<{^_^}>
#436 (by andir, 31 weeks ago, merged): Update to current nixpkgs-unstable
<andi->
ipdates php stuff O.O
<andi->
maybe
<andi->
let me see that link
<cole-h>
tbh I should probably update borg to flakes sooner rather than later...
<andi->
is flakes stable yet?
<cole-h>
OK nvm
<cole-h>
I forgot it's not
<cole-h>
because I've been using it so much lol
<andi->
I used it and I honestly am not sure of the benefits vs the costs
<cole-h>
Well, one benefit would be in this case: evals wouldn't be failing, because we'd be purely pinned to a specific nixos-unstable rev (but the reliance on NIX_PATH, I'm assuming, is causing it to use a "newer" nixpkgs)
<andi->
I'll have a look at php in a bit
<cole-h>
andi-++ Cheers, thanks.
<{^_^}>
andi-'s karma got increased to 36
<cole-h>
If you can't find time, I could try and just ask for your review :P
<andi->
also works for me ;)
<cole-h>
If only I had a way to test that the php stuff actually worked
<andi->
haha yeah
<andi->
I think I didn't even test it as gchristensen said there is no way for anyone but him to test it
<cole-h>
heh
<cole-h>
I think we should aim to increase the bus factor once he's generally available again. I hate bothering him every time Something™ goes wrong (especially now).
<andi->
What is that thing required for? Just the GitHub event arriving somewhere?
<cole-h>
Looks like it's subscription to some GitHub events, yeah.
<andi->
I am wondering if we couldn't just deploy that on an additional server and let that receive events as well. Eventually, once we verified things we could fail it over.
<andi->
I've no feeling for the volume of events tho..
<cole-h>
Probably about the same as number of PRs, I'd guess
<andi->
I'd expect the number to a a fair bit higher, all comments, edits, pushes, merges, … should arrive there, no?
<cole-h>
* Probably about the same as number of PRs times number of events, I'd guess
<cole-h>
:P
<cole-h>
gchristensen: I think something is up with buildkite + aarch64 -- seems like they're all "Waiting on concurrency group"