<{^_^}>
"The year 2022 is 11 months, 30 days, 23 hours, 50 minutes, 36 seconds away!"
<infinisil>
:o
<bbigras>
we could skip 2021 and go straight to 2022.
<makefu>
yeah, pretty sure that the first half of this year will directly go into the bin
<gchristensen>
optimistic I see
supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos-chat
<andi->
realistic.
<makefu>
it is not like all restrictions in germany will disappear after the 10.01. when the hard lockdown will come to an end ( just like the virus )
<andi->
You think they'll end it on the 10th? hahaha
<makefu>
the hard lockdown will be lifted, but surely a lot of restrictions will be kept in place
<gchristensen>
andi-: you're not buying my pessimism?
<makefu>
i will only buy gchristensens pessimism if he can send via contact-free delivery
<makefu>
:)
<andi->
I do but I also do not think it is pessimistic but realistic that the first 6 months are just as garbage as the last.
<gchristensen>
ah
<andi->
maybe in april thing will improve
<andi->
when the weather is better but well..
<gchristensen>
I was excited to hear a friend was vaccinated last week
<andi->
I hope we pick up on the vaccination rate. By the current pace it will require >10y to vaccinate everyone in my state.
<makefu>
andi-: exactly what i think. when the weather is warm and most of the people which are at risk are vaccinated
<gchristensen>
same
<andi->
I am in the last group and I really do not mind that. I can work from at home and only have to leave here every few weeks to buy food if I really have to..
<gchristensen>
same
<andi->
but this isn't healthy long term.. 3 days with people already exhausted me a lot during xmas.. used to be just fine.
<makefu>
same same - i even think that 2021 will not be my vaccination year
<gchristensen>
...same...
<gchristensen>
plot twist: we are all the same person
<andi->
/o\
<andi->
we are all the same machine
<andi->
stuck in the same simulation
<andi->
otherwise this internet thing wouldn't ever work.
<andi->
unrelated: I managed to buy a not so bad webcam. Turned out to be some shady chinese brand, ignores white balance correction and gets hot even when turned off... Might as well send it back /o\
<andi->
only took me 8months to decide for one product.
neeasade has quit []
<gchristensen>
;_;
<andi->
2021 positive note: only 20 more days until this idiots is leaving the white house in the USA>
<abathur>
don't jinx us
<andi->
:P
<andi->
I am just observing with some popcorn...
<bbigras>
I managed to buy a webcam too this year. It only took 5 months to ship. the date was pushed back once.
<samueldr>
vaccination, I know full well it won't be before the end of the spring, possibly during summer, since they have a priority ordering set
<samueldr>
but it's nice to know it's being followed
<bbigras>
yeah.
<bbigras>
my gf should have it quickly because she works in a hospital (not on the front line, but with old people). but at the same time I always had a bit more fear because of her work.
<samueldr>
yeah, understandably
<bbigras>
I can't begin to imagine how it is for people on the front line and their family. maybe it just becomes the new normal for them. at least the vaccine is coming soon for them.
<abathur>
I get the impression it's like an emotional steamroller
<bbigras>
yeah probably. there must be many scares. and some friends/colleague catches it.
<andi->
Yeah, at least here the doctors at the front line in my family are being tested regulary. As in every day morning/evening.
<bbigras>
yeah
<bbigras>
it's amazing that the vaccine were made in a record time.
<andi->
still doesn't rule out transmission but gives them some sense of safety
<bbigras>
yeah
<cole-h>
infinisil: vw indeed
<cole-h>
P
<cole-h>
:
<infinisil>
accidentally hit some keys while cozying up in bed :3
<lovesegfault>
RIP MF DOOM
tilpner has quit [Remote host closed the connection]
spudly1 has quit [Ping timeout: 246 seconds]
tilpner has joined #nixos-chat
supersandro2000 has quit [Quit: Ping timeout (120 seconds)]
<colemickens>
now that I qjackcontrol works and I have pipewire I can wire a soundboard into my jitsi calls and make people think their discord is blowing up.
<noneucat>
hahahaha
<noneucat>
what's the state of pipewire these days? is it stable? :)
<drakonis>
yes
<drakonis>
its so good
<ashkitten>
colemickens: what do you use for a soundboard?
<colemickens>
I was just playing around with qjackcontrol, I think that's the answer? I can't get Carla to do what I'd expect, but I also don't necessarily know what I'm doing.
rajivr has joined #nixos-chat
<ashkitten>
can that play sounds?
<colemickens>
oh oh oh
<colemickens>
I just googled "discord sound board" and spam open linked until I found out that did what I wanted.
<colemickens>
(Some webpages would close the audio connection after a single sound, so I wouldn't be able to connect it.)
<ashkitten>
and can it disable all unmapped keys on that keyboard?
<gchristensen>
you'll have to do your own research on that one :)
<ashkitten>
ah
<energizer>
dont know much about pipewire but for me it is notable that paul davis (Ardour) has expressed some skepticism "Given the apparent similarities to gstreamer I have notable doubts about their fundamental design (for example, I’m not convinced that they’ve understood the importance of using a pull model at the lowest layers). " https://discourse.ardour.org/t/is-pipewire-the-answer/89148
<ashkitten>
i mean, it works great for me but i'm not a pro audio user
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
spudly11 has joined #nixos-chat
LnL has quit [Quit: exit 1]
LnL has joined #nixos-chat
lunc has quit [Read error: Connection reset by peer]
endformationage has quit [Quit: WeeChat 2.9]
kalbasit has quit [Ping timeout: 240 seconds]
kalbasit has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
tilpner_ has joined #nixos-chat
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
<noneucat>
a little off topic but what's the proper way to do attribution if you are working on an abandoned patchset?
<noneucat>
should i be adding some header like signed-off-by?
<samueldr>
signed-off-by is when another individual is involved with the final patch
<samueldr>
I think it's Co-Authored-By in your case
<noneucat>
got it!
<JJJollyjim>
in my experience, in some projects signed-off-by is mandatory and signifies acceptance of a CLA
<samueldr>
oh right, I got it mixed up with acknowledged-by
cole-h has quit [Ping timeout: 240 seconds]
__monty__ has joined #nixos-chat
cole-h has joined #nixos-chat
cole-h has quit [Ping timeout: 256 seconds]
<sphalerite>
Happy new year all!
<sphalerite>
noneucat: yeah, you don't want to put a signed-off-by in someone else's name. I'm not sure about any legal technicalities, but at least in linux's interpretation of signed-off-by, that's signing a document in their name.
<danderson>
lovesegfault: pong
<danderson>
2-day RTT, not great, not terrible :P
<lovesegfault>
danderson: :D
<lovesegfault>
Did you see my message?
<danderson>
not sure I did - where was it?
<lovesegfault>
It was here
<lovesegfault>
I asked if you lived in the east bay b/c I saw the picture you posted on twitter of some turkey you cooked and I wanted some
<lovesegfault>
lol
<danderson>
oh, sorry, my irssi sucks and halts scrollback after a few days
<danderson>
sadly no, Victoria BC - the great white canadian, uh, south :)
<lovesegfault>
:D
<danderson>
but the secret is the various slabs of fat at the start of the cooking :)
<danderson>
turkey gravy w/ a bunch of butter and broth, plus extra saved turkey drippings
<lovesegfault>
I'm not allowed to make meat at home :(
<danderson>
reheat, dump in turkey chunks, reheat, serve
<danderson>
super easy
<lovesegfault>
my wife is a vegetarian
<danderson>
(not pictured: 6 hours of cooking the previous day to bake the turkey)
<lovesegfault>
ha!
<danderson>
ah, yeah, this recipe is, uh, not vegetarian
<danderson>
I'm struggling to think of a single vegetarian ingredient in it
<danderson>
... the croutons?...
<lovesegfault>
I got her eating fish recently
<lovesegfault>
made some damn good black cod tonight
<danderson>
ooh, nice
lunc has joined #nixos-chat
cosimone has joined #nixos-chat
metalune[m] has joined #nixos-chat
waleee-cl has joined #nixos-chat
<gchristensen>
awk is so good. I'm trying to parse the output of nix-diff and just get the names which have changed: awk 'begin { inputnames=0} ($1 != "+" && $1 != "-") { inputnames=0 } /^\s+• The set of input names do not match:$/ { inputnames=1 } inputnames==1 && ($1 == "+" || $1 == "-") { print $1, $2 }' < diff
<eyJhb>
Late happy new years to you all, another 2-3 months of writing 2020
<infinisil>
Same, late happy new year!
<MichaelRaskin>
Hm, would softdog support (module built, and I guess loaded by default) be a useful addition to nixos-rebuild test, I wonder…
<MichaelRaskin>
(It seems to be indeed built by default)
<MichaelRaskin>
Wait, how come there is still no mode for reconfiguration without module reload? Enabling nowayout temporarily and specifying the timeout specifically for an upgrade would be so nice.
spudly11 has quit [Ping timeout: 272 seconds]
<infinisil>
Hm, does the order of records in bind matter?
<MichaelRaskin>
I would assume it serves them in the same order by default, and some software might always try the first DNS record first?
<infinisil>
I'm thinking of making a bind NixOS module that allows setting records with e.g. `records.NS.foo = { target = "example.com"; }`
<infinisil>
I guess I could just add a `priority` option in case an ordering is needed
<MichaelRaskin>
Presumably order of options relevant for different queries does not matter as much as for a single one
<infinisil>
Ah yeah
<MichaelRaskin>
I guess for NS you want a list to begin with, though
<infinisil>
Oh yeah but how should I handle multiple entries for the same type and name
* infinisil
doesn't like lists
* MichaelRaskin
do not like infinisil
<infinisil>
xD
<MichaelRaskin>
If not lists, you end up with inventing fake keys.
<infinisil>
Yeah, but that doesn't have to be so bad
<MichaelRaskin>
Or using merge logic that doesn't even need to summon an eldritch abomination because it is one.
<infinisil>
Forces giving names to record entries
<infinisil>
Hm, maybe there should be a merging mode for lists where it orders them according to the hash of the list contents
<infinisil>
So it's at least deterministic
CRTified[m] has quit [Quit: Idle for 30+ days]
<infinisil>
Though it still doesn't allow overriding of existing entries
cosimone has quit [Quit: cosimone]
<lassulus>
I think an attrset with an optional priority field would be better than a list
<lassulus>
lists are always hard to override from multiple files and merging is somewhat confusing
* infinisil
nods
cosimone has joined #nixos-chat
<infinisil>
Hm should it be `records.foo.NS = ...` or `records.NS.foo = ...`
<infinisil>
I think the former
endformationage has joined #nixos-chat
cosimone has quit [Quit: cosimone]
cosimone has joined #nixos-chat
<infinisil>
Argh, naming each entry is annoying though, I'll just go for a list
<MichaelRaskin>
If only we had some better list processing functions…
<MichaelRaskin>
(although, of course, overriding single elements of lists would still need module system support, too)
<viric>
I wish rmlint could decide based on equal mtime too
<viric>
ah -Z 0 !
lassulus has quit [Ping timeout: 246 seconds]
lassulus has joined #nixos-chat
andi- has quit [Remote host closed the connection]
lassulus has quit [Ping timeout: 264 seconds]
andi- has joined #nixos-chat
lassulus has joined #nixos-chat
lassulus has quit [Quit: WeeChat 2.9]
lassulus has joined #nixos-chat
rajivr has quit [Quit: Connection closed for inactivity]
<__monty__>
Hmm, would make already long messages quite a bit longer potentially.
<__monty__>
Why is the indirection a problem?
<samueldr>
there's context in the irc logs links
<samueldr>
but basically the git.io urls are hard to distinguish when piped into a tool to select which one to open
<samueldr>
the main reason *for keeping* git.io is it keeps the already long-ish messages shorter, I don't see any other reason
<__monty__>
I personally prefer keeping the noise to a minimum. Makes it easier to follow conversations. Less intimidating for beginners. There's tools that do link hinting right in text so people aren't obligated to use a method that loses context.
<__monty__>
Best of both worlds would be ideal ofc. I could imagine a shortener that keeps the issue number or a truncated commit hash.
<samueldr>
yeah, git.io is from github outright, so it's not really an external dependency
<samueldr>
adding an external dependency that may or may not decide to go away is not an easy decision
<energizer>
tbh i prefer that bot feeds be kept in a separate channel for people who want to follow them without blowing away people's questions in the main channel
<samueldr>
energizer: that's another discussion
<__monty__>
I know. There's also the trust in a shortener that you know will end up on github.
<cole-h>
We do have nixos-bot or whatever
<samueldr>
(and it was already had in the past, though it doesn't mean it's not time to discuss that again)
<supersandro2000>
wouldn't that also trigger to post information about it?
<cole-h>
No
<supersandro2000>
s owe get 2 messages everytime?
<energizer>
samueldr: i think the context matters. if it's in a feed-only channel, it's less important that the link be short
<cole-h>
Oh
<samueldr>
it shouldn't given that it's the same bot
<supersandro2000>
thats good
<cole-h>
Yeah, I thought {^_^} ignored itself
<samueldr>
energizer: definitely
<samueldr>
energizer: but that's not what the question is about
<samueldr>
energizer: right now it's about changing *only* that
<samueldr>
it's more that the IRC protocol doesn't send you your own messages back
<cole-h>
ah
<cole-h>
Whatever :P
<samueldr>
same end-result
<cole-h>
colemickens: ping
aleph- has quit [Read error: Connection reset by peer]
Church- has joined #nixos-chat
<abathur>
could offset some of the volume growth if it's possible to combine the paired "merged pull request" "pushed 2 commits" pair (though I assume that they come in via webhook and it'd be nontrivial if it's just a simple input/output)
<abathur>
and maybe pare down some of the other text
<sphalerite>
I wonder what percentage of git.io's namespace is due to us :D
<samueldr>
IIRC some github services directly give git.io URLs so I wouldn't think it's that big
<samueldr>
abathur: yeah, that's something that's on the backburner
<samueldr>
main issue is that right now it reacts to events as soon as they arrive
<abathur>
it may also be meaningful to skip the shortener for PR/issue links, but not for the compare links; it looks like a PR would go from 20->44, but the compare links are 20->68
<samueldr>
adding something that'll wait a bit before sending the events is annoying
<abathur>
especially since the compare links don't have an issue/pr ID in them anyways
<abathur>
and obviously can't be the target of the request
<samueldr>
abathur: good observation
* samueldr
just had an idea
<samueldr>
another issue with coalescing merge/push events is identifying the proper event to elide
<abathur>
nod, I probably wouldn't consider the complexity just for this; I mainly just said it in case it already had a state layer for dealing with possibilities like a duplicate message, or a delay followed by a burst of deliveries
<samueldr>
well, it's really something we want
<lukegb>
The "merged PR" messages are more useful than the "pushed commits to master" ones... unless the commits weren't just the PR
<samueldr>
lukegb: yeah
<samueldr>
now how do we coalesce them? :)
<lukegb>
Although e.g. pushed commits to haskell-updates *is* useful
<samueldr>
exactly
<samueldr>
we only want to elide the "push button" merge events
<samueldr>
I have to check if merge_commit_sha matches the sha of the push event
<abathur>
that seems right after reading for a few
<abathur>
I don't know if they ever come out of order
<abathur>
there might be some generic logic, maybe aggregate anything that matches this commit within 10 seconds, with a template for composing them based on event type?
<samueldr>
abathur: if they come out of order it's less of an issue if at some point there's a stray double
<samueldr>
it was more about figuring out a scheme to do it
<infinisil>
How about just not showing commits that are associated with a PR
<samueldr>
infinisil: tell that to the github webhook
<abathur>
nod, not an issue if it's rare
<samueldr>
we're being sent the two events
<infinisil>
Well we can filter it out
<samueldr>
infinisil: then do it
<samueldr>
"just" filtering it out is the issue
<samueldr>
all events are being handled just-in-time
<samueldr>
no knowledge of the past, no knowledge of the future
* infinisil
checks the API docs for the commit event
<abathur>
can we hook anything custom on the merge action?
<samueldr>
yeah
<abathur>
i.e., could we stuff a magic value into the commit on merge, that we get back from the push event, and use that to filter?
<samueldr>
though my main hangup is about accidentally leaking memory by doing something dumb
<samueldr>
there's no alternate thread where I can cull stuff from a list
<abathur>
sorry, not sure action was a good word
<samueldr>
but exactly
<samueldr>
I'll keep whatever sha the latest merge had
<infinisil>
Anybody happen to have the API reference for the commit event ready?
<samueldr>
and filter any push with that sha
<samueldr>
one ID to keep around
<abathur>
when the monkey clicks the merge button, can we do anything to affect whatever's in the data structure GH will eventually send on the push event
<samueldr>
it'll be extremely simple, nothing huge to keep around, no processing that can go weird
<samueldr>
a simple 1:1 match on strictly the most recent merge
<infinisil>
Ah but I guess just using the pull requests associated with the commit isn't that great, because such a commit being pushed doesn't mean that the PR is merged
<samueldr>
yeah, there's also that
<samueldr>
also, it would require API access
<samueldr>
right now it's all done strictly from the incoming messages
<samueldr>
(and from my local tests suite it seems to just work
<samueldr>
:| colour me unimpressed... I was about to add a feature to ignore bot events but it looks like the way I implemented a previous feature already does the hard work
<colemickens>
There are certain program derivations that reach into nixpkgs config to see if PA is enabled
<colemickens>
and if not, might build the application differently. It's a mechanism that isn't very clean imo
<colemickens>
You _might_ be able to ripgrep for "config.pulseaudio" and see how its used.
__monty__ has quit [Quit: leaving]
<samueldr>
okay, so on my end things are set: git.io urls will not be used for pull URLs, hexa-, but still for push events, and then push events from a merge event have been removed
<hexa->
let's try that
<samueldr>
yeah, I don't know when it'll be deployed
cosimone has quit [Quit: cosimone]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-chat
cole-h has quit [Client Quit]
<infinisil>
"Nearly all animal life is dependent on bacteria for survival as only bacteria and some archaea possess the genes and enzymes necessary to synthesize vitamin B12, also known as cobalamin, and provide it through the food chain."