The CNAME record has been set for 2 days and it is still sitting there.
So I've move my site to Cloudflare now, but my domain doesn't work (yet).
So there are 2 options now:
- Wait for https://github.com/cloudflare/terraform-provider-cloudflare/pull/1835 to be merged and go Terraform all the way
- Package **wrangler** for AlpineLinux and use it in CI
Seems like doing option 2 in the mean time is reasonable.
I think I'll go with Cloudflare Pages since I already have a Cloudflare account for my work, and the Free tier seems more generous than Vercel.
What I lie to my friends: I'm a DevOps Engineer with high impacts on the company's services
What I actually do at work:
- Explain to people that Cloudflare is down when yarn doesn't work
- Teach devs how to use git
- Fix Gitlab CI job when devs can't release their code
- Approve any merge requests from devs with yaml files in them
- Delete people's databases (when upgrading Statefulsets)
- Calm people down when Cilium and Ceph can't heal themselves
Browse the task list maintained by @ta180m at https://gitea.com/Ta180m/gitea/issues/3, pick one and help #Gitea move towards #federation. If you are not sure how to help, just reply and someone in the @forgefriends community will guide you.
🚀 boost appreciated!
Quoted without comment or context from the answer I just wrote to a student question, requiring me to stare at the architecture manual for a full hour and then do several experiments:
In principle, pushl would decrement %rsp by 4. However, that instruction doesn’t actually exist; in 64-bit mode the only sizes of push that correspond to actual machine instructions are pushq and pushw, which decrement %rsp by 8 and 2 respectively.
I believe the only version of pushq that takes an argument smaller than 64 bits wide is pushq $xxxx where xxxx is a constant. It can be up to 32 bits wide and it’s zero-extended to 64 bits before pushing.
I can’t think of any circumstance where you would actually want to use pushw.
In 32-bit mode, pushw stays the same but pushq turns into pushl – yes, the exact same machine instruction changes its meaning depending on what mode the CPU is in. There are a bunch more cases like that. This is what you get when you start in 1980 with a 8-bit microcontroller, and keep adding extensions to it for forty years. It reminds me of my great-aunt-in-law’s house.
DT: systemd being "forced" on us is nonsense
- Remove systemd init. Oh this is easy, just choose 1 distro that doesn't use it.
- Now uninstall logind from your system. You use seatd? Great
- Next try getting rid of udevd. Good luck making wayland, pipewire, udisks2, lvm2, chromium, ... work!
I'll probably wait for the next mpv release and see how things go to make the switch.
For volume control, I wrote a small Lua script for the CLI usage, and use easyeffects Loudness module for preset profiles. Acceptable solution for me. There is also pw-volume which is a Rust wrapper for pw-cli.
The only obstacle left is whether there is a GUI / TUI / CLI for switching sinks / sources. Doing it with wpctl or pw-cli is pretty troublesome as there are a lot of device IDs.
FOSS devs soon: and before I wrap this README up, subscribe to my github account, smash the star button and don't hesitate to fork and submit that PR! Catch you all in the next repo!
❤️ FOSS and *kawaii* anime girls
Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.