| 22 Apr 2025 |
 emily | In reply to @qyliss:fairydust.space
and I'm not sure I've ever seen that come up in practice I think all of this would come up a lot less in a situation where we truly had a tier 1 kind of user/contributor split and general support quality | 17:48:22 |
 Alyssa Ross | (Not saying it was a good idea, just that I've never seen or had the experience of having to break Darwin out as part of contributing upstream from Linux.) | 17:49:00 |
| emily | In reply to @qyliss:fairydust.space
We basically did this for QEMU for Darwin for a period of years :P well, yes. I won't defend that. tbf I was thinking of bigger things than a handful of reverts though | 17:49:05 |
| emily | In reply to @qyliss:fairydust.space
(Not saying it was a good idea, just that I've never seen or had the experience of having to break Darwin out as part of contributing upstream from Linux.) I think this is partly a function of upstreams like Rust having more resources for those platforms though | 17:49:50 |
| emily | like if Rust had comparable Darwin investment it wouldn't be tier 1 there either | 17:50:07 |
| emily | also something something microarchitectural patents something something RISC-V guy who is convinced running an x86 QEMU VM is illegal :P | 17:51:18 |
| emily | not entirely a farce: Intel do shake down people over x86 emulation! | 17:51:42 |
| emily | (ok, generally corporate people) | 17:52:02 |
| emily | anyway, I respect people who don't want to test stuff on Darwin and favour being an environment that is welcoming to people with strong views on software freedom - albeit I don't think we will/should ever get to the point of Guix there. I'm just unconvinced there's a practical obstacle to maintaining pluralism on the matter even in a world where Darwin is tier 1. | 17:54:18 |
 Tristan Ross | So we want to consider x86_64-linux-gnu` as the only tier 1 platform? | 17:55:14 |
| Tristan Ross | * So we want to consider x86_64-linux-gnu as the only tier 1 platform? | 17:55:25 |
| emily | that's the only thing I can imagine making sense as a description of the status quo, yeah | 17:56:30 |
| Tristan Ross | Cool | 17:56:37 |
| Tristan Ross | https://pad.lassul.us/9yYTVp73QBul7dCm6tCWsw?view I've made changes to this now, I'm basically using it to draft things until I have a clear way to PR this into the documentation. | 17:57:56 |
| emily | which is x86_64-linux > aarch64-linux >> aarch64-darwin > x86_64-darwin >>> everything else | 17:59:00 |
| Tristan Ross | Yeah, I think we could see aarch64-linux become a tier 1 platform if adoption grows more. | 18:00:11 |
| Tristan Ross | It's not going to be anytime soon but it'll likely happen at some point. | 18:00:40 |
| emily | maybe depending on how the server market shifts | 18:01:25 |
| emily | desktops would require a change in Microsoft priorities | 18:01:39 |
| Tristan Ross | In reply to @emilazy:matrix.org
maybe depending on how the server market shifts Yeah, basically Ampere, Graviton, and NVIDIA's ARM chips lol. But I think it's mostly Ampere or Graviton. | 18:02:51 |
| Tristan Ross | In reply to @emilazy:matrix.org
desktops would require a change in Microsoft priorities Yeah... Apple Silicon is a thing but I'd consider it a niche | 18:03:16 |
| Tristan Ross | I think Pi's are an easily accessible option but not much performance there. However, I do see it being a simple setup for people. | 18:03:55 |
| emily | the difference is that there don't need to be any convenient options for x86 since it's what everyone already has | 18:05:17 |
| Tristan Ross | Yeah, which is how Apple was able to do it. Forced adoption. | 18:06:10 |
| Tristan Ross | If companies forced Arm more, there'd be more convenient hardware. | 18:06:41 |
| emily | well the arm machines being the same price or cheaper but faster probably helped :P | 18:07:25 |
| Tristan Ross | Definitely | 18:07:39 |
| Tristan Ross | Battery life for one is a thing. | 18:07:58 |
| emily | we're not going to see a substantial amount of users and contributors on desktop AArch64 Linux as long as it's paying more for worse perf | 18:08:04 |
| emily | we surely have more using Apple Silicon fulltime | 18:08:20 |
