<samueldr>
assuming all refs are in my git repo, any tricks to see all commits for which a given rev is the parent of?
patagonicus6 has joined #nixos-chat
<samueldr>
ah, turns out I'm a bit inattentive and I mostly have the info I want already
patagonicus has quit [Ping timeout: 264 seconds]
patagonicus6 is now known as patagonicus
<samueldr>
scratch that, I don't have the info I want
<samueldr>
but I found better, I think, since I was XYing myself
<samueldr>
git branch --contains is actually what I want I think
<samueldr>
I hope I have a branch for everything I want
<supersandro2000>
😂 note: currently hard linking saves -1.09 MiB
<supersandro2000>
also: 54837 store paths deleted, 317538.11 MiB freed
patagonicus7 has joined #nixos-chat
wirew0rm1 has joined #nixos-chat
wirew0rm has quit [Ping timeout: 240 seconds]
patagonicus has quit [Ping timeout: 240 seconds]
eyJhb has quit [Ping timeout: 240 seconds]
eyJhb has joined #nixos-chat
eyJhb has joined #nixos-chat
betawaffle has quit [Ping timeout: 240 seconds]
pasukon has quit [Ping timeout: 240 seconds]
patagonicus7 is now known as patagonicus
aszlig has quit [Ping timeout: 240 seconds]
hplar has quit [Ping timeout: 240 seconds]
pasukon has joined #nixos-chat
betawaffle has joined #nixos-chat
hplar has joined #nixos-chat
feepo has quit [Ping timeout: 242 seconds]
aszlig has joined #nixos-chat
feepo has joined #nixos-chat
eyJhb has quit [Ping timeout: 240 seconds]
eyJhb has joined #nixos-chat
eyJhb has joined #nixos-chat
<ashkitten>
how does passing a fd over a unix socket actually work? isn't a fd just an integer? how does the kernel know how to map the file in the other process environment?
slack1256 has quit [Remote host closed the connection]
BaughnLogBot has quit [Ping timeout: 260 seconds]
BaughnLogBot has joined #nixos-chat
sdier has quit [Ping timeout: 240 seconds]
sdier has joined #nixos-chat
<lovesegfault>
ashkitten: it triggers a special codepath in the kernel, AIUI
<lovesegfault>
well, the kernel _knows_ the sender is sending a fd b/c you construct the msg in a funky way
<Ke>
my system has F(n,m) where each power of 2 index has n entries for m smallest powers of 2
<Ke>
like if entry x is divisible by 2^k k<m, it can occupy on of the (n+1)k smallest index slots, if it can't fit it'll get deleted
<colemickens>
Is it possible that timedrift breaks systemd-resolved's ability to allow-downgrade from DNSSEC?
<colemickens>
This is the second time I've lost tailscale on my rpi4 from... what I think is time-drift, I guess? I just fixed clock and rebooted systemd-resolved and ... things started working again.
<colemickens>
not even sure why this is happening, timedatectl says it should be syncing
<lovesegfault>
colemickens: the rpi is really tricky with time
<lovesegfault>
and by really tricky I mean it sucks
<srk>
you can run chrony + gpsd for really precise time :)
<srk>
with PPS input as well
<lovesegfault>
srk: how does gpsd work?
<supersandro2000>
default chrony should be enough
<srk>
lovesegfault: you attach external gps over usb/uart, it can do nmea or ubx (ublox) protocol and it can feed time to chrony/ntpd
<srk>
to get even more precise time you can also hook PPS (pulse per second) pin from GPS to some rpi pin and kernels PPS support as additional source
<srk>
with that you get stratum 1 ntp server for your local network and it works even when internet goes down :)
<lovesegfault>
TIL srk is sorki :)
<srk>
:)
BaughnLogBot has quit [Ping timeout: 265 seconds]
BaughnLogBot_ has joined #nixos-chat
BaughnLogBot_ is now known as BaughnLogBot
rajivr has quit [Quit: Connection closed for inactivity]
<patagonicus>
So, yeah, turns out services work better with "foo.enable = true;". I had changed one to false in the config instead of removing the settings because it was temporary, but then I only checked if the service appears in the config when I wanted to check if it was running. Sigh.
<philipp[m]1>
> timeto mars
<{^_^}>
undefined variable 'timeto' at (string):477:1
<philipp[m]1>
> timeTo mars
<{^_^}>
"9 hours, 33 minutes, 17 seconds"
BaughnLogBot_ has joined #nixos-chat
BaughnLogBot has quit [Ping timeout: 260 seconds]
BaughnLogBot_ is now known as BaughnLogBot
supersandro2000 has quit [Read error: Connection reset by peer]
supersandro2000 has joined #nixos-chat
<infinisil>
> timeTo mars
<{^_^}>
"7 hours, 12 minutes, 34 seconds"
<LinuxHackerman>
>timeTo mars
<LinuxHackerman>
> timeTo mars
<{^_^}>
"6 hours, 51 minutes, 33 seconds"
<gchristensen>
eee
<sterni>
> timeTo 1666666666
<{^_^}>
value is an integer while a set was expected, at /var/lib/nixbot/lib/date.nix:45:13
<sterni>
> mars
<{^_^}>
{ day = <CODE>; hour = <CODE>; minute = <CODE>; month = <CODE>; second = <CODE>; year = <CODE>; }
<sterni>
does it support leap seconds 🤔
<sterni>
infinisil: is date.nix on github?
<gchristensen>
it supports leap seconds in the same sense thath unix does: 1 day is 8600 second
<LinuxHackerman>
8600 seconds? Are those Mars days or something?
<LinuxHackerman>
Days that fast would probably result in noticeably less perceived gravity.
<LinuxHackerman>
Unless you're on the poles.
<gchristensen>
throw a 4 in there somewhere, close enough :P
<sterni>
gchristensen: yep, this is mostly correct, but not for timeTo then
<sterni>
gchristensen: or it is correct in a sense because it returns it in unix secndos which may either be 0, 1 or 2 seconds iirc
<gchristensen>
I'll reimburse you the cost of the ISO-8601 spec if you implement it correctly in Nix
rajivr has joined #nixos-chat
<gchristensen>
(and account for leap seconds of course)
<gchristensen>
until then, hopefully nobody is doing leap second math in nix
<cole-h>
And then `git rebase -i --autosquash origin/<base>`
supersandro2000 has joined #nixos-chat
<cole-h>
(which will automatically move around the fixup and squash commits to where they belong, as well as changing their "rebase action" to match)
<cole-h>
Despite the name, it works for `fixup!` commits as well :P
<srk>
now there are drafts so it's good but I've had a couple fixup commits merged
<bbigras>
but couldn't github suggest using the "squash merge" or something when there's multiple commits? it's a bit silly having to do a force push after clicking an "accept suggested change" on github.
<cole-h>
GitHub's merge semantics are just silly to begin with
<samueldr>
story of my recent life: I have a `git fetch` going since over 12 hours ago
<cole-h>
lol
<supersandro2000>
bbigras: nope
<srk>
bbigras: accept suggested change is broken when there are multiple commits
<bbigras>
doesn't it just do a new commit?
<supersandro2000>
it does
<srk>
it does but then you need to manually split it if changes are not related
<supersandro2000>
squash merging does not work when we want to have 2 commits
<supersandro2000>
just do the changes locally and push them
<srk>
would be cool to be able to say `don't allow merging while there are fixup! or XXX/draft prefixed commits`
<bbigras>
"squash merging does not work when we want to have 2 commits" isn't most PR only 1 commit?
<srk>
supersandro2000: if it's part of suggestion someone already made the change so why repeat locally :)
<leons>
the sane option is to integrate those automatic commits based on suggestions such that they represent meaningful commits with atomic changes
<leons>
I fear any fancy integration GitHub would implement would just cause it to create an even worse workflow
<srk>
we need something like git-fixup-blame with github client :D
<leons>
srk: what's github client?
<srk>
anything using its API I guess
<supersandro2000>
bbigras: squash merging just merges all together which is not always desirable
<leons>
i have git for most of what I do? i don't see how one would need any API integration for a rebase?
<srk>
I mean, pull the suggestion locally, apply, add/fixup to respective commits
<veleiro>
wow yeah that's why i was asking
<supersandro2000>
srk: because I not always do complete suggestions and miss eg inputs
<veleiro>
it makes sense to amend your changes before having it merged, instead of adding to them
<supersandro2000>
and having automatically changes in nixpkgs is not a real option if we do not want to explode the commit count
<supersandro2000>
yeah but if you squash them when it is ready it just takes another round back to you
<supersandro2000>
if you take a few days or weeks then maybe a treewide change got in and you need to change something.
<veleiro>
would make reviewing easier to track the changes over a review
<bbigras>
supersandro2000: I know but it would be desirable most of the times. Maybe the merging person could handle it.
<bbigras>
it's not rocket science
supersandro2000 has joined #nixos-chat
<sterni>
srk: I wish
<veleiro>
guess i need to check out nixpkgs-review
<bbigras>
I heard it's 30 secs /s
<veleiro>
when you do reviews, do you add diffs? suggestions dont seem helpful if they lines weren't originally changed
<bbigras>
what do you mean? by suggestions I meant the diff ones that you can apply in 1 click and it marks them as resolved and create a commit.
<veleiro>
well in github they have suggestions but those are only allowed for the lines that are included in the PR's diff
<veleiro>
cant add a "suggestion" to something at the top of the file (like a forgotten dependency) when it had no changes on the diff
<veleiro>
just want to contribute them properly
<joepie91>
<samueldr> the basic "github-like PR" workflow is not an issue AFAIK, what is, is the complete inability to extend the tooling to our additional needs
<joepie91>
well, and the unicorn herds escaping every once in a while :P
<samueldr>
when talking about _workflow_, 500s are not exactly relevant
<samueldr>
what I was bringing on the table is that we don't need to uproot the process of "opening up a PR" if we explore other tools
<samueldr>
though, what really I want to see: multiple "views" and interfaces over the contribution process
<veleiro>
the problem is trying to build on these centralized toolings instead of ones that are platform-agnostic, just as said above about not being able to addon to github
<veleiro>
i dont know how bitcoin puts up with github either
Dotz0cat has joined #nixos-chat
<infinisil>
Booo they delayed the stream
<infinisil>
> :v mars
<{^_^}>
mars = date.parseDateTime "2021-02-18 19:15:00"
<infinisil>
> mars = date.parseDateTime "2021-02-18 19:30:00"
<{^_^}>
Perseverance landing stream start in ^^ https://www.youtube.com/watch?v=iPXuw8SJfFI: Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred]
<samueldr>
apparently github doesn't really cope well with huge repos (e.g. linux, maybe nixpkgs) being cloned --shallow, and then git fetch --unshallow
<bbigras>
I did a --shallow clone once. and made a PR. and it was a clusterfuck .github notified every single code owners.
pie_40 has quit [Ping timeout: 240 seconds]
pie_43 has joined #nixos-chat
<infinisil>
That guy has an odd accent
<infinisil>
German-like, but kinda also not
<ar>
infinisil: feels kind of odd to use "launch" for something that's explicitly a landing ;)
* infinisil
nods
<philipp[m]1>
Would "spaaaaaaace" be better?
<samueldr>
those spacex launches that don't leave the atmosphere don't go to space!
<samueldr>
other than the US designation of 50 miles which seems... pretty arbitrary... it's all about a single metric that matters, e.g. where atmospheric drag becomes a thing, or when lift stops being a thing
<philipp[m]1>
If I didn't cut my hair for that long, there would probably be hunters after me because they mistook me for a bear.
<infinisil>
Um um um
<samueldr>
I uh... started cutting my hair again *coincidentally* the same month that everything started happening
<infinisil>
That's supposed to happen!
<colemickens>
(this is random, sorry) - gchristensen: I was reading scrollback and saw your zfs tweet. It's a git-like source control system where the working copy is always committed, it keeps track of history across rebases (and tracks conflicts as well). Very different in implementation from what you're doing, but not a terribly dissimilar end result. https://github.com/martinvonz/jj It felt neat when I was kicking the tires on it,
<colemickens>
albeit lightly documented.
<samueldr>
maybe I shouldn't?
<gchristensen>
oh cute
<infinisil>
Yeah, long hair ftw
<bbigras>
philipp: they can't hunt all of us!
<infinisil>
An engineer!
<philipp[m]1>
Watch the technical stream that I linked. It's much more relaxed.
<infinisil>
philipp[m]1: Oh no it has mono audio, only on the right side
<philipp[m]1>
This is why you should have prepared by installing pipewire :)
<bbigras>
was it on the left side earlier?
<bbigras>
but yeah, thanks for the link philipp
<philipp[m]1>
The streams are quite a few seconds apart.
<colemickens>
lovesegfault: what did you end up doing? I am going to look into if systemd-timesyncd should be running more aggressively or something, I guess.
<philipp[m]1>
nice x11 mouse in their animation there.
<lovesegfault>
colemickens: I just use timesyncd and cross my fingers
<philipp[m]1>
Choo chooo chooo!
<colemickens>
lovesegfault: great minds think alike, thanks :)
<philipp[m]1>
Nice!
<bbigras>
\o/ yay for the rover! at least something didn't suck this year
<philipp[m]1>
There is a helicopter on mars!
<bbigras>
fuck yeah
<gchristensen>
"Attitude Control" I've worked with people who needed a whole Attitude Control department
<bbigras>
keep doing stuff like this USA.
<bbigras>
haha
<bbigras>
the first picture was fast
<sterni>
lol are they using kde
<gchristensen>
let's make sure nix is on the next rover
sorear has quit [Remote host closed the connection]
Guest88372_ has quit [Remote host closed the connection]
jared-w has quit [Remote host closed the connection]
<bbigras>
with or without flakes?
<lovesegfault>
Who will get there first, man to the moon or flakes to nix?
<lovesegfault>
s/moon/mars/
<lovesegfault>
I need coffee
<bbigras>
hehe
Guest88372_ has joined #nixos-chat
sorear has joined #nixos-chat
<cransom>
just give us 1% of nasa's budget
<AMG>
It's gonna be sick seeing that helicopter fly on mars. Hopefully it works!
<philipp[m]1>
Yeah, I wonder how soon they'll try it. They probably need some time to check all their systems out.
<bbigras>
yeah
slack1256 has quit [Ping timeout: 260 seconds]
jared-w has joined #nixos-chat
pie_43 has quit [Ping timeout: 240 seconds]
<andi->
so did they land on mars or was it all just a simulation? Has anyone checked? :P
<omnd>
as if martians doesn't have anything to fly or travel.. they waited for humans to help lol
pie_86 has joined #nixos-chat
slack1256 has joined #nixos-chat
<omnd>
you've made me want to watch The Martian again
<omnd>
i loved the book.. way more science in it than the movie included.. but the movie was great too
<cole-h>
+1
Dotz0cat_ has joined #nixos-chat
Dotz0cat has quit [Ping timeout: 256 seconds]
<flokli>
lovesegfault: there's already ice on mars, probably flakes too.
<philipp[m]1>
omnd: The movie will always have a place in my heart because they included the airlock scene.