libreplanet-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: media.libreplanet.org non-requested confusing auto download issue


From: Jean Louis
Subject: Re: media.libreplanet.org non-requested confusing auto download issue
Date: Wed, 5 May 2021 18:51:02 +0300
User-agent: Mutt/2.0.6 (2021-03-06)

As for campaigns@fsf.org in copy of this email I suggest that on FSF
Mediagoblin pages of Libreplanet, preloading should be turned off, and
poster should be placed instead.

To generate a poster I am scaling video down by using Common Lisp, and
then generating video thumbnail by using `ffmpeg':

(defun media-scale (width height new-width)
  "Returns the height proportional to the width
height, and based on the new width"
  (let* ((ratio (/ width height))
         (height (truncate (float (/ new-width ratio)))))
    (list new-width height)))

(format nil "ffmpeg -y -i '~a' -vframes 1 -s ~ax~a '~a'"
            video width height temp-thumbnail-320)

For me is important that width and height of the screenshot of video
corresponds to the same scale as original one. But one could create
picture first and then resize it with Imagemagick:

Example:

$ ffmpeg -y -i xvideo8.mp4 -vframes 1 xvideo8.jpg
$ mogrify -resize 320 xvideo8.jpg

that would result in resized video. However MediaGoblin is not fit for
that, it would require maybe some configuration or even patches.

Then the URL to the final file xvideo8.jpg would be used as
poster="https://www.example.com/xvideo8.jpg"; that would show in place
of video.

* Miroslav Rovis <miro.rovis@croatiafidelis.hr> [2021-05-05 17:08]:
> But since you're on bugs at gnu (as your email says), I hope
> my lead and your explanation is sufficient to make the
> change, and I hope it will just happen, silently is fine for
> me. 

Definitely not, I do not represent GNU project, I just support GNU
project. 

GNU project while supported by FSF is in its own, it is not same as
FSF, though much related. 

Libreplanet.ORG is not GNU project, it is FSF's project. You can
try reporting to: campaigns@fsf.org

> > In many countries people pay good amounts for data, and often Internet
> > is not fast, it could take many hours to load such video.
> It is also that analysis is more than an order of magnitude
> slower than simple browsing. I can't know what I take into
> my machine quickly even with year 2020 launched modern
> (commodity) processor machine (AMD Ryzen 7 Pro 4750G), with
> unnecessary preloading like this, that only analysis can
> tell what it (likely) is.

As I work with Website Revision System in background, I have today
reworked a plugin for my own publishing that by default no video will
preload. None was preloaded by default even before, but discussion as
here helped me realize what is important.

> What I mean is, it took a couple of minutes to preloading-imposed
> download into browser cache a good portion of grandsun1715.webm
> file, but when, seeing the quick growing of the network trace and
> understanding that some unexpected traffic was happening, I cut the
> network connection (physically), and went on to analyse with
> Wireshark and some scripts of mine, it took me many hours to reach
> to my conclusion, because Wireshark, Tcpdump, and other are good
> tools, but the network is not optimized for analysis, it's optimised
> for quick use, not analysis...

To avoid that confusion, next time you can simply use right mouse
click and option Inspect Element or just F12 in Firefox-based
browser, then you go to Network tab, reload and watch for things.

> So, many hours it took me to analyze and reach to my (probable)
> conclusion, including the failed decryption of exactly the huge
> unexpected download. That basically means that possibly there was
> MiTM and spoof that happened as well. Hope not, but thearetically
> possible.

That is good approach for unknown connections on computer, but the
above one is so much simpler when you know it was invoked by browser.

> To me, knowing what gets into my machine --and the browser is the
> most used for intrusion, has the attack surface ridiculously huge
> and hard to control-- is as important as free software and
> hardware. Free software and hardware must be safe, else my freedom
> can easily be compromised and hence it's not freedom anymore. [*]

Recommended reading:

How to Run a More Secure Browser
https://www.dragonflybsd.org/docs/handbook/RunSecureBrowser/

When I wish to invoke URL by using the above mentioned DragonFlyBSD
tip on safe browsing, I am sometimes using this Emacs function:

(defun browse-safe-url (url &optional arg)
  "Browse URL with b"
  (let ((username "ano")) ;; different username than my own
    ;; Insecurity settings for personal DISPLAY only
    (shell-command "xhost +")
    ;; Browse URL with different username
    (async-start-process "sudo" "sudo" nil "su" "-c" "--" username "-c"
                         (format "exec iceweasel \"%s\"" url))))

Otherwise you may construct command for you:

This allows access to your X Server from other user's space, but is
insecure if you are in some local network. If you are behind firewall
it is pretty alright.

$ xhost + 

You must be sudoer in /etc/sudoers and have to configure another user
name on the system. You then launch a browsre like `iceweasel'

$ sudo su -c -- ANOTHER-USER -c iceweasel

Then browser will launch in other user's space. 

The assumption is that in that user's space you should not keep any
sensitive private files.


-- 
Jean

Take action in Free Software Foundation campaigns:
https://www.fsf.org/campaigns

Sign an open letter in support of Richard M. Stallman
https://stallmansupport.org/
https://rms-support-letter.github.io/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]