Hi, main maintainer of Flare here. Nice to see some competition in the space, especially with regards to voice calls.
Unfortunately, there is lots of wrong information in this thread.
Are you worried about Signal taking you down for TOS violations, since obviously your client is a TOS violation? I assume now that you’re advertising your app they’ll see it and consider planning accordingly to kill it, although maybe the captcha is enough.
Signal hasn’t been going after unofficial clients for years at this point. I have been talking to Whisperfish developers for developing Flare for a few years now, and they have never received an angry E-Mail from Signal. And Whisperfish is now almost a decade old, although it changed primary maintainers to the persons I talked to in 2019. So there has been no TOS E-Mail from Signal for unofficial clients since at least 7 years, at least none that I know of.
To my knowledge, the current stance of Signal with regards to third-party clients is that they ignore them and pretend that they don’t exist.
I think the way it went for Flare was that the primary dev of the project was able to keep themselves on it but Signal updated to where they would ban you from all of Signal servers for 7 days if you tried to switch to Flare from a normal Signal account.
No idea where you got that from. I was never banned from Signal servers, and have never heared of anybody being banned for using Flare (or any other libsignal-service-rs based clients like Whisperfish or gurk-rs). And given there are ~600 people per month downloading Flare from Flathub, I would certainly have heared it if people are being banned from Signal. If you had such an experience, please report such things.
I mean a huge part is that Flare is written in Rust and using a whole reimplementation of the Signal protocol written in Rust that has caused issues in terms of compatibility with the official implementation in the past.
That one is complicated. Let’s compare the stacks of Flare and Linal as far as I can see:
- Flare:
- flare-backend: My own library.
- presage: Co-maintained mostly by gurk-rs and Flare developers, sometimes also Whisperfish developers)
- libsignal-service-rs: Co-maintained mostly by Whisperfish developers, sometimes also by Flare and gurk-rs developers.
- libsignal: Official Signal library.
- Linal:
- signal-cli: Not sure who maintains it, probably mostly people that directly use signal-cli.
- libsignal-service-java fork: According to the README of signal-cli; 7 years ago officially maintained by Signal, now an unofficial fork.
- (libsignal: Their wiki also links to aforementioned libsignal, so maybe they also use that; not sure though.)
So best-case for Linal we depend on the same libsignal, except for a different stack between it. Worst-case for Linal, it depends on an unofficial community fork of a library that has been deprecated 7 years ago from Signal. I cannot tell which of the two (or some middle ground) is true, as I don’t know signal-cli too well. But Flare is in that point always at least as good as Linal.
So that should have less compatibility issues in general, I assume (depends a bit on changes necessary from Android to Java in this case).
That seems unlikely to me. From the history of big breakages I have experienced with Flare, most have also happened with signal-cli (and I think I also had cases where signal-cli caused an issue but Flare did not).
In my opinion, as long as Signal doesn’t provide easy-to-use some library (on the level of flare-backend, not libsignal) for most of the stack, there will always be breakages, except soft-forking official applications like Molly. The current stance of Signal that they ignore the existence of third-party clients doesn’t help here. There has been some movement on official higher-level libraries like libsignal-net, but it’ll still be a while until we can use that (currently it leads to conflicts regarding BoringSSL vs OpenSSL if I recall correctly).
I would argue that both clients are not ideal since in an ideal world we could have the official Signal desktop application support arm64 with the feature to act as main device build-in. But since we don’t get that, the differences are mainly in performance and reliability.
Agree with that part, but even if Signal would publish for arm64, with primary device, and mobile-friendly, Flare and Linal would still be interesting for providing a more consistent UI with the rest of the system (Phosh, other GTK apps, …).
I mean technically porting the Signal desktop app for Linux to aarch64 would help a lot others too.
There is currently an official beta of Signal Desktop on arm64 Linux. Don’t know the current state of that though.
Aside: Wanted to provide links to all projects I mentioned and claims I made, but can’t add more than two links due to being a new Purism Form member. If you can’t find something I mentioned, please ask and I’ll link it.