While following the instructions for 10.9 and NixOS/nixpkgs.git everything works fine - until I run `nix-env -u` at the very end (with only cacert and nix 1.8 currently installed).
After successfully compiling a ton of stuff, it errors in this step with a 404 from curl: building path(s) ‘/nix/store/8x6hblrx1vhsmvizy7sd2pga9bs4p5i0-nix-1.9pre4021_f46e329.tar.xz’ trying http://hydra.nixos.org/build/18964230/download/4/nix-1.9pre4021_f46e329.tar.xz % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- 0:00:01 --:--:-- 0curl: (22) The requested URL returned error: 404 Not Found error: cannot download nix-1.9pre4021_f46e329.tar.xz from any mirror builder for ‘/nix/store/8vpjq29iddn6vkm8l3208acfv7lhbvra-nix-1.9pre4021_f46e329.tar.xz.drv’ failed with exit code 1 cannot build derivation ‘/nix/store/g0hvm17z7x4hzlz7lqhjbyrlzsmdwwla-nix-1.9pre4021_f46e329.drv’: 1 dependencies couldn't be built error: build of ‘/nix/store/g0hvm17z7x4hzlz7lqhjbyrlzsmdwwla-nix-1.9pre4021_f46e329.drv’ failed Should I file an issue? If so, with nix or with nixpkgs? Thanks. Am 04.02.2015 um 06:24 schrieb Henry Till <henryt...@gmail.com>: > I've updated the wiki with new instructions: > > https://nixos.org/wiki/Nix_on_OS_X > > -Henry > > > On Tue, Feb 3, 2015 at 12:52 PM, Daniel Peebles <pumpkin...@gmail.com> wrote: > > > > Not much I can think of, but I'm on my phone and can't check now. Perhaps > > some of our patches to individual packages will affect Linux, and some of > > those packages might be used by the Linux stdenv bootstrap. Definitely > > something we should look out for and batch up. > > > > > > > > On Feb 3, 2015, at 12:24, Shea Levy <s...@shealevy.com> wrote: > > > > Sure, tonight works. This sounds great! Besides the patches to cc-wrapper, > > are there any changes you expect to hit linux stdenv? > > > > On Feb 3, 2015, at 5:07 PM, Daniel Peebles <pumpkin...@gmail.com> wrote: > > > > I think we can pull it off this or next week, with a few people > > contributing to testing the semantically meaningful PRs we break into > > chunks. > > > > Fully agree on your proposed workflow, but I've been pretty undisciplined > > so far. For the current big merge, my plan was to break my history into > > major chunks along the following lines (not necessarily in this order): > > - LLVM 3.5 > > - Assorted minor Mac-flavored patches to cc-wrapper (hook to detect > > frameworks in buildInputs, etc.) > > - A large collection of source builds for Apple-specific stuff from > > opensource.apple.com. I've bundled them together in a subdirectory to make > > it a fairly clean independent PR > > - The actual bootstrapped stdenv using clang and my bootstrap tools > > archive. If there are trust issues with people pulling in my binaries, we > > can regenerate them, but it'll be a bit of a pain to do sensibly. We should > > definitely shift the binary bootstrap tools off my Dropbox, either way :) > > - A whole load of assorted patches to individual packages to make them work > > with clang and/or Darwin (should probably be broken up further into smaller > > categories of breakage being fixed, since there are patterns) > > > > I think those are the main categories of PR, and I was planning on > > squashing my messy git history into roughly that shape in the next few days. > > > > Once we've rewritten that history, I'm quite happy to be more disciplined > > in my commits to make them merge more easily. > > > > Want to talk this evening (probably won't be home until 10:30 eastern) on > > ##nix-Darwin about specifics? I need people to help break out individual > > parts (e.g., clang) > > > > Thanks, > > Dan > > > > > > On Feb 3, 2015, at 10:13, Shea Levy <s...@shealevy.com> wrote: > > > > As far as I know, there is still a lot of churn on pure darwin. > > > > Dan, if we wanted to merge pure darwin soon (say, maybe this or next week) > > could you a) get it ready in time and b) adopt a new workflow where future > > major changes were made and merged in single-feature chunks? I want to > > avoid a long-term situation where a lot of unconnected changes come in as a > > set and switch to a feature branch model once pure-darwin is in. > > > > ~Shea > > > > On Feb 3, 2015, at 3:03 PM, Wout Mertens <wout.mert...@gmail.com> wrote: > > > > Ouch. So users on 10.10 should use Zalora and disable hydra binaries? > > > > Perhaps we should fast-track the pure-darwin builds. It works swimmingly on > > my system, and looks like the current Darwin situation is not tenable. So > > anything in pure-darwin that breaks Linux would be good to find early. > > > > On Tue Feb 03 2015 at 4:00:48 PM Shea Levy <s...@shealevy.com> wrote: > >> > >> I’m not sure about butters. Either way a single cache will not do the > >> trick, as until pure-darwin lands binaries will be specific to the version > >> of OS X they are compiled for. In particular, the zalora cache builds for > >> 10.10. > >> > >> On Feb 3, 2015, at 2:59 PM, Wout Mertens <wout.mert...@gmail.com> wrote: > >> > >> I see that the darwin machine butters is online (and idle) in Hydra, does > >> that mean the zalora cache is no longer needed for nixpkgs on Darwin? > >> > >> On Tue Feb 03 2015 at 3:51:47 PM Domen Kožar <do...@dev.si> wrote: > >>> > >>> :beers: > >>> > >>> On Tue, Feb 3, 2015 at 3:40 PM, Shea Levy <s...@shealevy.com> wrote: > >>>> > >>>> Hi all, > >>>> > >>>> For a while now, there has been a lot of work to improve darwin support > >>>> on the master branch of joelteon’s nixpkgs fork. With the latest staging > >>>> merge, the core of that work is now on upstream master. The rest of the > >>>> changes available on joelteon’s fork are very disparate, and it is not > >>>> clear whether or why they are all needed, so I’d like to request that > >>>> those who have been working off of the fork test their setups against > >>>> upstream now. If something is still not working, please at least open an > >>>> issue, but if you can find the relevant fix on the fork and port it as a > >>>> PR that would be greatly appreciated. > >>>> > >>>> Hopefully we can soon move on from that fork entirely and have further > >>>> changes (including the ongoing pure-darwin work) done directly against > >>>> upstream master. > >>>> > >>>> Thanks, > >>>> Shea > >>>> _______________________________________________ > >>>> nix-dev mailing list > >>>> nix-dev@lists.science.uu.nl > >>>> http://lists.science.uu.nl/mailman/listinfo/nix-dev > >>> > >>> > >>> _______________________________________________ > >>> nix-dev mailing list > >>> nix-dev@lists.science.uu.nl > >>> http://lists.science.uu.nl/mailman/listinfo/nix-dev > >> > >> > > > > > > > > _______________________________________________ > > nix-dev mailing list > > nix-dev@lists.science.uu.nl > > http://lists.science.uu.nl/mailman/listinfo/nix-dev > > > _______________________________________________ > nix-dev mailing list > nix-dev@lists.science.uu.nl > http://lists.science.uu.nl/mailman/listinfo/nix-dev _______________________________________________ nix-dev mailing list nix-dev@lists.science.uu.nl http://lists.science.uu.nl/mailman/listinfo/nix-dev
