Ah, `$(history -a)` runs in a subshell, so that value is never reset in the main shell process. The way to go is
```sh
declare -a PROMPT_COMMAND
PROMPT_COMMAND+=("history -a")
```
Unlike PS0, it would write history only after the command has finished, but I can live with that.
bash question: how to get it to append to history file immediately, w/o automatically making those commands propagated to all running shell instances?
`PS0='$(history -a)'` kinda does it, but it doesn't move "already appended" pointer, so every time it writes all commands from the current session.
Weird, looking at the source code - `history -a` should work without duplicating entries:
`PS0='$(history -a; history -c; history -r)'` (or other variations that include `history -r`) would avoid duplicating commands, but would also add to history commands from other shell instances.
bash question: how to get it to append to history file immediately, w/o automatically making those commands propagated to all running shell instances?
`PS0='$(history -a)'` kinda does it, but it doesn't move "already appended" pointer, so every time it writes all commands from the current session.
www.youtube.com/watch?v=4lKy...
*sigh*
Aww, thank you 🥹
Today is: the 1460th of February, 2022
Huh, that's an idea
How do I turn this crap off? Short of forking the frontend
I think A/B testing is the red herring here, and I get annoyed every time some web UI changes for any reason outside my control 🥲
With desktop software I can usually get some comfort of "oh fuck, rolling it back!", to at least get a bit of time to mentally prepare for the inevitable.
(I also have started the work to employ Redmine as the moderator-facing side github.com/bsky-watch/m..., but have lost interest a while ago)
I've got github.com/bsky-watch/l... for the "serving labels" side of things, but it has no moderator-facing side on it's own
People who use Forgejo: is there some special dance required for Actions to push docker images to the built-in container registry?
I'm authenticating with `${FORGEJO_TOKEN}`, but still get `unauthorized: reqPackageAccess`
One issue I take with them though is that the heatsink absolutely sucks. Just a piece of aluminium, no heat-pipes, and heat transfer rate is very bad. It kinda works with the default 15W power limit, but at 35W it can't cool down the CPU below 80-90°C even with less than 1 core worth of CPU load.
And I want, at some point, to take them out of their cases, design and 3D-print required parts, and shove all 4 of them into a single 1U enclosure.
++
Couple months ago I was able to get 4x de.gmktec.com/en/products/... (while they were 300 EUR each). Now I'm running OpenEBS, CloudNativePG, MinIO, and other stuff on them.
Update: account restored, reason for suspension was automated sending of moderation reports that was running for almost 2 years (I even forgot that it was there, LOL).
🍿
Yup, and nobody is going to do anything about it. We've already seen this w.r.t. russia since 2014.
Same
Thanks, I did send an appeal shortly before posting, but haven't got any reply yet. I'll keep waiting, since there's no other options available to me.
I still don't see a point in doing so 🙃
In practical terms there's very little difference between an account being taken down at a PDS level, AppView, or the relay.
Hey, @bnewbold.net , may I bother you to check what's going on?
one of the weirder things AI has given us is aggressively commented and rigourously formatted malware payloads
At my current stage of life, exceptions to "mostly" are much more painful to deal with on Windows than on Linux 😅 But that's due to me being comfortable using various diagnostic tools and not being shy to read the source code (including kernel) when I don't understand from docs how it should work.
Not doing any of that, to the best of my knowledge. I've paid quite a bit of attention to properly handle rate limiting in particular.
