• 0 Posts
  • 72 Comments
Joined 1 year ago
cake
Cake day: July 5th, 2023

help-circle



  • Sure, but my first month of Lemmy was via the web. I still use Mastodon on the web (no app). Whenever I find myself on Reddit (via search) it’s web only. If someone sends me a random YouTube link, I open it in the web. On the off chance someone sends me a random TikTok link, web only.

    Again I don’t use any of these apps, but if someone sends me a loops link and it says install the app, I’ll just ignore it.



  • So far, there is only some user settings accessible in the web interface. I don’t have an iPhone, hence I cannot say for their app. The android app will come later, apparently.

    I don’t give a shit about TikTok or short form video but legitimately how is this expected to succeed with these limitations? Clearly people like this stuff but can a fediverse service really afford to be iPhone only? This truly seems doomed to fail.







  • So first off I appreciate your “about:config” dump, that is helpful.

    However, having ONLY that dump makes it difficult to actually diagnose your issue. It would be helpful to have an idea of what fails.

    • Are you only unable to join meetings?
    • Can you create meetings?
    • Only audio?
    • Only video?
    • Are there any error messages?
    • Are there any messages at all on screen?
    • Are there errors in the JavaScript console?
    • If there are a lot, can you compare it against the working vanilla version and see where the errors are unique/start to differ?
    • When did this issue start to happen? (It worked a week ago, but not now?)
    • Is your config based off a guide online?

    Second, you have “media.peerconnection.enabled” set to “false”. That setting controls WebRTC. Google Meet uses WebRTC. If I set that value to false my Google Meet meetings fail to load, it just stays stuck at the joining meeting phase.

    So maybe that’s your culprit. Of course I would expect if you’ve had this set for a long time I would have expected it to fail for a long time. Also both alternatives you mentioned use WebRTC, so I would expect those to fail with that setting disabled.

    Of course if it used to work with that disabled then it’s possible Google Meet changed and had a fallback no longer in use or something else.

    As an aside, I understand that WebRTC has (had?) issues leaking local IP addresses while connected to a VPN. So I understand why you might have it disabled.


  • in my hardened config firefox browser

    The only conclusion I can draw is that google is so desperate for my data that they refuse to service me unless I give them this.

    What is your config? It’s possible you’ve got things configured in an unexpected, but valid, way that causes Google Meet to fail.

    How does it fail? Are there errors in the console?

    You’re immediately assuming malice, but it may simply be a mistake (Google’s or yours).



  • So a company provides infinite protection?

    “I didn’t murder that man, the company did.”

    “The company paid individual X to murder them, not me.”

    No, that’s ridiculous.

    There is a line you can cross. Musk has crossed that line. Is it exactly written, if your name is Elon Musk and you own companies X, Y, & Z, and you perform actions A, B, C, you I’ll be fined in this exact way? No. There is a grey area, and a group within the EU is allowed to make a more specific determination.

    And do you know who agreed to these rules? Elon Musk. He chose to do business in the EU. He agreed to their rules.




  • To clarify the cost of creating an LLC is a hundred bucks more or less depending on the jurisdiction. So Elon should be allowed to create “Musk Corp Oct2024 LLC” and then say or do anything under the guise of Musk Corp Oct2024 LLC, then if he’s sued or fined just declare bankruptcy and create “Musk Corp Nov2024 LLC” and do whatever he wants?

    At some point you have to recognize the individual is at fault. You can’t just hide behind “Oh that wasn’t me, that was the company” or " That was Musk of SpaceX having an opinion of Musk of Tesla, they are different entities."

    If someone is attempting to be genuine and truthful when it comes to personal statements, fine, we can consider the protections. But if someone is flagrant and malicious then those protections no longer apply.


  • Re Google Safe Browsing

    I would argue it’s a security feature with potential privacy concerns, however I would agree it is more of a failsafe or suggestion.

    However it being disabled by default or not included at compile time versus enabled by default may also be relevant when it comes to security. As a hypothetical a high severity bug with Google Safe Browsing could arguably make a browser less secure. However even as a failsafe/suggestion, the small security benefit may make the overall browser more secure, e.g. filtering known bad websites that attack known vulnerabilities.

    I’m also just using Safe Browsing as an example here, it may or may not be worth focusing on since a browser is basically an operating system.

    You mentioned sandboxing, which I think is perhaps a more reasonable scope.


    1. Do you have your current list of sources? You mentioned you want more, but where are you looking to start? For example are you looking at the CVE database? Are you looking at competitions like Pwn2Own? Or detailed project group like Google Project Zero?
    2. Is it fair to compare Chromium, which is not an end user product, to Firefox which is? Do you plan to look at or compare forks of the software? As an example both Google Chrome and Mozilla Firefox enable “Google Safe Browsing” by default, however the fork “ungoogled-chromium” does not include “Google Safe Browsing” (and they provide their reasoning).