gchristensen changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | 18.03 release managers: fpletz and vcunat
{^_^} has quit [Remote host closed the connection]
gchristensen has quit [Ping timeout: 240 seconds]
{^_^} has joined #nixos-dev
gchristensen has joined #nixos-dev
ma27 has quit [Ping timeout: 246 seconds]
Lisanna has quit [Quit: Lisanna]
pxc has joined #nixos-dev
<ekleog> hmm… is there a policy about packaging one's own programs for nixpkgs? I've just completed https://github.com/Ekleog/todiff/ , and know at least a friend also using nixos who plans on using it, but it feels unfairly biased, as it'll maybe only ever have like two users
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm has joined #nixos-dev
orivej has joined #nixos-dev
Sonarpulse has quit [Ping timeout: 256 seconds]
pxc has quit [Ping timeout: 276 seconds]
orivej has quit [Ping timeout: 240 seconds]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 240 seconds]
<sphalerit> Sonarpulse: instructinos, are those instructions that have nearly no observable effect and execute extremely quickly?
<sphalerit> ekleog: I wouldn't hesitate to add it anyway, although you could of course also just provide a default.nix in the repo, then it can be installed with nix-env -f https://github.com/ekleog/todiff/archive/master.tar.gz -i
<sphalerit> Or an overlay
<MichaelRaskin> Then again, third-party software packaged in Nixpkgs is also not always crazily popular.
pie_ has quit [Ping timeout: 240 seconds]
jtojnar has joined #nixos-dev
pie_ has joined #nixos-dev
ashgillman has joined #nixos-dev
MichaelRaskin has quit [Quit: MichaelRaskin]
ashgillman has quit [Ping timeout: 256 seconds]
ashgillman has joined #nixos-dev
davidlt_ has joined #nixos-dev
davidlt has quit [Ping timeout: 240 seconds]
pie_ has quit [Ping timeout: 265 seconds]
vdemeester` has quit [Ping timeout: 256 seconds]
terrorjack has quit [Ping timeout: 256 seconds]
ocharles has quit [Ping timeout: 256 seconds]
vdemeester` has joined #nixos-dev
ocharles has joined #nixos-dev
terrorjack has joined #nixos-dev
vdemeester` has quit [Changing host]
vdemeester` has joined #nixos-dev
__Sander__ has joined #nixos-dev
Synthetica has joined #nixos-dev
<Mic92> We have a program that just draws a horizontal bar in the terminal.
pie_ has joined #nixos-dev
ashgillman has quit [Ping timeout: 264 seconds]
ashgillman has joined #nixos-dev
ma27 has joined #nixos-dev
pie_ has quit [Ping timeout: 264 seconds]
pie_ has joined #nixos-dev
<dtz> Lolwut
<dtz> Which is that? Haha
<LnL> there are multiple packages that just print hello world
<makefu> like hello, but this can do so much more
<dtz> :(. Well "hello" at least is a build and autotools test, didn't know about the others
<dtz> Haha
<makefu> like `hello --traditional`
<dtz> Well hopefully they're at least small expressions. Soon (fetchGit?) we'll need to start sending channel deltas lol :P
orivej has joined #nixos-dev
davidlt_ has quit [Remote host closed the connection]
davidlt_ has joined #nixos-dev
davidlt_ has quit [Remote host closed the connection]
davidlt has joined #nixos-dev
ashgillman has quit [Ping timeout: 264 seconds]
<Mic92> dtz: hr
* dtz half-expected it to require multiple versions of GHC
<Mic92> dtz: I procrastinated a bit and made this: https://github.com/Mic92/x86_64-linux-cheatsheats
<dtz> err not to bash haskell, other than pandoc manages to be required for too many things! :D
<Mic92> to no longer wasting time on google
<Mic92> soon we need pandoc in our boostrap binaries
pie_ has quit [Ping timeout: 252 seconds]
<dtz> lol right
<dtz> or maybe nixpkgs "generations" O:)
<dtz> "nix: the next generation"?
<dtz> "nix: the final frontier"
<dtz> "nix: the wrath of khan"? lol
<Mic92> one need the previous release to build the next
<Mic92> just like rust
<Mic92> go 42
<ekleog> sphalerit: ok, thanks :)
davidlt_ has joined #nixos-dev
davidlt has quit [Ping timeout: 240 seconds]
<Profpatsch> Gah, nix 2 doesn’t have a switch to turn off color ANSI codes for error messages. :(((
<Profpatsch> Or at least for nix-instantiate --parse.
<Profpatsch> That broke the emacs flycheck syntax check.
<Profpatsch> Mic92: What’s go 42 mean?
<Profpatsch> catch-22? :)
<Mic92> Profpatsch: /go 42 -> go 42 (weechat channel switcher plugin)
<dtz> Profpatsch: quick, add an option for it and submit as a PR!
<Profpatsch> dtz: It’s not that simple, it seems like lots of places add ansi-error codes all over the place.
<dtz> bonus points if it allows forcing color to be used when writing to non-terminal
<dtz> hmm
<dtz> was hoping would/could be fixed like: https://github.com/NixOS/nixpkgs/issues/37114
<dtz> similar use case I think
<Profpatsch> Search for ANSI_
<Profpatsch> Or for ESC_, there’s two different places that have #defines for the same codes apparently.
<Profpatsch> #define ESC_RED "\033[31m"
<Profpatsch> #define ANSI_RED "\e[31;1m"
<Profpatsch> readonly RED='\033[38;31m'
<dtz> does the commit fixing the linked issue not address your problem? https://github.com/NixOS/nix/commit/486872150638d56483c2bc429ba9e137d9974ee8
<Profpatsch> I think the correct way would be to make --json work for --parse
<Profpatsch> Oh wow, there’s a function that filters those out again.
<Profpatsch> Yeah, that should fix it.
<dtz> xD
<dtz> <3
<Profpatsch> Syntax checking works again, nice.
<LnL> Sonarpulse: didn't we get rid of NIX_CXXSTDLIB_COMPILE?
<clever> Profpatsch: oh, one min
<clever> what was that package called...
<clever> Profpatsch: there is a header-only c++ package, that gives you ansii color codes
<clever> ah, found it
<clever> std::cout << termcolor::red << "Hello, Colorful World!" << std::endl;
<Profpatsch> clever: Needed the inverse, though. :)
<clever> parsing?
<clever> dtz: oh, nix why-depends always uses color, it doesnt auto-detect
<Profpatsch> Removing the escape sequences from strings.
<clever> Profpatsch: ahh
<clever> idealy, the program should istty(stdout) and detect
<clever> and also allow forcing it to any state
<clever> grep and many other tools do that
<clever> ah
<dtz> file issue for why-depends? and maybe for forcing colors on? O:)
<clever> oh, i also get this one in namespaces: https://github.com/NixOS/nix/issues/1996
<clever> yep, nix-user-chroot ones, same as him
pie_ has joined #nixos-dev
Sonarpulse has joined #nixos-dev
<Sonarpulse> dtz: OK fixed the test infra
<Sonarpulse> those packages (or their deps) need meta.platforms adjusted to not fail
JosW has joined #nixos-dev
orivej has quit [Ping timeout: 256 seconds]
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aszlig has quit [Quit: leaving]
aszlig has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
ma27 has quit [Ping timeout: 240 seconds]
pie_ has quit [Ping timeout: 268 seconds]
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
ma27 has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
Lisanna has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
__Sander__ has quit [Quit: Konversation terminated!]
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
Lisanna has quit [Quit: Lisanna]
aminechikhaoui has quit [Read error: Connection reset by peer]
ma27 has quit [Ping timeout: 256 seconds]
aminechikhaoui has joined #nixos-dev
ma27 has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
aminechikhaoui has quit [Read error: Connection reset by peer]
aminechikhaoui has joined #nixos-dev
ma27 has quit [Ping timeout: 256 seconds]
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-dev
acowley_ is now known as acowley
Synthetica has quit [Quit: Connection closed for inactivity]
tilpner has quit [Quit: :wq]
tilpner has joined #nixos-dev
<Mic92> dtz: Do you know can I get a musl-gcc like wrapper in nix? I would like to switch in nix-shell between the glibc/musl.
<Mic92> *know how
tilpner has quit [Quit: :wq]
tilpner has joined #nixos-dev
<dtz> Mic92: I think "easiest" is something like: $ nix run -f channel:nixos-18.03 --arg localSystem '{config="x86_64-unknown-linux-musl";}' stdenv.cc
<dtz> maybe Sonarpulse has a shorter incantation, lol
<Sonarpulse> no shorter one yet
<dtz> it'd be easier if nix tools didn't insist on using runCommandCC or whatever; they always want to pull in default stdenv :(
<Sonarpulse> yeah
tilpner has quit [Quit: :wq]
tilpner has joined #nixos-dev
<Mic92> this could be enabled (optionally?) in our musl expression I guess?
<dtz> maybe, but those wrappers won't have deps/etc handled and built w/musl
tilpner has quit [Client Quit]
<dtz> I forget why we disable wrappers, but there's a reason. I think something breaks that's non-obvious lol
tilpner has joined #nixos-dev
<dtz> but yeah those should work pretty well for local use
<dtz> I use llvmPackages_6.stdenv and llvmPackages_6.libcxxStdenv in a musl-based nixpkgs all the time ^_^
<Mic92> I know
<Mic92> For the time being I need as little libc as possible
<Mic92> and glibc does malloc()/brk() before calling main(). I cannot have this at the moment.
<dtz> Mic92: maybe it was just the multiple outputs thing, which I think your gist fixes? hmm
<dtz> even so I'm a smidge reluctant to provide a half-way solution when a proper solution is available :3
<dtz> but I suppose that RFC is still lingering.. :)
ma27 has joined #nixos-dev
<Mic92> dtz: well, it could be also just for compatibilty sake with projects that expects `musl-gcc` to be availab.e
<Mic92> or users...
<Sonarpulse> angerman: I fixed the overrides thing on nixpkgs master
<Sonarpulse> we should get my PR to your nix windows example working to your liking
<Sonarpulse> merged a few bits of your mingw+ghc nixpkgs pr too
pxc has joined #nixos-dev
<LnL> Sonarpulse: didn't we remove NIX_STDLIBCXX_COMPILE?
<Sonarpulse> LnL: no we didn't, I don't think
<LnL> oh really?
<Sonarpulse> nop!
<Sonarpulse> could we make pkg-config wrapper a thing? :)
<Sonarpulse> then we can properly get rid of it
<LnL> thougt we did, noticed because of that recent pr
<Sonarpulse> LnL: NIX_CXXSTDLIB_COMPILE
<Sonarpulse> LnL: it does sound a little familiar
<LnL> :p
<Sonarpulse> basically it has nothing to do with cxxstdlib
<Sonarpulse> just that one of the libc++ put some of the headers in a subdirectory
<LnL> well but is the environment variable actually used anywhere?
Synthetica has joined #nixos-dev
<Sonarpulse> LnL: let me keep grepping
<Sonarpulse> LnL: nah in cc-wrapper
<LnL> from what I recall we got rid of it, just didn't remove the variable itself to avoid a rebuild
<Sonarpulse> on master
<LnL> :/
<Sonarpulse> :(
<Sonarpulse> IT'S ALIVE
<LnL> darnit
JosW has quit [Quit: Konversation terminated!]
<Sonarpulse> LnL: OTOH now that I fixed meta.platforms
<Sonarpulse> I guess we can go work on darwin->linux again
<dtz> Sonarpulse: when trying to build ghc cross-musl (native doesn't work because we are using a glibc-based bootstrap GHC, so figured I'd start with trying to obtain such a thing...) I'm getting what /seems/ like a non-musl error-- it's.. failing to find ncurses?! lol
<dtz> do you know anything about that? haha
<LnL> if we could get past the kernel build that would be breakt!
<LnL> great* :)
<Sonarpulse> dtz: huhhhh
<Sonarpulse> weird
<Sonarpulse> a long time ago seen such things
<Sonarpulse> but I thought I've sinced fixed that
<dtz> yeah it seemed like some strangeness from cross infra, at least more so than obviously anything else. Curiously (!) this does not happen w/aarch64-musl lol :(
<Sonarpulse> dtz: also cross, you mean?
<dtz> yes
<Sonarpulse> dtz: I know ghc itself has its own dynamic linker
<Sonarpulse> for some ghci things
<dtz> it actually was working ~5d ago, and I think works on master (!)
<Sonarpulse> don't think that would be at play
<Sonarpulse> master is super broke on hydra at least
<Sonarpulse> that "lto is not a language" bug again, sigh
<Sonarpulse> I fixed all the tests that went away at least
<dtz> cached path: /nix/store/lcw90icbaakys5c1vi4ixvjsyriycxk1-hello-1.0.0.2-aarch64-unknown-linux-musl
<dtz> lol :(
<dtz> okay don't mean to nag just thought I'd ask since I'm ghc and espec ghc-cross+nixpkgs noob and that's your domain ^_^
<dtz> JIC it's "oh yeah I forgot to push that fix" :D
<dtz> comment re:having its own dynamic linker is a possible lead, brb after some builds test
<dtz> xD
<dtz> or test builds, who can say
<Sonarpulse> dtz: the cached path is from before?
<Sonarpulse> sorry alt-tabbing without ping and I didn't see it
<dtz> np, and yeah, ~5d ago
<dtz> apparently
<dtz> it started failing on aarch64-musl b/c gcc7 broke (??) lol but AFAICT that has been fixed or was transient
<dtz> has some builds in the queue for that job, so fingers crossed :)
<Sonarpulse> dtz: ok! hope it just works out
<Sonarpulse> if not
<Sonarpulse> ugh regressions
jtojnar_ has joined #nixos-dev
<dtz> whoa lol so ... cross-musl has ncurses w/o any heads lmao WHAT
jtojnar has quit [Read error: Connection reset by peer]
<dtz> that'd, uh, explain why ghc can't find it. but... when.. how... why.... !? haha
* dtz puts in debugging hat
<shlevy> Hmm job queued 14 days ago is just building now...
<dtz> erm nvm lol
<shlevy> gchristensen: darwin job hanging at configure :( https://hydra.nixos.org/build/71402813
jtojnar_ has quit [Ping timeout: 256 seconds]
<Dezgeg> someone should probably debug why sdl-config hangs on darwin :P
<Sonarpulse> dtz: heads?
ma27 has quit [Ping timeout: 265 seconds]
<gchristensen> shlevy: hmm it seems my deploy didn't "take"
<gchristensen> (1) fixing the community/ofborg aarch64 box (2) scaling up the evaluators for now (3) the hydra macs is my priority list
<gchristensen> oh (0) fixing the impending inode crisis on the existing evaluators
MichaelRaskin has joined #nixos-dev
pie_ has joined #nixos-dev
orivej has joined #nixos-dev
<shlevy> \o/ glibc-2.27 all green
<shlevy> Moving on to binutils 2.30 then into staging we go
jtojnar has joined #nixos-dev
<Sonarpulse> shlevy: did you ever make trustedBuildRemotes?
<shlevy> Nope
<Sonarpulse> mm ok
* Sonarpulse wonders what will happen if all the office machines have each other as buildRemotes :D
<LnL> Dezgeg: we've worked around it in a few builds, but I've seen a whole bunch of cases when going through zhf
<Sonarpulse> shlevy: is there an issue for it?
<shlevy> Sonarpulse: not sure, sorry
<Sonarpulse> no worries
<Sonarpulse> especially while the triage stuff is unresolved :)
<shlevy> :o not that there was anything private there but I didn't realize those threads were public
<shlevy> Good to know now before I start trash-talking your stuff there :D
<Sonarpulse> :D
orivej has quit [Ping timeout: 265 seconds]
davidlt__ has joined #nixos-dev
davidlt_ has quit [Ping timeout: 265 seconds]