Lets start a (hopefully not too painful!) #Guix adventure thread!
I downloaded the latest VM image (1.4.0), started up QEMU like this:
qemu-system-x86_64
-nic user,model=virtio-net-pci,hostfwd=tcp::10022-:22
--enable-kvm
-m 2048
-device virtio-blk,drive=guix -drive if=none,file=guix-system-vm-image-1.4.0.x86_64-linux.qcow2,id=guix
Boots up fine, gets an IP, nice. But there's no SSH running, so I can't log in. Lets try adding one!
Side note: if you just want to package software, you don't need a guix system and thus no VM. You can install guix alongside your Linux distribution without any interfering. See "foreign distribution" in the manual.
(Of course to test services, etc. you need a guix system.)
@algernon
This way, from your current system, you can run "guix system" to build QEMU images and run them. It greatly smooths the experience, IMHO. 😁
When you are done and not sold to Guix, you can just uninstall it from your current system.
https://guix.gnu.org/manual/devel/en/guix.html#Binary-Installation
The documentation says I should add an SSH server. The linked page there doesn't tell me much about how, though.
I figured I'll copy /run/current-system/configuration.scm somewhere, edit it, and guix system reconfigure /etc/config.scm.
But alas, that ends with:
Git error: unsupported URL protocol
In the traceback above, I see references to /home/ludo/. I suppose it is trying to fetch channels from there or something?
guix describe indeed says the repo URL is /home/ludo/src/guix/+version-1.4.0. I guess I'll need to change that, but... how?
I'm running guix pull in another terminal, but oh dear, that is slow as fuck.
Ok, pulling from Codeberg is considerably faster. From Savannah, it got to 12% in like 12 minutes. Codeberg is 25% through after half a minute.
Somewhat surprisingly, the download part was the fastest. It was done in about 2 minutes, pulling from Codeberg. It took a lot longer to do the post-download git stuff, and now it is running Authentication on ~58k commits, and that's taking a bit too.
But at least it is moving, unlike when I tried to pull from Savannah, so that's something!
Maybe I should have given the VM more CPU and RAM.
Had to restart the VM, because it did run out of ram. After guix pull --url https://codeberg.org/guix/guix, guix system reconfigure /etc/config.scm no longer fails early. It is busy computing the derivation now, so I will have SSH soon!
And then I can continue doing what I wanted to do with it in the first place.
30 minutes later, it seems to have finished downloading packages (as part of guix system reconfigure), and is busy hogging the CPU without any output whatsoever.
I'm not entirely happy about this situation.
Just as I wrote that, it logged "updating substitutes", proceeded to download a bunch more stuff, and is now hogging the CPU again.
This is a tiny bit slow. But hopefully I won't have to do this too often.
Ooof. It looks like packaging Rust apps in #Guix involves listing all dependencies in the derivation. That's going to be An Experience™.
A space for Bonfire maintainers and contributors to communicate
