<veleiro>
why the hell would a services.nginx.virtualHosts.<host>.extraConfig add 2 of the same entry to an nginx file???
<veleiro>
okay, spend all day trying to solve this
<veleiro>
maybe yesterday too
<veleiro>
why would an extraConfig field for a virtualHosts entry just duplicate? i removed the option completely and it finally installs but only because it removed one of the duplicates defined in my config
<veleiro>
wtf
maljub01 has quit [Quit: maljub01]
maljub01 has joined #nixos-chat
<veleiro>
i dont understand at all, dont build commands rebuild everything and not append?
<elvishjerricco>
veleiro: That means some other part of your configuration is defining the exact same extraConfig
endformationage has quit [Quit: WeeChat 2.9]
<veleiro>
that's a good point, its this damn pull request module in progress
omnd has quit [Ping timeout: 246 seconds]
abathur has quit [Ping timeout: 260 seconds]
abathur has joined #nixos-chat
abathur has quit [Quit: abathur]
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui6 has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
abathur has joined #nixos-chat
buckley310 has quit [Quit: Connection closed for inactivity]
<siraben>
similar flavor (but not sure if it's related) is the shrinking of failing tests by the QuickCheck library
<siraben>
to produce a minimal counterexample such that a test fails
Dotz0cat has quit [Ping timeout: 264 seconds]
<elvishjerricco>
TIL you can't use hibernate if you have any zfs pools imported. Like, at all; you don't have to be trying to put swap on zfs (which you already shouldn't do). You just can't hibernate with a pool imported, or else your pool is likely to be destroyed when you import it.
<elvishjerricco>
Apparently does have the necessary suspend hooks, but not the hibernate hooks
<elvishjerricco>
I don't really understand how this destroys a pool though; in theory zfs is an atomic fs that never creates an invalid disk state that needs repair.
<elvishjerricco>
I guess if there were some in-flight writes, and upon resuming, the uber block is updated under the assumption those writes made it to disk?
<LinuxHackerman>
elvishjerricco: I'd guess that it's more related to the in-memory state of zfs. Though my dad used hibernate on his zfs laptop for a while and it never blew up 🤷
<elvishjerricco>
LinuxHackerman: I thought the point of hibernate was to serialize all allocated memory. How would zfs memory state get messed up?
<LinuxHackerman>
elvishjerricco: potentially in that the filesystem is still doing things while its memory is being persisted, resulting in an inconsistent state between the on-disk pool and the filesystem driver's state.
<LinuxHackerman>
That's just speculation though, for competent answers you might have to ask #zfsonlinux :D
<LinuxHackerman>
I have had some catastrophic issues with inconsistency between filesystem driver state and on-disk filesystem state before, though that was with windows and ntfs
<LinuxHackerman>
(hibernating, then taking the storage device out, playing around with the filesystem on another machine, then reinserting the storage device and restoring the hibernated state is a very bad idea)
<LinuxHackerman>
(and I'm sure that applies to linux as well as windows)
<elvishjerricco>
Oh wow yea that would be a bad idea lol
maljub018 has joined #nixos-chat
cole-h_ has joined #nixos-chat
eyJhbV2 has joined #nixos-chat
lassulus_ has joined #nixos-chat
maljub01 has quit [*.net *.split]
cole-h has quit [*.net *.split]
eyJhb has joined #nixos-chat
eyJhb has quit [*.net *.split]
lassulus has quit [*.net *.split]
eyJhbV2 is now known as eyJhb
lassulus_ is now known as lassulus
maljub018 is now known as maljub01
cole-h_ is now known as cole-h
__monty__ has joined #nixos-chat
BaughnLogBot has quit [Ping timeout: 246 seconds]
cole-h has quit [Ping timeout: 260 seconds]
BaughnLogBot has joined #nixos-chat
<eyJhb>
I have asked this again, but is there any Sonos alternative ? That doesn't cost my first porn
<eyJhb>
born* not first porn!! Damn it...
<__monty__>
WWFS
<joepie91>
eyJhb: I assume you mean the multiroom connected audio thing?
supersandro2000 has quit [Disconnected by services]
supersandro20003 has joined #nixos-chat
cosimone has joined #nixos-chat
<eyJhb>
__monty__: WWFS?
<eyJhb>
joepie91: Looked at that, and will look at it further. Sounds promising, but the... information about it is sparse/weirdly provided.
<eyJhb>
LinuxHackerman: Haven't found that before, will add it as well
<eyJhb>
Also joepie91 , last commit was in 2019 :/
<__monty__>
eyJhb: What Would Freud Say
supersandro20003 is now known as SuperSandro2000
Dotz0cat has joined #nixos-chat
SuperSandro2000 is now known as supersandro2000
<philipp[m]1>
lol, just wanted to text something to a friend and I noticed that I typed out "Oooooh, I finally understdood, how" and I have no idea any more what I did finally understand.
<philipp[m]1>
That message is from today, no less.
<joepie91>
LinuxHackerman: nah I recall the other thing having a considerably shinier website
<eyJhb>
philipp[m]1: I have done that before with writing on a blackboard, and then forgetting what I wrote
<eyJhb>
And not being able to read whate I wrote. Queue the game of unintended draw and guess
waleee-cl has joined #nixos-chat
slack1256 has quit [Ping timeout: 260 seconds]
cosimone has joined #nixos-chat
andi^ is now known as andi-
<Mic92>
How can compile tests in rust with a predictable name so I can run them in gdb?
<Mic92>
If cargo test --no-run would at least print the binary it compiled...
cole-h has joined #nixos-chat
rajivr has quit [Quit: Connection closed for inactivity]
<cole-h>
mmm, building noto-fonts-emoji from scratch
<cole-h>
nice
<veleiro>
wow so much work getting nixpkgs on ur own server
<sterni>
cole-h: never would have expected a font build system to have a huge python closure
<sterni>
cole-h: and every single one of the directly used library has a release every few weeks it's crazy
<cole-h>
Right? I was wondering why I was building some python... lol
<cole-h>
Though I guess I'm not really shocked, cuz Iosevka has a node build system...
<sterni>
cole-h: I remember when I noticed last year that our noto-fonts-emoji was one unicode emoji version behind
<sterni>
cole-h: I was like *clueless* next thing I know I maintain over a dozen new python packages :D
<cole-h>
hahaha
<sterni>
but the bad thing really is how slow the compression step is of the build system
<samueldr>
conventional package management doesn't care that much for the build closure for such "leafy" packages as fonts
<samueldr>
so font toolchains just go ham with all the dependencies they can
<sterni>
yep we could've just downloaded the ttf and be done with it
<sterni>
I think we switched back to downloading the built font for one emoji font recently
<samueldr>
it's a policy thing we should settle, whether we should prefer prebuilt fonts, or build them, and explore the trade-offs that implies
<samueldr>
(this is getting a bit off-topic for here)
<samueldr>
but compare this to source packages: as a policy it is preferred to use released source tarballs rather than going with a source control checkout
<sterni>
yep I think in that case there was an additional issue which contributed to the decision
<sterni>
I think it's a good policy overall and see no reason against it for fonts, really
<samueldr>
since fonts are much less of an executable... and official releases are prebuilt artifacts, it kind of makes sense
<samueldr>
squarely sitting on the fence!
<sterni>
I mean libreoffice and chromium still take much longer
<veleiro>
is 644 typical permissions for file creation with normal users
<cole-h>
yes
<sterni>
it's just really weird when things go wrong
<cole-h>
assuming a umask of 0022 (which is the default AFAICT)
<sterni>
my nixos-rebuild took half an hour on my notebook because of an emoji font? that's kinda bad
<veleiro>
thanks, never really realized i guess
<samueldr>
there is an assumption that "it should be cached, so why bother?"
<samueldr>
keyword: should :)
<samueldr>
and then there's package customization
<samueldr>
fonts are hard
* samueldr
goes back to writing a cursed binary
<cole-h>
veleiro: and then directories are 0755
* colemickens
yeets sway off his computer
* colemickens
"yeet" sure fell off fast
evanjs has quit [Ping timeout: 265 seconds]
<veleiro>
why why colemickens ?
evanjs has joined #nixos-chat
<cole-h>
colemickens: What're you switching to?
<cransom>
i realized that i'm old enough that when i finally catch wind of things like 'yeet', just by knowing it exists, it's already uncool.
<colemickens>
because my **** monitor is flashing multiple times a minute, just shuts off, stops working, yatta yatta. It's been reported on wlroots for ages, it broke sometime around 1.2->1.3 and just never ever has worked right since.
<colemickens>
(more rant elided)
<colemickens>
I'm going to try out the Plasma wayland PR, and try out the tiling script that was posted on reddit the other day.
<veleiro>
some fancy monitor tech like freesync ?
<colemickens>
It is a gsync enabled monitor, but I'm not sure that that would affect anything. i915 surely doesn't know anything about it or support it.
<veleiro>
hmm, sad
<veleiro>
i like sway because its minimal and always seems stable to me, cant imagine going back to the desktopmanager or x11 world
<cole-h>
Woot, running the zrepl module now that it's in nixos-unstable-small! Working great so far -- finally happy to have set up local replication to my backup disk
<cole-h>
infinisil: It was very easy thanks to your RFC42 work :)
<infinisil>
Hehe, was just wondering what this was for
<infinisil>
Although, I'm just seeing that you used type = types.submodule { freeformType = format.type; };
<infinisil>
Which is equivalent to just `type = format.type` :)
<infinisil>
No need for freeform modules if you don't need to define any type-checked options
<infinisil>
Although, I guess using freeform modules in such a case means that other modules (e.g. user-defined ones) can still specify options, even if this module doesn't
<veleiro>
I'm just looking at RFC-42 for this sourcehut PR
<cole-h>
infinisil: IIRC that's what the RFC body suggested, so that's what I did :P
<infinisil>
Oh darn
<cole-h>
infinisil: Or maybe I misunderstood it
<cole-h>
Looks like the freeformType is used when you want to specify other typed settings lol