for those other times. They're easy to find. But
you'll see at the bottom of the ffprobe report what ffprobe thinks of timed_ID3 data.
And that is nothing. It has no codec for interpreting that data. I sure wish I could
read some of this famous timed_ID3 data. I'd like to know what it is. But I have never
found any clear description of what that stuff is & I have never found a tool that reads
it. Ffmpeg ignores it, VLC ignores, & VDH ignores it but . . . It seems that sometimes
VDH doesn't completely ignore it & it causes downloaded content to skip & jump. In any
case, we hate timed_ID3 data so we will ignore it.

With the information gleaned from the ffprobe report, I constructed an invocation of
ffmpeg & recorded 3 minutes of this livestream. I have attached the log from that
invocation as Korean Livestreammp4.txt. (If you pay close attention to the ffmpeg
command I ran, you'll note that the log file has a different name. But Google won't let
you attach a file with just any old extension. So I changed the file name to .txt before
I attached it here.) That shows the complete ffmpeg command, how long the recording
lasted, the size of the resulting file, & the average download speed. That isn't all at
the top of the file. You need to look for some of it. It's near the bottom.

You should not be surprised to know that the file I recorded plays fine with audio all
the way through. I didn't sit & watch it, just sampled it at intervals to make sure it
was OK.

Now, a lot of this sounds like I'm just picking this stuff out of my back pocket. But
anybody can do this stuff. There's lots of documentation here in this forum explaining
the full background so you can do this, too. Click this:

https://groups.google.com/g/video-downloadhelper-q-and-a/c/BzPLK2YyL-s

Be sure you carefully read the first 2 paragraphs there. Once you have read that, you
can check out the references within that web page that I am about to give you. But don't
just find the relevant link & click it. READ the context around the link.

Here are the references you need to look for within the text of that web page:

cannot download
evasive maneuvers
stealth quote

Do a string search for each of those strings on that web page. Each one of those
references gives you a link to click that will tell you all the background stuff you need
to learn if you want to record these livestreams. I mentioned above that this livestream
had been running a while. It is possible with ffmpeg to skip back to the beginning of
the stream. I discuss that in the ffmpeg tutorial. Be sure to pay attention to the
caveats I explain there. I also discuss the special parameters to use on ffmpeg when you
are recording a livestream, examples of which are in the command I ran. Do be sure to
read the stealth quote reference. People seem to miss that one when I mention it. It's
important. I wouldn't bother mentioning it if it were frivolous.

Now, you say this worked until you updated Firefox. One pat answer I like to give is
this: too bad. Web sites change all the time. They don't seem to care much about those
of us who want to download or record their content. If the site doesn't have a clearly
marked button for downloading their content, they don't care. You have to forget how
things used to work. That is ancient history, even if it was working fine just this
morning. Forget about it. Suck it up. Face the new reality.

That's my usual pat answer. But then, there is this issue:

https://groups.google.com/g/video-downloadhelper-q-and-a/c/O_bRwxrfqS8

These days, that is quite likely causing problems. You can try the circumvention
mentioned there. Actually, the circumvention is mentioned on Bugzilla.

Happy reading!

[Name Redacted]

Willy,

Per your linked bugzilla description, setting dom.workers.pFetch.enabled to false restores my ability to download videos from WinkTV and PandaTV. I am not sure yet about the repercussions of making this change in my browser settings, but my girls are more important to me than many other things.

I was going to test more extensively (Chrome on Linux, trying Winblows), but now I am fairly convinced it isn't the website making a change. There wasn't more than a day between being able to download, and not being able, and smack in the middle of that window was me updating Linux.

Thanks for the super-article otherwise, but I work 12 hours a day, and it will be a long time before I could put time aside for a science project to switch to a much better but much harder to use tool....................... Although ffmpeg has been cropping up in just about everything I need to do (lossless editing of videos, fixing video files that this tool breaks by forgetting to write the end of file, downloading stuff, transcoding, etc..........)

[Wild Willy]

According to recent activity on Bugzilla, it appears that our solution to this problem
will not be in the next release of Firefox, which will be 113. But it has been scheduled
for Firefox 114. I'm guessing we're looking at a delay of about 2 more months. So in
the interim, either you need to continue to use the circumvention described on Bugzilla
or you need to learn how to use ffmpeg.