<{^_^}>
nix#3006 (by matthewbauer, 4 weeks ago, open): Run test to determine if sandboxing is available
<gchristensen>
it would be really nice to have this in for the next release, whenever that is :P
noonien has quit [Quit: Connection closed for inactivity]
<infinisil>
Profpatsch: Yeah I agree with you, that's a bad idea, left a comment too
orivej has quit [Ping timeout: 272 seconds]
<yorick>
the only issues you could autoclose is updates to packages that are already updated by something else
<ekleog>
Profpatsch: this proposal minus the “auto-close” part of it is not that bad, though
<andi->
I just installed a new NixOS machine based on master and it fails to communicate with cURL via SSL/TLS. It looks like the OpenSSL 1.1 change might be responsible for that. Nix is unable to download packages /o\
drakonis1 has quit [Ping timeout: 246 seconds]
drakonis1 has joined #nixos-dev
<gchristensen>
uh oh, my RFC39 integration might have been blocked by github
<clever>
gchristensen: ?
<Taneb>
Adding a suspicious amount of organization members in a short time?
<gchristensen>
I think so
<gchristensen>
going to give it a rest and then ask support about it
<globin>
andi-: hmm, works fine for me, any special config?
<gchristensen>
globin: quite the timing for the nomination :)
<gchristensen>
I really wish the NixOS project had a solid contact, like an account manager, with GitHub
<yorick>
seems like RFC-0051 is hugely divisive
<gchristensen>
it is certainly not clear cut
<globin>
yeah it needs some clarification and further work :)
abbradar[m] has joined #nixos-dev
abbradar[m] is now known as abbradar
<andi->
globin: it seems to be something unrelated to the changes... Still not sure where that came from. I used tethering from my phone so might be some spooky middlebox there.
<adisbladis>
ekleog: Agreed, I think a triage bot pinging issues infrequently would be pretty nice.
<clever>
gchristensen: ah, the cause of the invites i keep hearing about
__monty__ has joined #nixos-dev
<emily>
I'm packaging libfx2, which consists of two parts: embedded C firmware to be compiled with sdcc and then flashed to a device, and a Python module plus programs to interact with said device. I can easily make a libfx2 package that contains a bunch of firmware (like nixpkgs.blackmagic), and I can easily package the Python module and programs, but should these be merged into the same package or kept separate? (they're in
<emily>
the same git repo, just different subdirectories)
<emily>
and if they should be merged, how should I achieve that? not sure how to mix buildPythonPackage and a custom build together
<Taneb>
My gut is they should be separate, but have the git repo only specified onces
<emily>
so `inherit (otherpkg) src` basically? that seems reasonable
drakonis has joined #nixos-dev
<Taneb>
To make sure that the two are kept in sync
<emily>
yeah
orivej has joined #nixos-dev
<timokau[m]>
Huh, as a nixos org member I didn't even have to accept the nixpkgs maintainers invitation. Just got added automatically. Interesting.
FRidh has joined #nixos-dev
carter has quit [Excess Flood]
carter has joined #nixos-dev
<thoughtpolice>
emily: Yeah, `inherit (upstream) src` is a good way to do it
<thoughtpolice>
emily: Normally, if the same src package ships both the module and the python thing, I will just make a 'default.nix' and 'python.nix' in the same directory in the tree. Then just `import` them from somewhere else
<emily>
it turns out setup.py actually does the build itself, although I should maybe override that
<thoughtpolice>
(That package uses a slightly different mechanism for unpacking the source, but you can ignore that)
<emily>
(because e.g. it doesn't put it in a very useful location and I can't figure out how to hook into the actual pre-install with buildPythonPackage)
<emily>
(since the "build" gets run at "install time", afaict)
<thoughtpolice>
Yeah, that happens sometimes and is annoying. A lot of packages that do that will normally let you use the "system installed" version of some C code, but if setup.py drives the whole build, it sounds like that won't be doable.
<jtojnar>
maybe mention in the posts that it should only be pinged when a serious problem occurs
<ryantm>
Seems like they should use @NixOS/nixpkgs-committers in that case
<FRidh>
which should also be avoided
<gchristensen>
we should not be pinging ~100 people to merge a single PR
<gchristensen>
that is the purpose of the team, and the purpose of ofborg being able to request reviews of maintainers
orivej has quit [Ping timeout: 258 seconds]
ixxie has joined #nixos-dev
drakonis has quit [Quit: WeeChat 2.5]
drakonis has joined #nixos-dev
drakonis1 has quit [Ping timeout: 246 seconds]
drakonis has quit [Ping timeout: 250 seconds]
__monty__ has quit [Quit: leaving]
orivej has joined #nixos-dev
evanjs has quit [Quit: Configuring ZNC, sorry for the joins/quits!]
evanjs has joined #nixos-dev
FRidh has quit [Quit: Konversation terminated!]
ixxie has quit [Ping timeout: 245 seconds]
orivej has quit [Ping timeout: 268 seconds]
drakonis has joined #nixos-dev
orivej has joined #nixos-dev
<Profpatsch>
I get a lot of notification spam, but I think most of it is because of wrong rebases or similar
ajs124 has quit [Quit: Gateway shutdown]
psyanticy has quit [Quit: Connection closed for inactivity]
ajs124 has joined #nixos-dev
<emily>
when a Python package has a build_ext (i.e. native code) stage in its setup.py, it seems like running tests for it is kind of broken; build_ext gets executed another time (breaking in my case since I need to do the build manually to customize it slightly)
<emily>
am I missing something there?
<emily>
(ah, never mind, I reproduced it without Nix locally so I guess it's some deeper problem)