<srhb>
I just read the Hackage downtime article and it mentions Rackspace terminating FOSS donations. Is this a general thing, and if so does it affect us (Hydra?) as well?
<domenkozar>
srhb: I think it says "in case of"
<domenkozar>
This is important so we can have an accurate inventory of the services we use, in preparation for the eventual termination of Rackspace’s donations of infrastructure to FLOSS projects
<srhb>
domenkozar: Okay, I'm reading that like "they've warned it's going to happen" - I hope you're right. :)
<domenkozar>
srhb: I think we have a few machines as slaves
Lisanna has quit [Remote host closed the connection]
Jackneilll has joined #nixos-dev
Jackneill has quit [Read error: Connection reset by peer]
taktoa has quit [Remote host closed the connection]
JosW has joined #nixos-dev
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
JosW has quit [Quit: Konversation terminated!]
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
orivej has joined #nixos-dev
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
<gchristensen>
domenkozar, srhb: rax IS terminating their FOSS contributions
<srhb>
:(
<shlevy>
:(
<domenkozar>
I guess it's the typical 2-3 yearly cycle
<domenkozar>
when new mangement sets new policies
<gchristensen>
we've had them a long time, and it has been like ~1yr since they said they were going to shut them down last december :P
<domenkozar>
correction: new ineffective policies
<gchristensen>
:D
<gchristensen>
my suspicion is they don't want to be the ones to shut down our boxes, and I'm pretty sure we won't shut them down early ...
<domenkozar>
new ineffective, not thought-through policies
<domenkozar>
gchristensen: did they give us a date when they will go offline?
<gchristensen>
last December
<gchristensen>
iirc :)
orivej has quit [Ping timeout: 256 seconds]
<domenkozar>
maybe we'll get an invoice next december
* domenkozar
googles if rax was bought by oracle
<domenkozar>
nope, we're good :)
<gchristensen>
:D
<domenkozar>
trivia: owner of rax bought painting "the scream" for $220m
<gchristensen>
lol
<gchristensen>
one of the founders is why I didn't go to college
<domenkozar>
gchristensen: which one?
ma27 has quit [Ping timeout: 255 seconds]
ma27 has joined #nixos-dev
<gchristensen>
Pat Condon
<gchristensen>
* not _why_ exactly, but part of why :)
ma27 has quit [Ping timeout: 240 seconds]
ma27 has joined #nixos-dev
pxc has joined #nixos-dev
drakonis has joined #nixos-dev
pxc has quit [Ping timeout: 240 seconds]
ma27 has quit [Ping timeout: 276 seconds]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 240 seconds]
ma27 has joined #nixos-dev
tilpner has quit [Quit: :wq]
tilpner has joined #nixos-dev
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
<copumpkin>
shlevy: you around? I'm seeing lots of errors in 2.0.1 about "claims to be content-addressed but isn't" and I'm wondering if you have any hunches about what might cause it. Everything was fine in 2.0
<copumpkin>
haven't figured out a simple repro yet
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
ma27 has quit [Client Quit]
ma27 has joined #nixos-dev
Lisanna has joined #nixos-dev
<bgamari>
is there any way to determine the ordering which extraConf options are merged in?
<bgamari>
where the cups services.printing.extraConf overrides LogLevel
<bgamari>
samueldr, hmm
<samueldr>
(that's a guess though, not sure if it applies to the string types)
<samueldr>
well, lines in that case
<LnL>
pretty sure it applies to all types that can be merged
<bgamari>
I'm not sure the right way to deal with this issue
<samueldr>
great, that was what I was thinking, but couldn't find actual proof in the types source file
<pierron>
bgamari: it is.
<bgamari>
would there be opposition too just making LogLevel an option of its ownw?
<bgamari>
pierron, what is it?
<pierron>
mkOrder is a way to ensure that you have some ordering between the definition, before it gets merged.
<bgamari>
right
<bgamari>
I didn't know about this
<bgamari>
that being said, it does seem a bit fragile
<pierron>
it is part of the module system.
<LnL>
I'd say it makes sense to have separate options for things that depend on ordering
<bgamari>
yes, I see that
<bgamari>
LnL, alright
<pierron>
if you want something less fragile, look at the activationScript mechanism, and you will get back to mkOrder ;)
<LnL>
since it's not clear what the 'default' ordering will be
<bgamari>
LnL, right
<samueldr>
I had a small question, about the 18.03 channel advance: build 73334587 has a blank space for chromium, does it mean that chromium isn't cached for those updating?
* bgamari
spent quite a while trying to work out why his changes weren't taking effect before I looked at the cupsd.conf
<LnL>
but for special cases mkAfter/mkBefore can be used
<srhb>
samueldr: No, it's cached.
<srhb>
samueldr: The test was removed, so it wasn't _guaranteed_ that it was cached though
<srhb>
That change was reverted.
<srhb>
samueldr: (It would be clearer if nixpkgs.chromium was a direct constituent of nixos.tested instead of transitively through nixos.tests.chromium)
<samueldr>
so, if I understand correctly, even though the build step #2 in build 73202439 "timed out", it still built it?
<srhb>
samueldr: To be honest I'm no longer sure how it got there, but you can verify that the path is indeed there with nix-store --realise /nix/store/h1pbpcwjmzqzr13r5ng3m4cdx3wf9q0i-chromium-66.0.3359.117,
<srhb>
samueldr: Because I'm reading it like you are.
<samueldr>
the log that "timed out", if it helps, looks like the previous good build; it finishes with the same lines
<srhb>
I... Hydra must be confused. :P
<samueldr>
though, d'oh, realising the store path is a good way to verify thanks!
<srhb>
We did put it through quite a bunch of confusing exercises, and hopefully it picks up the new timeout next time an evaluation triggers. Unfortunately the change to meta.timeout doesn't trigger evaluation on itself.
zybell has joined #nixos-dev
pxc has joined #nixos-dev
pxc has quit [Ping timeout: 264 seconds]
peti has quit [Ping timeout: 248 seconds]
peti has joined #nixos-dev
pxc has joined #nixos-dev
ma27 has quit [Ping timeout: 256 seconds]
ma27 has joined #nixos-dev
mingc has quit [Quit: WeeChat 2.0.1]
pxc has quit [Ping timeout: 240 seconds]
drakonis has quit [Remote host closed the connection]
<bgamari>
What is the protocol for requesting backport of a version bump to a nixos release?
zybell has quit [Ping timeout: 240 seconds]
<samueldr>
minor revision or major revision? as for minor, I have been told to either ask for a cherry-pick or create a MR with the cherry pick; cherry pick protocol explained here https://nixos.org/nixpkgs/manual/#idm140737315816544
<thoughtpolice>
bgamari: There's an odd amount of actual commits from cups-filters 1.20 to 1.20.2 strangely, since a few bumps were done and whatnot and some maintainer lists changed, etc.
<LnL>
not sure how it's implemented, it's possible inherit gets translated pretty early in the process so it might not be possible to distinguish between inherit foo; and foo = foo;
<ekleog>
whew this sounds like it'd bump memory consumption by quite a lot, like word + {line, col, path} per variable assignment
<gchristensen>
probably so :')
<ekleog>
likely no big slowdown, though?
<ekleog>
depends on whether you'd want a = b + c to have two parents b and c or not, though :°
<gchristensen>
that would be a new thing
<gchristensen>
I wouldn't mind if this was a run-time switch, or even a patch on top of nix :P
<gchristensen>
but I think being able to ask (and answer) "where the heck did this come from?" is a useful thing
<ekleog>
hmm actually nix isn't garbage-collected, so it'd likely be something like 28 bytes per variable assignment… that's python-level :D
<gchristensen>
nix is GC'd... it has boehm
<ekleog>
oh it is? so 24 bytes (never looked inside the source code, thought it was “vanilla” c++)
<srhb>
gchristensen: I wonder if some of the new annotation work on hnix could help there.
* srhb
is super hyped for that
<gchristensen>
what is the annotation work?
<LnL>
yeah, maybe it would need to be a new builtin but something that tries to find the position of a value recursively until it finds a nontrivial expression should be possible
<ekleog>
oh I hear annotation and hnix in the same sentences… do you know if comments are coming to the AST annotations? :D