09:17
rakkable joined
|
|||
patrickb | melezhik: I've had a look. Arm64 is pretty doable (I lean towards use hetzner to create a throwaway server for the build, alternative: Google cloud, RPi in basement). | 12:15 | |
tellable6 | patrickb, I'll pass your message to melezhik | ||
patrickb | melezhik: RISCV is a different beast. I don't know if anyone ever managed to get Rakudo building or running on RISCV at all. That would be the first step. | 12:17 | |
tellable6 | patrickb, I'll pass your message to melezhik | ||
patrickb | melezhik: Do you know whether the rocky Linux people managed to build or run Rakudo? | 12:20 | |
tellable6 | patrickb, I'll pass your message to melezhik | ||
patrickb | build.opensuse.org/package/show/de...aku/rakudo | 12:22 | |
tellable6 | 2025-10-02T11:10:34Z #raku <melezhik_> patrickb: it'd be cool to add binary for arm64and risc-v - it'd be extremely helpful for folks from rocky linux team. No pressure though )) | ||
patrickb | melezhik: it seems riscv actually does compile already. Yay! | 12:23 | |
tellable6 | patrickb, I'll pass your message to melezhik | ||
patrickb | ab5tract: Do you know whether zig allows cross compiling across architectures? | 12:30 | |
according to zig.guide/build-system/cross-compilation/ it does... | 12:31 | ||
ab5tract | Yeah I was the under impression that it does | 12:32 | |
patrickb | ab5tract: Then I won't do anything about arm64 and RISCV. Let's see how zig turns out. If that works out it seems all of our platform related wishes are automatically solved. | 12:43 | |
ab5tract Thanks for bringing this up! | |||
ab5tract | patrickb: cool! I’ll prioritize organizing my patch so that we can get explore its possibilities | 12:48 | |
patrickb | ab5tract: Did you investigate whether the perf difference of the binaries of zig vs other compilers is acceptable or not? | 12:50 | |
Zig seems to be the perfect solution for us with the one potential downside being perf. But I'd be very willing to accept a few percents of perf loss. | 12:51 | ||
ab5tract | Re: perf differences, zig users clang exclusively for production builds at the moment | 12:52 | |
That is, their self hosted compiler is only enabled in debugging for now, and I think even then it’s still behind a feature flag | 12:53 | ||
patrickb | I guess the best way forward is to just create a first zig Rakudo build and run a few measurements on our own. | ||
ab5tract | github.com/ab5tract/MoarVM | 12:54 | |
Ahh, dang, I was hoping it would link to the branch | |||
patrickb | ab5tract Please don't feel too much pressured by this feature request. In general it's always possible to build a Rakudo oneself if no precomp binary is available. It is a convenience thing. | 12:55 | |
ab5tract | zig-support is the branch | ||
patrickb: can you give it a try and see if it works for you? | 12:56 | ||
patrickb | I'll do! (I hope to do so in the evening. pretty busy right now) | 12:58 | |
ab5tract | I don’t know whether there are zig compiler features that are “unlocked” by using a native zig build (eg using their build.zig approach rather than simply swapping in ‘zig cc’ for $CC | ||
Likewise no pressure :) | |||
13:08
melezhik_ joined
13:20
melezhik_ left
13:25
melezhik joined
14:07
melezhik_ joined
14:12
melezhik_ left
17:25
melezhik left
|
|||
Geth | rakudo/r4991-feed-methods: 36fc56dae1 | ab5tract++ | src/Raku/ast/expressions.rakumod RakuAST - Fix indirect methods with feed operators There is an indirect method call syntax, eg: foo-method($foo: |@args) This should work with feeds, such that: ... (9 more lines) |
18:24 | |
rakudo: ab5tract++ created pull request #5967: RakuAST - Fix indirect methods with feed operators |
|||
20:17
librasteve_ joined
|