MichaelRaskin has quit [Ping timeout: 256 seconds]
mbrgm has quit [Ping timeout: 256 seconds]
mbrgm has joined #nixos-dev
contrapumpkin has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
contrapumpkin has joined #nixos-dev
<gchristensen>
I have a simple way to implement PR sampling, I wonder what y'all think: during the PR's evaluation, if the user is a trusted user, parse commit messages for the subject package (ie: "postfix: foo bar baz" -> postfix) and if the package is a buildable attribute according to the evaluation, schedule a build job for it
<gchristensen>
this is dead simple, and comes with the benefit of encouraging strict following of the commit format. if there is a real benefit from being precise, perhaps andi- would have committed "gitAndTools.grv: init at 0.1.0" instead of "grv: init at 0.1.0" ;)
<andi->
gchristensen: d'oh, yes /o\
<andi->
but yes I would like that..
<gchristensen>
it is a shame my unemeployment is ending :)
<andi->
hrhr, same here.. only 16 days left of freedom
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm has joined #nixos-dev
s33se1 has joined #nixos-dev
s33se has quit [Ping timeout: 264 seconds]
orivej has quit [Ping timeout: 240 seconds]
tilpner_ has joined #nixos-dev
tilpner_ has quit [Remote host closed the connection]
tilpner_ has joined #nixos-dev
s33se has joined #nixos-dev
s33se has quit [Client Quit]
s33se1 has quit [Quit: WeeChat 2.0]
s33se has joined #nixos-dev
goibhniu has joined #nixos-dev
<domenkozar>
this job in is "sending inputs" for 16h
<pie_>
would it be possible to have a hash for downloaded archives that is of the original archive and not the nar, so as to check against hashes or the archive provided by the distributors?
<viric>
?
<pie_>
i mean fetchurl has the sha256, but sometimes developers/mirrors will provide .sha256 files and the like which are hashes of the original file
<LnL>
it's the same unless fetchzip is used, but we only do that for dynamically generated archives like the github tarballs
<pie_>
oh huh i guess i remembered wrong then
<pie_>
so its just the store hashes that are differeny
<LnL>
ah yes a lot of the hashes in nixpkgs are base32, you'll want to convert those when comparing
<pie_>
0i mean the store hashes will be different anyway because theyre a hash of something else
orivej has joined #nixos-dev
<viric>
pie_: you want to store the hash of a hash?
<dtz>
Sonarpulse: hey, what's the deal with "stdenv.glibc"?
peterhoeg has quit [Ping timeout: 248 seconds]
peterhoeg has joined #nixos-dev
<pie_>
dtz, i think i ran into that at some point
<pie_>
iirc something to do with being able to use other libc later
<pie_>
(iirc)
<pie_>
for example musl
* dtz
is working on nixpkgs musl goodness
<dtz>
but still, seems strange that it's not stdenv.cc.libc, not sure what it's actually meant for
<pie_>
mmm mu(e)sl(i)
<dtz>
might be cruft from gnu/hurd support? lol
<dtz>
xD
tilpner has quit [Disconnected by services]
tilpner has joined #nixos-dev
<pie_>
im probably mixing it up with somethng else
<pie_>
dtz, see if github blame takes you to an issue or pr or something on its definition :P
<dtz>
:D
<dtz>
fwiw "git blame" is totally a thing, don't need github
<pie_>
yeah but does git blame give you prs?
tilpner_ has quit [Quit: :wq]
<pie_>
then again i guess this would probably be pre-github lol
tilpner_ has joined #nixos-dev
tilpner has quit [Remote host closed the connection]
tilpner_ is now known as tilpner
<dtz>
xD
<LnL>
isn't stdenv.libc a thing?
<LnL>
oh, only on darwin :/
<dtz>
stdenv.cc.libc
<dtz>
since w/o a compiler you don't have a libc. Maybe.
<pie_>
unrelated, i cant wait to see what nix will be like in a couple years
ma27 has joined #nixos-dev
__Sander__ has joined #nixos-dev
Sonarpulse has joined #nixos-dev
FRidh has joined #nixos-dev
<dtz>
LnL: random but fingers crossed vim bump helps your rendering lag O:)
<LnL>
nope :/
<LnL>
funny thing is that I can't reproduce it with my vim-nix highlighting, only other languages
<LnL>
I think it's related to relativenumber somehow
<shlevy>
Sonarpulse: Any reason anything you did recently might break haskell.lib.justStaticExecutables on darwin? Getting an error "-r and -dead_strip cannot be used together"
<shlevy>
Not sure why we'd suddenly be buildiing with -r...
<Sonarpulse>
shlevy: what is "-r"?
<Sonarpulse>
shlevy: I don't think so
<Sonarpulse>
but its possible
<shlevy>
Relocatable
<Sonarpulse>
shlevy: hmm odd
<shlevy>
I have this sinking feeling I'm about to bisect through a million world-rebuilds...
<gchristensen>
shlevy: you can tell git to skip them if you'd like
<gchristensen>
basically something like (nix-build . -A stdenv -j0 || git bisect skip)
<domenkozar>
shlevy: some libs build with -r
<domenkozar>
like if you have debugging/profiling on
<domenkozar>
it's the reason why it's not the default on darwin
__Sander__ has quit [Quit: Konversation terminated!]
<shlevy>
Aaah profiling is the issue, OK
<shlevy>
Thanks
<shlevy>
niksnut: How hard would it be to respect NIX_CURL_FLAGS inside nix itself? It's looking like we need --negotiate to deal with an annoying proxy issue
<domenkozar>
shlevy: ideally we'd catch this in generic builder
<domenkozar>
and assert
<domenkozar>
shlevy: hopefully my hie opinion wasn't harsh, I didn't want to descredit your work
<shlevy>
Nah, no worries
<shlevy>
I wish they'd just cut proper releases though
<domenkozar>
ye
<niksnut>
shlevy: sounds hard, unless libcurl happens to provide a function to parse curl command line flags
<shlevy>
Damn, OK
<shlevy>
This is the kind of thing that really should go through env vars like http_proxy etc... :/
<niksnut>
also it would tie us to curl
<shlevy>
Yeah
taktoa has joined #nixos-dev
* contrapumpkin
waves weakly
ma27 has quit [Ping timeout: 252 seconds]
<jtojnar>
according the 1.12 slides, `nix-build --hash` is supposed to replace `nix-prefetch-*` but I am getting "tar: This does not look like a tar archive" when I want to hash a patch file (https://bugzilla.gnome.org/attachment.cgi?id=367356)
<fpletz>
will anyone here be at the fosdem pre-beering at delirium tomorrow evening? :)
propumpkin has joined #nixos-dev
contrapumpkin has quit [Ping timeout: 256 seconds]
propumpkin is now known as contrapumpkin
<gchristensen>
contrapumpkin: you ok?
<contrapumpkin>
yup, why?
<gchristensen>
your weak waving :)
<contrapumpkin>
oh, a bit sick, but otherwise fine :)
<gchristensen>
ah :)
viric has quit [Quit: memtest]
<clever>
jtojnar: `nix-build --hash` has to be pointed to a fixed-output derivation that downloads something
<jtojnar>
clever: how is it upposed to replace nix-prefetch-url then?
<clever>
jtojnar: ignoring prefetch-url, the old option was to run nix-build, wait for the fixed-output to fail its hash check, then read the error, fix it, and re-build
<clever>
jtojnar: but, it would discard that first download, and have to download it a second time after fixing
<clever>
jtojnar: nix-build --hash, will just rename the 1st download, to match what its hash is, so the 2nd build finds it, if you fix the hash in the nix right