Accessing new architectures are back burner projects. Some of the most recent progress on RISC-V was done mostly by one developer using a simulator.
AMD64 and x86 are still the main focus still and with all the recent progress on initiating graphics acceleration with Mesa3D, I really don’t see that changing. I’ve got a Cubox i4P waiting for the day that AArch64 support begins though. Its on-die Vivante graphics core is already partially supported by Mesa under Linux. I understand that ARM SBCs are a bonus rather than a central feature in the meantime.
As for Chrome, I won’t hold my breath because Google stopped its password sync to Chromium, I ended up importing all my config files into Brave instead on my Linux box. Only Debian based distributions are supported by mainline Chrome. Manjaro Linux isn’t supported and Haiku doesn’t have much hope of getting Chrome either.
Firefox is being rewritten in Rust, slowly. If the package manager needs to go cross-architecture, Wasmer is mostly written in Rust also. Do you program in Rust?
I saw this comment on OSNews. That site used to be great, but I don’t really rate the content any more. It is a glorified Blog these days. The quality of the comments shows that. I saw also a few people said “well report the bug then” but the OP didn’t respond last time I saw it. The anecdotes are easy to push as reality on places like OSNews with minimal moderation. But because there is no context and not link to an unfixed bug report, the OP was just venting without helping fix the “10 minute” job.
They also completely missed the fact that the people porting Haiku to other architectures are experts in those architectures, and that by porting the OS, you will also find and fix more bugs related to “assumptions” made about how stuff should be done. I think I have seen that @X512 has found “bugs” (or at least cases) that no one would have otherwise seen and might have caused issues even for Intel down the line.
2021-05-22 12:34 PM Brisvegas
People would rather spend six months porting to a new architecture or developing a new package
manger instead of spending 10 minutes fixing a ‘show stopper’ bug. eg. You can’t use an NDIS wifi
connection out of the box with Haiku. Yet is just a few minutes work to fix. Absurd.
I agree with this. I was extremely dissapointed when Thomas started to rant about political things at osnews.
With this too. Some people would like to start some kind of revolution and flame war in the Haiku scene because they are displeased about the introduction of the package management. There were hot discussions about this here on this forum, where most people showed support for the package management idea. So they now use every possible way to share their (also biased) opinion about this on other sites (like hackernews, osnews, probably at reddit to, etc.). This is like throwing excrement everywhere.
It was told them already they are free to fork, but man, that would require work. Throwing half-truths, biased opinions (even if you think is true) is so much easier! And you can get likes too! And you can use multiple nicks even!
And then they do not elaborate on why adding NDIS WiFi support would onlytake a couple minutes when asked to elabiorate on it. Typical of these kinds of commenters.
Recognized one of the Haiku detractors there as someone who’s one of these kind of folk described in the quote above. They were one of those against package management here in the forums and seems to be actively spreading FUD about the OS now elsewhere. They’re banging on about bad desktop defaults now for instance, when Haiku’s desktop hasn’t changed much at all. Absolute nonsense.
NDIS means Windows NT network drivers? That will require NT compatibility layer in Haiku kernel with PE module loader and partial NT kernel and drivers API implementation. That is definitely not “only take a couple of minutes”.
Also I don’t understand why it is so important. Most WiFi adapters are working on Haiku and it is often possible to replace mPCIe WiFi adapter with compatible one.