Do you have to do anything special to build M1 native-image in github workflows?
Iโve browsed some of your excellent workflow files on github but I donโt understand how it works yet.
Here for example the amd64 image is built: https://github.com/clojure-lsp/clojure-lsp/blob/e1adabd1a12d1b9893cf5e88ac2b03e87875e9f9/.github/workflows/nightly.yml#L218 But I donโt see similar ones for arm64..
ATM there is no way for that work on github actions :/
for clojure-lsp, we build https://github.com/clojure-lsp/clojure-lsp/blob/master/.cirrus.yml
that macos on github actions is only for amd64
There is a github issue with lots of people asking for that, but it's opened for a long time and github doesn't seem to prioritize
Thanks eric!
If only it was finally the year of the linux desktop then we could have a simple life!
Linux comes in arm64 too so you would still have this problem
yep, asus started to build arm64 notebooks, I'm looking forward to replace mine soon
And donโt tell me about qemu, itโs too ffing slow
Haha I was just about to tell you about qemu ๐
I use CircleCi for aarch64 Linux and cirrus for macOS arm64 but you could use Cirrus for both
Yeah they both seem to be using graviton2 from aws in the background which is pretty cool
oh, good to know about that
Github probably has to wait for azure to duct tape together some off the shelf laptops to create their runners
At Nubank we are starting to test services on graviton aws instances as well, but it's pretty early yet like only staging stuff
It looks awesome to be honest
Huge thanks btw to all you guys for the inspiration and guidelines on how to get graalvm up and running with clojure.