<cole-h>
gchristensen: I'm gonna hold off on PRing this until I get in touch with srk to see if they're cool with me cherry-picking their hacking section from the Cachix PR while I'm already touching the README
<gchristensen>
cool :)
<gchristensen>
sounds good
<cole-h>
Quick sanity check: tests run on all possible machines, right? (e.g. if it's `handleTest`, will run on darwin, x64, and aarch)
<cole-h>
The "Note: this will only run on x86_64-linux machines." attached to the test subcommand is what's confusing me
<gchristensen>
the tricky bit is no PR with currently failed checks will get a green check if it is magically fixed by an ew eval
<gchristensen>
but that is probalb yfine
<gchristensen>
do it! :)
<cole-h>
Could also do it in a follow-up
<gchristensen>
meh, go for it
prusnak has joined #nixos-borg
<MichaelRaskin>
Could you _remove_ grahamcofborg-eval checks but not create new ones?
<MichaelRaskin>
Or maybe always set them to gray-box for the first stage of migration?
<cole-h>
Harrumph. I don't want to deal with that now. I'll undo that change for this PR and think about it more... *waves hands* in the future
* cole-h
adds TODO item
<cole-h>
gchristensen: Mind taking another look at the README changes (especially the most recent `fixup`, which slightly rewords the "Trusted Users" section) when you have a chance?
<cole-h>
MichaelRaskin: You so frequently have good insight. I'd be interested to hear your thoughts too, if you have some spare time: #482
<MichaelRaskin>
Oh nice, now GitHub diffs are also JS-only.
<MichaelRaskin>
cole-h: is draft status detected now?
<cole-h>
Do you mean for the WIP stuff? No, we don't use that as a metric for WIP.
<cole-h>
s/metric/whatever suitable word you prefer/
<MichaelRaskin>
Also, maybe drop @GrahamcOfBorg mention from documentation even if it still works for those who remember it?
<MichaelRaskin>
Maybe mention that Draft status does not matter then
<cole-h>
Only reason I hesitate doing so is because `@GrahamcOfBorg` does indeed still work (and probably will for the forseeable future)
<cole-h>
Will mention draft status, good point.
<MichaelRaskin>
Well, this is the user documentation, you tell people how to use it, not the precise limits of what works
<MichaelRaskin>
You do not mention the precise timeouts either
<cole-h>
Fair enough.
<MichaelRaskin>
Actually, you do not mention the timeout _existence_
<MichaelRaskin>
(So building Chromium is not just rude, it is pointless!)
<cole-h>
Huh, I hadn't thought about that. Totally forgot we have timeouts :D
<MichaelRaskin>
All committers are trusted users, right?
<cole-h>
They must be added manually via PR.
<MichaelRaskin>
Still?
<MichaelRaskin>
Oh
<cole-h>
I don't know if there are any committers that aren't a trusted user, but for any new ones they must be added to the list in config.public.json
<MichaelRaskin>
Actually, given the regular discussions about FO derivations, I am not sure how good an idea it was to build everyone's PRs, not just limited. say, to maintainers
<cole-h>
(brb, just froze my entire session)
cole-h has quit [Quit: Goodbye]
<MichaelRaskin>
Hmm. Does anyone without commit access to ofborg run any builders nowadays?
cole-h has joined #nixos-borg
<cole-h>
OK, I am back. re builders: Not that I know of. Pretty sure that stopped a while ago (before I got involved)
<MichaelRaskin>
That's also what I assume
<MichaelRaskin>
I mean, maybe then mark this chapter as «if you do not know, you do not need this»
<cole-h>
Meaning the "Running a builder" section, yeah? Maybe it could just go to the wiki page linked at the end of it.
<MichaelRaskin>
Maybe that
<cole-h>
That page is probably due for a cleanup, anyways