• 1 Post
  • 40 Comments
Joined 1 year ago
cake
Cake day: June 20th, 2023

help-circle


  • Linksys WRT54G

    The Linksys WRT54G did not run OpenWrt by default and the original OS does not even remotely resemble OpenWrt. What OpenWrt did use from the original OS was the Broadcom wireless driver because it was closed source (and a similar kernel version, so the driver could be used), since there was no driver in the mainline kernel.

    But to try to answer the question, this device has been designed by the OpenWrt developers to fit their needs (and their users needs). Other routers running some variant of OpenWrt on them by default were designed by companies unrelated to the project. They most likely used OpenWrt because it was convenient to them. Their intentions weren’t usually the same as the OpenWrt team’s (repairability, easy to unbrick, etc.). Not that there is anything wrong with that. I like GL.Inet routers.


  • Actually AMDs mobile parts are pretty good at idle power consumption and so are their desktop APUs. Their normal CPUs, which use the chiplet design are rather poor when it comes to idle power consumption. Intel isn’t really any better when compared to the monolithic parts at idle and Intel CPUs have horrible power consumption under load. Their newest CPUs are better when it comes to efficiency than 13th and 14th gen CPU, bus still don’t match or even exceed AMD.




  • Unless you require the dynamic features of Wordpress, you could have a look at some of the static site generators out there (such as Hugo). Having a static site would reduce the attack surface considerably. Also due to the shenanigans happening with Wordpress at the moment, I would be weary of using it.

    About SSL, what others have already mentioned, SSL certs are available for free these days, thanks to letsencrypt.





  • The head of BitWarden has come out and stated the SDK being required to compile BitWarden was a mistake, however, and if this proves to be true (which I have no reason to doubt) then I see no reason why any of this is an issue.

    I don’t see why this should make any difference at all. Sure, I get why he is are saying they are going to fix it - he thinks that this gets them in compliance with the GPLv3. But from a practical point of view there is no difference at all. The software is useless without that SDK part. Even if it does indeed get them in the clear from a legal point of view (which I am not convinced that it actually does), it is still a crappy situation.

    I think, it would look way less shady, if they said they are going fully source-available and not pretend that they are keeping the client open source. I would still dislike that, of course. At least that wouldn’t have eroded the trust in them as much as it did for me.


  • Proprietary is a strong word IMO. Here’s the repo, it’s not FOSS, but it is source available.

    Yeah, that’s what I meant by “proprietary”. I guess having the source to look at is better than nothing, but it still leaves me uneasy. Their license lets them do anything they want (ignoring that - as it stands - their license is void due to the linkage with GPLv3 code, but they said they want to fix that). I have no idea what their plan is. I don’t think it is in their best interest to go the route they appear to be going. Having truly open source clients seems to be a selling point for quite a few customers. But what do I know…



  • I really hope that this is actually the case, but I am not very optimistic. This doesn’t seem to be a mistake. They intentionally move functionality of their clients to their proprietary SDK library. The Bitwarden person stated this in the Github issue and you can also check the commit history. Making that library a build-time dependency might actually have been a mistake. That does not change the fact, that the clients are no longer useful without that proprietary library going forward. Core functionality has been move to that lib. I really don’t care if they talk to that library via some protocol or have it linked at build time. I wouldn’t consider this open source, even if that client wrapper that talks to that library technically is still licensed under GPLv3.



  • It is really not just a packaging bug. If you read that comment of the Bitwarden person a little further, you’ll notice that he’s talking about that proprietary “SDK” library that they are integrating with their clients. Even if they manage to not actually link it directly with the client, but rather let the client talk to that library via some protocol - it doesn’t make the situation any better. The client won’t work without their proprietary “SDK”, no matter if they remove the build-time dependency or not.


  • Perhaps the hard dependency was a mistake, but not them moving more and more code to their proprietary library. It appears that their intent is to make the client mostly a wrapper around their proprietary library, so they can still claim to have an open source GPLv3 piece of software. What good is that client if you can only use it in conjunction with that proprietary library, even if you can build it without that dependency?



  • BitWarden already has lots of clients.

    Does it? I’d be very much interested to know. I’ve been looking for other clients before, because I didn’t like the sluggishness of the Electron client, but couldn’t find any usable clients at all. There are some projects on Github, none of which seemed to be in a usable state. Perhaps I have been missing something.

    This is being blown a bit out of proportion though. All they are saying is the official SDK may have some non-free components going forward. So what? It’s a private company, they can do what they want. Or the community can just fork it and move forward with a free one if they want, but it’s just not going to be in the official BitWarden clients. Hardly news or a big deal.

    Nobody said that they can’t do that (although people rightfully questioned that their changes are indeed comatible with the GPLv3). I very much disagree that this isn’t a big deal, though.



  • When you use a typical 74 Wh (“20000 mAh”) power bank, you can expect more than 12 hours of runtime, if your average power draw stays at or below 5 W. Of course you aren’t going to do much transcoding with a Pi in any case, but multiple concurrent streams shouldn’t be much of an issue.