lopsided98 has quit [Read error: Connection reset by peer]
lopsided98 has joined #nixos-dev
orivej has quit [Ping timeout: 245 seconds]
{^_^} has quit [Read error: Connection reset by peer]
{^_^} has joined #nixos-dev
<ekleog>
have there already been talks about adding a nixpkgs-18.09 channel, that wouldn't be tested?
Ericson2314 has quit [Ping timeout: 250 seconds]
<LnL>
that doesn't sound like a good idea, and I'm not sure what the usecase would be
<srhb>
samueldr: One possible explanation is that happy originally failed, so servant-lucid originally failed, and then servant-lucid was restarted with happy succeeding?
<srhb>
The time stamps seem to corroborate this
<srhb>
(It would be nicer if somehow we didn't reuse build numbers on restart...)
<ekleog>
then, the use case I quoted isn't really the actual one I'm thinking for myself: what I am currently thinking of is setting myself up a subset of the tests that I want to pass before my machine upgrades nixpkgs, and disregard tests outside of these
<ekleog>
(typically, I don't use chromium, so don't care if this test fails)
<ekleog>
so in order to have updates reaching me faster, it'd require having some way to point at the latest commit built by hydra, hence nixpkgs-18.09 :)
<LnL>
if you look at the channel update graphs you'll see it updates very frequently already
<LnL>
an untested channel won't get you things faster
<LnL>
unstable is a different story, because of the large changes that happen there daily
<sphalerite>
ekleog: sounds like nixos-18.09-small is what you want to me..?
<sphalerite>
It's not completely untested, but it does update faster
<ekleog>
sphalerite: hmm well, indeed that's the second-best option, I guess :)
<ekleog>
LnL: well, looking from the status tab, 18.03 has not seen a successful update for 4 days, while there have been 5 evaluations since then :/
<ekleog>
sphalerite: actually yeah, looking at 18.03-small it looks like it almost never fails a bump, so that indeed looks like what I'm looking for, thanks! :)
goibhniu has joined #nixos-dev
<LnL>
yes, it happens but those edgecases are not a good enough reason to add a dangerous channel IMHO
<ekleog>
well, 18.03 regularly is a week behind 18.03-small -- though that does appear to imply that 18.03-small actually updates frequently enough :)
<ekleog>
(maybe not a week but at least a few days, according to your graphs)
copumpkin has quit [Ping timeout: 244 seconds]
contrapumpkin has joined #nixos-dev
__Sander__ has quit [Ping timeout: 240 seconds]
__Sander__ has joined #nixos-dev
<andi->
I've tried to work on the 18.03 channel failures yesterday but some of tests are just so flaky.. I did run the `ferm` test locally 10x and it failed 4x :/ (on an idle machine)
<srhb>
andi-: Do you have a log of the failures? Perhaps we can robustify the tests.
<andi->
srhb: at home on my notebook probably.. Will try to check in the evening
<srhb>
andi-: Please do, hopefully we can find a common pattern and try to improve any weak spots. Feel free to ping me. :)
<andi->
i tried to figure it out and it just times out on a curl request that isn't bein reset it seems.. There isn't really much to it. Sometimes it works. Somtimes ti fails. THe logs were (almost) identical to those on hydra.
<srhb>
I would say if you had 4 failing out of 10 locally, that's a great start, because usually the issue is that they're irreproducible locally.
<andi->
I initially thought that the default policy for that chain changed and started adding an explicit ACCEPT/REJECT line at the bottom of the rules until I saw that it didn't really influence the outcome just caused the tests to be run again (since input changed)
<andi->
anyway, have to get some (boring) work done :(
aminechikhaoui has quit [Ping timeout: 252 seconds]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Ping timeout: 245 seconds]
lopsided98 has quit [Ping timeout: 252 seconds]
aminechikhaoui has joined #nixos-dev
lopsided98 has joined #nixos-dev
<nbp>
srhb: Ok, after having a short discussion with the github support, they are going to forward a private message to the user. I also contacted the support of Factorio and they scrambled the password of the user and sent him an email too.
<sphalerite>
nbp: the world needs more people like you! :D
<nbp>
sphalerite: I respect others as much as I would expect other to do for me ;)
<sphalerite>
hm does this work here too?
<sphalerite>
nbp++
<{^_^}>
nbp's karma got increased to 1
<sphalerite>
it does!
<srhb>
nbp: Great work :)
jtojnar has quit [Ping timeout: 252 seconds]
jtojnar has joined #nixos-dev
lopsided98 has quit [Ping timeout: 264 seconds]
lopsided98 has joined #nixos-dev
orivej has quit [Ping timeout: 252 seconds]
contrapumpkin has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
copumpkin has joined #nixos-dev
orivej has joined #nixos-dev
* ekleog
curious: who is this user you're talking about? (if it's ok to tell)
<ekleog>
ok, looks like the parser is more strict than I thought
<samueldr>
or not++
<{^_^}>
or not's karma got increased to 2
<ekleog>
wtf
<ekleog>
is that space a thin space or something like that?
<samueldr>
(off-topic, but explanation is: using a non-breakable space, as long as there are no spaces (ascii spaces) it will increase karma)
<ekleog>
:)
<sphalerite>
to type a nbsp, shift+altgr then space twice, iirc
<samueldr>
depends on many things, but COMPOSE, then double space does it
<samueldr>
or for the "ca" keyboard, altgr+space
<sphalerite>
I think shift+altgr is compose by default?
<samueldr>
no idea if there is a default
<ekleog>
here on bepo it is maj plus space iirc++
<{^_^}>
here on bepo it is maj plus space iirc's karma got increased to 1
<nbp>
ekleog: I avoided mentioning the name, nor the issue, nor replying on the issue to avoid getting more attention to this issue than needed.
<nbp>
ekleog: as I do not know if this user used this password for something else, I think it would be best to not bring too much attention to it, even if one of his/her password got invalidated.
<clever>
sphalerite: my keyboard also has a dedicated compose key
<clever>
but both compose and shift+altgr dont work in xterm
<nbp>
clever: Do you use one of the default xkeyboard-config, or managed to get a custom keyboard layout?
<clever>
nbp: only change to my keyboard layout is to make capslock into shiftlock
<ekleog>
nbp: indeed :)
<clever>
compose just types 9~
<clever>
and shift+altgr+space does nothing
<nbp>
clever: I tried to add a custom xkeyword-config layout, but I do not want to recompile everything which depends on it :/
<clever>
422 xserver = {
<clever>
423 xkbOptions = "caps:shiftlock";
<ekleog>
ISTR Nadrieril eventually found the solution to this one
<ekleog>
hmm he's neither here nor on #nixos, though :(
<Dezgeg>
if there's a solution I'm interested as well (for a particular finnish dvorak layout)
<nbp>
clever: I am using a modified version of https://qwerty-lafayette.org/ , which is loaded with xkbcomp at the beginning of my session
<clever>
nbp: ah
<nbp>
So much better for coding.
<samueldr>
nbp: compared to? (azerty?)
<nbp>
samueldr: compared to AZERTY and QWERTY.
<samueldr>
I'd be curious to see an azerty user compare the "canadian french" layout, the one with the dead keys, and some accented keys
<nbp>
it has french accents with a composed key, symbols with RAlt
<samueldr>
(hi nbp, I'm from Québec)
<nbp>
(hi)
<nbp>
I do not know the Québec layout.
<nbp>
or canadian french, because there is more than Québec which is speaking French.
<samueldr>
there are two, one is a mess, the other is closer to a basic qwerty (and none are specific to québec, but the one that's imho a mess is the one preferred by the OQLF)
<samueldr>
(we can move to #nixos-chat because it's now really off-topic :))
<nbp>
What is on topic for me, is how can we add new keyboard layout without recompiling everything?
<ekleog>
nbp: I'll be poking you as soon as I get an answer on whether nadrieril did manage to do it or not and if so how :)
__Sander__ has quit [Quit: Konversation terminated!]
<andi->
can someone abort / restart the vm test that have been in "sending inputs" for >10h? https://hydra.nixos.org/build/81214248 I doubt they'll finish some time soon :)