From e.dolstra at tudelft.nl Wed Feb 3 14:03:17 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Wed, 03 Feb 2010 14:03:17 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19778 - simons - in nixpkgs/trunk/pkgs: tools/system/ipmitool top-level In-Reply-To: <201002031247.o13ClV4X017920@proliant.st.ewi.tudelft.nl> References: <201002031247.o13ClV4X017920@proliant.st.ewi.tudelft.nl> Message-ID: <4B697415.9070905@tudelft.nl> Hi, Peter Simons wrote: > pkgs/tools/system/ipmitool: added optional support for generating statically linked binaries Rather than add a "static" option to every package that one might want to build statically (i.e., all of them), you should have a look at the `makeStaticBinaries' stdenv adapter. Then you can just call a function like ipmitool with a stdenv that builds static binaries: ipmitoolStatic = import ../tools/system/ipmitool { stdenv = makeStaticBinaries stdenv; inherit fetchurl openssl; }; (Maybe makeStaticBinaries should also set AM_LDFLAGS=-all-static, which it doesn't currently do.) -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From e.dolstra at tudelft.nl Wed Feb 3 22:19:35 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Wed, 03 Feb 2010 22:19:35 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19795 - sandervanderburg - nixpkgs/branches/stdenv-updates/pkgs/stdenv/native In-Reply-To: <201002032041.o13KfSci023876@proliant.st.ewi.tudelft.nl> References: <201002032041.o13KfSci023876@proliant.st.ewi.tudelft.nl> Message-ID: <4B69E867.7010702@tudelft.nl> Hi, Sander van der Burg wrote: > Author: sandervanderburg > Date: 2010-02-03 20:41:28 +0000 (Wed, 03 Feb 2010) > New Revision: 19795 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=19795&view=rev > > Added: > nixpkgs/branches/stdenv-updates/pkgs/stdenv/native/derivealiases-freebsd.sh > Modified: > nixpkgs/branches/stdenv-updates/pkgs/stdenv/native/default.nix > > Log: > Added all necessary aliases to match the Linux stdenv as closely as possible on FreeBSD Isn't it better to just add coreutils and so on to the stdenv on FreeBSD? (Basically, use stdenvNix on FreeBSD.) This change breaks in our own build farm: http://hydra.nixos.org/build/280266/nixlog/1/tail -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From S.vanderBurg at tudelft.nl Wed Feb 3 22:37:43 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Wed, 3 Feb 2010 22:37:43 +0100 Subject: [Nix-dev] RE: [Nix-commits] SVN commit: nix - 19795 - sandervanderburg - nixpkgs/branches/stdenv-updates/pkgs/stdenv/native References: <201002032041.o13KfSci023876@proliant.st.ewi.tudelft.nl> <4B69E867.7010702@tudelft.nl> Message-ID: <8A23E531C514ED43B1B469040352E7BA12947A@SRV503.tudelft.net> In order to fix that I have to install additional packages on that machine... hmmm... Maybe, it's indeed better to support stdenvNix. I will look into that now... Btw. I'm now testing on FreeBSD 8.0. I think it's also a good idea to update the FreeBSD machine. -----Oorspronkelijk bericht----- Van: Eelco Dolstra [mailto:e.dolstra at tudelft.nl] Verzonden: wo 3-2-2010 22:19 Aan: nix-dev at cs.uu.nl; Sander van der Burg - EWI Onderwerp: Re: [Nix-commits] SVN commit: nix - 19795 - sandervanderburg - nixpkgs/branches/stdenv-updates/pkgs/stdenv/native Hi, Sander van der Burg wrote: > Author: sandervanderburg > Date: 2010-02-03 20:41:28 +0000 (Wed, 03 Feb 2010) > New Revision: 19795 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=19795&view=rev > > Added: > nixpkgs/branches/stdenv-updates/pkgs/stdenv/native/derivealiases-freebsd.sh > Modified: > nixpkgs/branches/stdenv-updates/pkgs/stdenv/native/default.nix > > Log: > Added all necessary aliases to match the Linux stdenv as closely as possible on FreeBSD Isn't it better to just add coreutils and so on to the stdenv on FreeBSD? (Basically, use stdenvNix on FreeBSD.) This change breaks in our own build farm: http://hydra.nixos.org/build/280266/nixlog/1/tail -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100203/817a83ca/attachment.html From ludo at gnu.org Wed Feb 3 23:35:10 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Wed, 03 Feb 2010 23:35:10 +0100 Subject: [Nix-dev] Dependency on libiconv In-Reply-To: <201002032020.o13KK1Zr023786@proliant.st.ewi.tudelft.nl> (Sander van der Burg's message of "Wed, 03 Feb 2010 20:20:01 +0000") References: <201002032020.o13KK1Zr023786@proliant.st.ewi.tudelft.nl> Message-ID: <878wbaeyn5.fsf@gnu.org> Hi, Sander van der Burg writes: > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=19793&view=rev > > Modified: > nixpkgs/branches/stdenv-updates/pkgs/development/libraries/dbus-glib/default.nix > nixpkgs/branches/stdenv-updates/pkgs/top-level/all-packages.nix > > Log: > Fixed dbus-glib compilation on FreeBSD [...] > --- nixpkgs/branches/stdenv-updates/pkgs/development/libraries/dbus-glib/default.nix 2010-02-03 20:12:18 UTC (rev 19792) > +++ nixpkgs/branches/stdenv-updates/pkgs/development/libraries/dbus-glib/default.nix 2010-02-03 20:20:00 UTC (rev 19793) > @@ -8,7 +8,7 @@ > sha256 = "0nv4gxcbpa9f0907dmzmfm222w8y45z19cx27l85f5qknf8hncxm"; > }; > > - buildInputs = [pkgconfig expat gettext]; > + buildInputs = [pkgconfig expat gettext libiconv]; [...] > dbus_glib = makeOverridable (import ../development/libraries/dbus-glib) { > inherit fetchurl stdenv pkgconfig gettext dbus expat glib; > + libiconv = if (stdenv.system == "i686-freebsd") then libiconv else null; > }; Libiconv is for all non-GNU systems, and only them. So the Right Thing would be to have: buildInputs = ... ++ stdenv.lib.optional (! (stdenv ? glibc)) libiconv; and leave ?all-packages.nix? part free of any conditional (does having a build input equal to ?null? always work?). Thanks, Ludo?. From lkraider at gmail.com Wed Feb 3 23:58:34 2010 From: lkraider at gmail.com (Paul Eipper) Date: Wed, 3 Feb 2010 23:58:34 +0100 Subject: [Nix-dev] Mac Installer Message-ID: <2ee02671002031458k476e430au3af4b3f42e81fa8e@mail.gmail.com> Hi, I've built a binary installer package for nix 0.13 for mac 10.5 (leopard), if anyone is interested. I have not tested it yet on other systems, so if anyone wants to try it please give some feedback. link: http://lkraider.eipper.com.br/files/nix/nix_0.13_installer.pkg size: 42mb It also includes a checkout of the unstable repo (wich it could probably do without, tho). att, -- Paul Eipper From e.dolstra at tudelft.nl Fri Feb 5 11:17:32 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Fri, 05 Feb 2010 11:17:32 +0100 Subject: [Nix-dev] Broken links in site/manual In-Reply-To: <8b2a1a961001310832v4ecfb822h9cf1a3ff6a6388@mail.gmail.com> References: <8b2a1a961001310832v4ecfb822h9cf1a3ff6a6388@mail.gmail.com> Message-ID: <4B6BF03C.9040001@tudelft.nl> Hi, Bjorn Buckwalter wrote: > FYI I found these two broken links on the site (page with link --> > destination that doesn't exist): > > http://hydra.nixos.org/build/118589/download/1/manual/ --> > http://nixos.org/releases/nixpkgs/nixpkgs-unstable/ > http://nixos.org/nix/ --> http://nixos.org/nix/nixos/ Thanks, fixed. -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From e.dolstra at tudelft.nl Fri Feb 5 11:19:18 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Fri, 05 Feb 2010 11:19:18 +0100 Subject: [Nix-dev] Nix 0.14 released Message-ID: <4B6BF0A6.9000900@tudelft.nl> Hi, I'm pleased to announce the availability of a new stable release of the Nix package manager. Release 0.14 can be found at http://hydra.nixos.org/release/nix/nix-0.14 *** Release notes *** The release notes can also be found at http://hydra.nixos.org/build/281118/download/3/release-notes/. This release has the following improvements: * The garbage collector now starts deleting garbage much faster than before. It no longer determines liveness of all paths in the store, but does so on demand. * Added a new operation, nix-store --query --roots, that shows the garbage collector roots that directly or indirectly point to the given store paths. * Removed support for converting Berkeley DB-based Nix databases to the new schema. * Removed the --use-atime and --max-atime garbage collector options. They were not very useful in practice. * On Windows, Nix now requires Cygwin 1.7.x. * A few bug fixes. -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From e.dolstra at tudelft.nl Fri Feb 5 11:31:07 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Fri, 05 Feb 2010 11:31:07 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19823 - simons - in nixpkgs/trunk/pkgs: development/libraries/sqlite top-level In-Reply-To: <201002041607.o14G7F7o008093@proliant.st.ewi.tudelft.nl> References: <201002041607.o14G7F7o008093@proliant.st.ewi.tudelft.nl> Message-ID: <4B6BF36B.5070905@tudelft.nl> Hi, Peter Simons wrote: > Author: simons > Date: 2010-02-04 16:07:15 +0000 (Thu, 04 Feb 2010) > New Revision: 19823 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=19823&view=rev > > Modified: > nixpkgs/trunk/pkgs/development/libraries/sqlite/default.nix > nixpkgs/trunk/pkgs/top-level/all-packages.nix > > Log: > Added an interactive variant of sqlite3 that has readline support enabled. It may be worth noting that you can get the same effect by using the "rlwrap" program (just run "rlwrap sqlite3"). -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From jason.dusek at gmail.com Fri Feb 5 18:38:31 2010 From: jason.dusek at gmail.com (Jason Dusek) Date: Fri, 5 Feb 2010 09:38:31 -0800 Subject: [Nix-dev] Managing users, configuration files and filesystem mounts. Message-ID: <42784f261002050938s31bf485tf9dd2586960f0837@mail.gmail.com> I am curious as to how Nix manages users and groups outside of NixOS. In perusing the NGinX derivation, for example, it's not clear to me that one ever creates a user `nginx'. Is this just not usually done with Nix? It seems reasonable that this is not possible, due to Nix's stated goals of purity. An alternative way I thought of is to create a `local-shadow' derivation that creates `/etc/passwd' and so on. However, these files would never leave the Nix store! So I am stuck again. While NixOS is appealing, I can't move my systems to it right this minute -- I'm looking at managing CentOS boxen. Maybe I'll have to use Bcfg2 or Puppet for the time being to manage users and other system global configuration. -- Jason Dusek From ludo at gnu.org Fri Feb 5 19:32:55 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Fri, 05 Feb 2010 19:32:55 +0100 Subject: [Nix-dev] Re: Managing users, configuration files and filesystem mounts. References: <42784f261002050938s31bf485tf9dd2586960f0837@mail.gmail.com> Message-ID: <87iqabcz3c.fsf@gnu.org> Hello, Jason Dusek writes: > I am curious as to how Nix manages users and groups outside of > NixOS. Normally, all users and groups are declared in NixOS. Users (actually sysadmins) can specify a list in the ?users.extraUsers? attribute, in their ?configuration.nix?. Likewise, services provided by NixOS can declare users/groups; for instance, the Avahi service declares that it wants an ?avahi? user and group, so that ?avahi-daemon? can run under that user. Does that answer your question? Thanks, Ludo?. From jason.dusek at gmail.com Fri Feb 5 19:42:48 2010 From: jason.dusek at gmail.com (Jason Dusek) Date: Fri, 5 Feb 2010 10:42:48 -0800 Subject: [Nix-dev] Re: Managing users, configuration files and filesystem mounts. In-Reply-To: <87iqabcz3c.fsf@gnu.org> References: <42784f261002050938s31bf485tf9dd2586960f0837@mail.gmail.com> <87iqabcz3c.fsf@gnu.org> Message-ID: <42784f261002051042q677a2abbh927c3e07b6724473@mail.gmail.com> 2010/02/05 Ludovic Court?s : > Jason Dusek writes: > > I am curious as to how Nix manages users and groups outside > > of NixOS. > > Normally, all users and groups are declared in NixOS. > > Users (actually sysadmins) can specify a list in the > ?users.extraUsers? attribute, in their ?configuration.nix?. > Likewise, services provided by NixOS can declare > users/groups; for instance, the Avahi service declares that it > wants an ?avahi? user and group, so that ?avahi-daemon? can > run under that user. Where can I browse the NixOS specific packages? I poked around under `http://nixos.org/releases/' but have not found them. > Does that answer your question? It sounds like users are managed in NixOS but when using Nix overlayed with Ubuntu or something else, you must manage users outside of Nix. Is that correct? -- Jason Dusek From viriketo at gmail.com Fri Feb 5 20:52:50 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Fri, 5 Feb 2010 20:52:50 +0100 Subject: [Nix-dev] Re: Managing users, configuration files and filesystem mounts. In-Reply-To: <42784f261002051042q677a2abbh927c3e07b6724473@mail.gmail.com> References: <42784f261002050938s31bf485tf9dd2586960f0837@mail.gmail.com> <87iqabcz3c.fsf@gnu.org> <42784f261002051042q677a2abbh927c3e07b6724473@mail.gmail.com> Message-ID: <45219fb01002051152x74fbbaaej81dc586546c24d1c@mail.gmail.com> 2010/2/5 Jason Dusek : > 2010/02/05 Ludovic Court?s : >> Jason Dusek writes: >> > I am curious as to how Nix manages users and groups outside >> > of NixOS. > > ?Where can I browse the NixOS specific packages? I poked around > ?under `http://nixos.org/releases/' but have not found them. As you know, the useful result of nix expressions are 'derivations', and in the case of nixpkgs, the derivations are thought coinciding with what other software installation systems call 'packages'. In the case of nixos, I find it harder to name the nix derivations there "packages". You will find the derivation of the avahi start script, the derivation involved in producing the /etc/passwd file, the derivation producing the /etc/resolv.conf, the derivation producing the initrd image, ... And, important, the derivation of the "activation script". That script puts whatever needed outside of the store, so the system works. That makes a strong usage of the other derivations of nixos, of course. For example this is what ends up *manipulating* the /etc/passwd file so it has all the users the nixos modules requested (keep on reading, about nixos modules). How the nix expressions in nixos map to file derivations, you will find that in the nixos manual, namely through a nix abstraction we call "nixos modules". A nixos module will define some attribute sets, receiving some others, and will be evaluated so from it can result multiple derivations. Those derivations may come not only from a single module, but from more than one. So, the "avahi" and the "vsftpd" modules both contribute to the derivation producing the /etc/passwd file (both define their users). And the result of the "avahi" nixos module is not a single derivation, but a contribution to many: the users, the init script, maybe some files at /etc related to avahi, adding a kernel module, ... > ?It sounds like users are managed in NixOS but when using Nix > ?overlayed with Ubuntu or something else, you must manage users > ?outside of Nix. Is that correct? Yes From marco-oweber at gmx.de Fri Feb 5 23:18:52 2010 From: marco-oweber at gmx.de (Marc Weber) Date: Fri, 05 Feb 2010 23:18:52 +0100 Subject: [Nix-dev] Re: Managing users, configuration files and filesystem mounts. In-Reply-To: <42784f261002051042q677a2abbh927c3e07b6724473@mail.gmail.com> References: <42784f261002050938s31bf485tf9dd2586960f0837@mail.gmail.com> <87iqabcz3c.fsf@gnu.org> <42784f261002051042q677a2abbh927c3e07b6724473@mail.gmail.com> Message-ID: <1265408212-sup-2410@nixos> Excerpts from Jason Dusek's message of Fri Feb 05 19:42:48 +0100 2010: > 2010/02/05 Ludovic Court?s : > > Jason Dusek writes: > > > I am curious as to how Nix manages users and groups outside > > > of NixOS. > > > > Normally, all users and groups are declared in NixOS. > > > > Users (actually sysadmins) can specify a list in the > > ?users.extraUsers? attribute, in their ?configuration.nix?. > > Likewise, services provided by NixOS can declare > > users/groups; for instance, the Avahi service declares that it > > wants an ?avahi? user and group, so that ?avahi-daemon? can > > run under that user. > > Where can I browse the NixOS specific packages? I poked around > under `http://nixos.org/releases/' but have not found them. Get the nixpkgs tar.gz or SVN repository and use grep or gnu id utils. This is best you can do. This way you won't miss anything. (Or join ask here or on irc). If you already have nix installed you can use nix-env to grep derivations by name. We should add this feature to the homepage ! > > Does that answer your question? > > It sounds like users are managed in NixOS but when using Nix > overlayed with Ubuntu or something else, you must manage users > outside of Nix. Is that correct? It depends. Of course MySQL, Apache users etc are created for you. However doing a useradd -m name .. does work equally well. It's not worth for some accounts adding them to NixOS because they may be of temporary nature. Marc Weber From ludo at gnu.org Sat Feb 6 01:16:35 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Sat, 06 Feb 2010 01:16:35 +0100 Subject: [Nix-dev] Tcpdump mirrors In-Reply-To: <201002052201.o15M1PJL032300@proliant.st.ewi.tudelft.nl> (Eelco Dolstra's message of "Fri, 05 Feb 2010 22:01:25 +0000") References: <201002052201.o15M1PJL032300@proliant.st.ewi.tudelft.nl> Message-ID: <87k4ur9q1o.fsf@gnu.org> Hello! Eelco Dolstra writes: > - # Tcpdump (see http://www.tcpdump.org/mirrors.html). > - tcpdump = [ > - ftp://ftp.wiretapped.net/pub/security/packet-capture/tcpdump.org/ > - ftp://mirror.aarnet.edu.au/pub/tcpdump/ > - ftp://gd.tuwien.ac.at/infosys/security/tcpdump.org/ > - ftp://ftp.gwdg.de/pub/misc/tcpdump/ > - ]; Why remove them? They appear to be still valid. Thanks, Ludo?. From juhpetersen at gmail.com Sat Feb 6 10:16:14 2010 From: juhpetersen at gmail.com (Jens Petersen) Date: Sat, 6 Feb 2010 19:16:14 +1000 Subject: [Nix-dev] [patch] small fixes for nixpkgs/doc Message-ID: <9436bffe1002060116x2ff203cfk479db9526e9a796d@mail.gmail.com> Hi, Here is a small patch to fix a few typos in the nixpkgs manual. Thanks, Jens -------------- next part -------------- A non-text attachment was scrubbed... Name: nixpkgs-doc-small-fixes.patch Type: text/x-patch Size: 1533 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100206/13d80ae5/attachment.bin From e.dolstra at tudelft.nl Sat Feb 6 14:14:19 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Sat, 06 Feb 2010 14:14:19 +0100 Subject: [Nix-dev] Stdenv branch merged Message-ID: <4B6D6B2B.4090903@tudelft.nl> Hi Nixers, Last night I merged the stdenv branch into the trunk. This gives us GCC 4.4, Glibc 2.11, support for cross-compilation, and a bunch of other updates to stdenv. Thanks to Llu?s and Ludovic for doing most of the hard work. Note that quite a few packages are still broken: http://hydra.nixos.org/jobset/nixpkgs/trunk/jobstatus Also, Mac OS X is completely broken. This should be fixed ASAP. -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From juhpetersen at gmail.com Sat Feb 6 16:07:12 2010 From: juhpetersen at gmail.com (Jens Petersen) Date: Sun, 7 Feb 2010 01:07:12 +1000 Subject: [Nix-dev] info wrapper patch Message-ID: <9436bffe1002060707j3bb78c4bj1ab49709ce08217e@mail.gmail.com> Tweak the info wrapper to not assume all info files end in ".info" (eg emacs). -Jens ps Still not sure how/if to generate share/info/dir in the default profile. -------------- next part -------------- A non-text attachment was scrubbed... Name: nixos-info-wrapper-files.patch Type: text/x-patch Size: 447 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100207/678fd6f2/attachment.bin From ludo at gnu.org Sat Feb 6 16:27:19 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Sat, 06 Feb 2010 16:27:19 +0100 Subject: [Nix-dev] Re: Stdenv branch merged References: <4B6D6B2B.4090903@tudelft.nl> Message-ID: <87636a9yg8.fsf@gnu.org> Hello! Eelco Dolstra writes: > Note that quite a few packages are still broken: > http://hydra.nixos.org/jobset/nixpkgs/trunk/jobstatus TeXLive is among them. Any takers? :-) Thanks, Ludo?. From nicolas.b.pierron at gmail.com Mon Feb 8 10:47:37 2010 From: nicolas.b.pierron at gmail.com (Nicolas Pierron) Date: Mon, 8 Feb 2010 10:47:37 +0100 Subject: [Nix-dev] Stdenv branch merged In-Reply-To: <4B6D6B2B.4090903@tudelft.nl> References: <4B6D6B2B.4090903@tudelft.nl> Message-ID: <787b29b91002080147g2cbb679dt7098da7c5038b88@mail.gmail.com> On Sat, Feb 6, 2010 at 14:14, Eelco Dolstra wrote: > Last night I merged the stdenv branch into the trunk. ?This gives us GCC 4.4, > Glibc 2.11, support for cross-compilation, and a bunch of other updates to > stdenv. ?Thanks to Llu?s and Ludovic for doing most of the hard work. I had problems with gcc 4.4 and C++ anonymous namespaces, like generated by recent bison versions. -- Nicolas Pierron http://www.linkedin.com/in/nicolasbpierron - http://nbp.name/ If you are doing something twice then you should try to do it once. From ludo at gnu.org Mon Feb 8 11:42:08 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Mon, 08 Feb 2010 11:42:08 +0100 Subject: [Nix-dev] Re: Stdenv branch merged References: <4B6D6B2B.4090903@tudelft.nl> <787b29b91002080147g2cbb679dt7098da7c5038b88@mail.gmail.com> Message-ID: <87ljf46mbj.fsf@gnu.org> Salut Nicolas, ;-) Nicolas Pierron writes: > I had problems with gcc 4.4 and C++ anonymous namespaces, like > generated by recent bison versions. Can you elaborate on the problems and solutions (if any), for future reference? Thanks, Ludo?. From viriketo at gmail.com Mon Feb 8 14:43:20 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Mon, 8 Feb 2010 14:43:20 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19862 - eelco - nixpkgs/trunk/pkgs/lib In-Reply-To: <201002081335.o18DZcBb017853@proliant.st.ewi.tudelft.nl> References: <201002081335.o18DZcBb017853@proliant.st.ewi.tudelft.nl> Message-ID: <45219fb01002080543i42c6f4f6m780bd7d09ff0159a@mail.gmail.com> 2010/2/8 Eelco Dolstra : > Author: eelco > Date: 2010-02-08 13:35:38 +0000 (Mon, 08 Feb 2010) > New Revision: 19862 > > You can view the changes in this commit at: > ? https://svn.nixos.org/viewvc/nix?rev=19862&view=rev > > Modified: > ? nixpkgs/trunk/pkgs/lib/customisation.nix > > Log: > * Fix the VM builds. ?They were broken because a `hostDrv' and > ?`buildDrv' attribute gets added to every derivation in > ?`makeStdenvCross'. ?So filter out those attributes. ?But the real > ?question is why makeStdenvCross is applied when I'm not doing a > ?cross compilation. Good question! Worth an investigation! I will take a look in the evening. From viriketo at gmail.com Tue Feb 9 00:22:47 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Tue, 9 Feb 2010 00:22:47 +0100 Subject: [Nix-dev] hydra-nixos-trunk does not have isos Message-ID: <45219fb01002081522s796f99e8vc3fb3f7f2b4b23b0@mail.gmail.com> Why the current job set does not have isos? I could instantiate them at my copy of nixos/nixpkgs trunk. Regards, Llu?s. From rob.vermaas at gmail.com Tue Feb 9 07:46:30 2010 From: rob.vermaas at gmail.com (Rob Vermaas) Date: Tue, 9 Feb 2010 07:46:30 +0100 Subject: [Nix-dev] hydra-nixos-trunk does not have isos In-Reply-To: <45219fb01002081522s796f99e8vc3fb3f7f2b4b23b0@mail.gmail.com> References: <45219fb01002081522s796f99e8vc3fb3f7f2b4b23b0@mail.gmail.com> Message-ID: <5a9e75b11002082246v17440aabgd516e47d252ebfbf@mail.gmail.com> Because there are some evaluation errors: at `tests.installer.simple' [nixpkgs = ..., services = ...]: attribute `buildDrv' missing at `tests.installer.lvm' [nixpkgs = ..., services = ...]: attribute `buildDrv' missing at `tests.installer.separateBoot' [nixpkgs = ..., services = ...]: attribute `buildDrv' missing at `iso_graphical' [officialRelease = false, nixpkgs = ..., system = "i686-linux", nixosSrc = ...]: attribute `buildDrv' missing at `iso_graphical' [officialRelease = false, nixpkgs = ..., system = "x86_64-linux", nixosSrc = ...]: attribute `buildDrv' missing at `makeIso' [nixpkgs = ...]: job `makeIso' requires an argument named `module' at `iso_minimal' [officialRelease = false, nixpkgs = ..., system = "i686-linux", nixosSrc = ...]: attribute `buildDrv' missing at `iso_minimal' [officialRelease = false, nixpkgs = ..., system = "x86_64-linux", nixosSrc = ...]: attribute `buildDrv' missing at `manual' [officialRelease = false, nixpkgs = ..., nixosSrc = ...]: attribute `buildDrv' missing On 9 February 2010 00:22, Llu?s Batlle wrote: > Why the current job set does not have isos? I could instantiate them > at my copy of nixos/nixpkgs trunk. > > Regards, > Llu?s. > _______________________________________________ > nix-dev mailing list > nix-dev at cs.uu.nl > https://mail.cs.uu.nl/mailman/listinfo/nix-dev > -- Rob Vermaas [email] rob.vermaas at gmail.com From viriketo at gmail.com Tue Feb 9 08:35:26 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Tue, 9 Feb 2010 08:35:26 +0100 Subject: [Nix-dev] hydra-nixos-trunk does not have isos In-Reply-To: <5a9e75b11002082246v17440aabgd516e47d252ebfbf@mail.gmail.com> References: <45219fb01002081522s796f99e8vc3fb3f7f2b4b23b0@mail.gmail.com> <5a9e75b11002082246v17440aabgd516e47d252ebfbf@mail.gmail.com> Message-ID: <45219fb01002082335o1736be52vdf16089fd71892dc@mail.gmail.com> Ahm! Let me look into this. :) 2010/2/9 Rob Vermaas : > Because there are some evaluation errors: > > at `tests.installer.simple' [nixpkgs = ..., services = ...]: > attribute `buildDrv' missing > > at `tests.installer.lvm' [nixpkgs = ..., services = ...]: > attribute `buildDrv' missing > > at `tests.installer.separateBoot' [nixpkgs = ..., services = ...]: > attribute `buildDrv' missing > > at `iso_graphical' [officialRelease = false, nixpkgs = ..., system = > "i686-linux", nixosSrc = ...]: > attribute `buildDrv' missing > > at `iso_graphical' [officialRelease = false, nixpkgs = ..., system = > "x86_64-linux", nixosSrc = ...]: > attribute `buildDrv' missing > > at `makeIso' [nixpkgs = ...]: > job `makeIso' requires an argument named `module' > > at `iso_minimal' [officialRelease = false, nixpkgs = ..., system = > "i686-linux", nixosSrc = ...]: > attribute `buildDrv' missing > > at `iso_minimal' [officialRelease = false, nixpkgs = ..., system = > "x86_64-linux", nixosSrc = ...]: > attribute `buildDrv' missing > > at `manual' [officialRelease = false, nixpkgs = ..., nixosSrc = ...]: > attribute `buildDrv' missing > > > > On 9 February 2010 00:22, Llu?s Batlle wrote: >> Why the current job set does not have isos? I could instantiate them >> at my copy of nixos/nixpkgs trunk. >> >> Regards, >> Llu?s. >> _______________________________________________ >> nix-dev mailing list >> nix-dev at cs.uu.nl >> https://mail.cs.uu.nl/mailman/listinfo/nix-dev >> > > > > -- > Rob Vermaas > > [email] rob.vermaas at gmail.com > From nicolas.b.pierron at gmail.com Wed Feb 10 20:04:42 2010 From: nicolas.b.pierron at gmail.com (Nicolas Pierron) Date: Wed, 10 Feb 2010 20:04:42 +0100 Subject: [Nix-dev] Re: [FOSDEM-Dist2010] Post mortem: feedback please! In-Reply-To: <20100210131415.GB3441@celtic.nixsys.be> References: <20100210131415.GB3441@celtic.nixsys.be> Message-ID: <787b29b91002101104k11c65f34r92963fc2c0b3eede@mail.gmail.com> Hi, On Wed, Feb 10, 2010 at 14:14, Wouter Verhelst wrote: > Since this is the first time we had a multi-distro devroom at FOSDEM, I > To do that properly, we will need feedback from as many people as > possible: distro coordinators, visitors, speakers, everyone. Please be > so kind as to let us know what you thought of the distro devroom. Even > if you send us a mail which consists of nothing more than something like > 'I love it, please do so again' or 'it was crap, I hate it, please do > the distro devrooms again next year', it would be valuable to know that. As a speaker, I enjoyed doing the talk. With these multi-distro devrooms, I've got a chance to share my work on NixOS and I'll be gald to see some derivative integrated in other projects. As a visitor, I have got some new ideas from Config::Model to add inside our current system. I love what you did by grouping similar subjects together. This give us to opportunity to meet people working on the same topics. I think all projects can benefit from other designs and at least compare and share experiences. I think that you should highlight groups in the schedule as a suggestion (if you are interested by this one, then you may have a look at the other ones in the same group). > Also, it would be nice if people could forward this request in their > respective communities. Doing so ... Cheers, -- Nicolas Pierron http://www.linkedin.com/in/nicolasbpierron - http://nbp.name/ If you are doing something twice then you should try to do it once. From viriketo at gmail.com Sat Feb 13 01:04:19 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sat, 13 Feb 2010 01:04:19 +0100 Subject: [Nix-dev] kde 4.4 - still trying to get it working Message-ID: <45219fb01002121604n77a5c1b0s438fe333e30caef8@mail.gmail.com> Hello, I made some effort today trying to get kde 4.4 built. kde-workspace still needs some work (specially in myPatchPhase, which Sander will know better). On the other hand, I think I found a blocker: the runtime requires phonon-4.4.0, and it still has not been released. http://old.nabble.com/Which-Qt-version-required-for-KDE4.4--td27415116.html I could not find it in the mirrors. Let's wait for it. Regards, Llu?s. From viriketo at gmail.com Sat Feb 13 12:20:46 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sat, 13 Feb 2010 12:20:46 +0100 Subject: [Nix-dev] hydra and the plus sign Message-ID: <45219fb01002130320y2cd62fa0o84ea718f07a4f94a@mail.gmail.com> Hello, I'm trying to get the closure for gtk+ from hydra, and I get: *** Step 1/1: downloading `http://hydra.nixos.org/jobset/nixpkgs/trunk/channel/latest/nar/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk%2b-2.18.6' into `/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk+-2.18.6' downloading archive... % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 4620k 0 4620k 0 0 253k 0 --:--:-- 0:00:18 --:--:-- 288k error: invalid character `%' in name `yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk%2b-2.18.6' download of `http://hydra.nixos.org/jobset/nixpkgs/trunk/channel/latest/nar/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk%2b-2.18.6' failed at /nix/store/rglv9ibyj0ajkm5gc7h0sqmgsh6kq3za-nix-0.14/libexec/nix/substituters/download-using-manifests.pl line 267. substitution of path `/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk+-2.18.6' using substituter `/nix/store/rglv9ibyj0ajkm5gc7h0sqmgsh6kq3za-nix-0.14/libexec/nix/substituters/download-using-manifests.pl' failed: builder for `/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk+-2.18.6' failed with exit code 1 path `/nix/store/yi5gida0yg8fhcmchy9wd4ihnri7i9nj-gtk+-2.18.6' is required, but there is no substituter that can build it There is some translation to URL escape codes where it should not be, I imagine. I have to go without manifests, I guess, until someone fixes this? Thank you, Llu?s. From 7c6f434c at mail.ru Sat Feb 13 18:07:29 2010 From: 7c6f434c at mail.ru (Michael Raskin) Date: Sat, 13 Feb 2010 20:07:29 +0300 Subject: [Nix-dev] Gnome semi-auto-updater script Message-ID: <4B76DC51.60307@mail.ru> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Medium-dirty hack, requires slight supervision. Used for my massive glib-gtk-gnome update. Llu?s asked me to send it to the list... -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJLdtxOAAoJEE6tnN0aWvw3+DUH/iB0hPbpnY/M2uY1YpFPL8lF O8gZA+VxTUQuqVDTgHaUWz6TOcmYe7+fdT0kDiBvgnJv/j2sUG0CnGsdtiJZ0CSJ FN57mM/HVqWfoa9N1sKtEVGUA10tsAN1qQPOVveaUA8MVnzk2puyJKawl38ppVS6 zg4jEuqap2SMulxdfem2YJ4GX2ExZ/YJMWPJjCwERExHeekO3M9GdtSwYTnH+swM /ZuxJqmc7GBaLCZceEyLBcDE+b4cyTdGkNbmlj27CtCZ+qsxUveZt1X9bI7LCNNk zb9wUsARke7eWAV0uQmEuzRZkxStIRheN/sunOvdGkP/6cGPnM7iARvBPOTZfYg= =/OJf -----END PGP SIGNATURE----- -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: gnome-package-update Url: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100213/aab73955/attachment.pl -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: gnome-package-update-separate Url: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100213/aab73955/attachment-0001.pl -------------- next part -------------- A non-text attachment was scrubbed... Name: gnome-package-update.sig Type: application/octet-stream Size: 287 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100213/aab73955/attachment.obj -------------- next part -------------- A non-text attachment was scrubbed... Name: gnome-package-update-separate.sig Type: application/octet-stream Size: 287 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100213/aab73955/attachment-0001.obj From e.dolstra at tudelft.nl Sun Feb 14 17:20:24 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Sun, 14 Feb 2010 17:20:24 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric - nix/branches In-Reply-To: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> Message-ID: <4B7822C8.5080107@tudelft.nl> Llus Batlle wrote: > Author: viric > Date: 2010-02-13 16:27:50 +0000 (Sat, 13 Feb 2010) > New Revision: 19980 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=19980&view=rev > > Added: > nix/branches/noaterm/ > > Log: > Creating a branch to try to get a nix without aterm. Can you give some background on this? What is the reason for getting rid of it? -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From viriketo at gmail.com Sun Feb 14 17:28:56 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sun, 14 Feb 2010 17:28:56 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric - nix/branches In-Reply-To: <4B7822C8.5080107@tudelft.nl> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> Message-ID: <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> Hello Eelco, I had the feeling that it becomes hard for us to follow the aterm development. You mentioned about they not accepting your fixes, we have newer aterm versions provoking segfaults around, either in i686 and x86_64, ... And so we are meanwhile stuck to an old aterm library. As the aterm functionality is not that big, and most of its benefits rely on its efficiency (correct me if I understood wrong), I thought that we could have a "slower-but-reliable" implementation of that. Pierron also got cought by the idea of getting rid of aterm, and being already two, we decided to start on something. At least, making a branch. Pierron has been thinking on all this the weekend, and he even thinks that a new implementation can beat the performance of aterm (I assume, "easily"). I don't know his ideas concrete, maybe whenever he can commit something, we can keep on talking on this. I'm pushing him to get first something api-compatible with aterm, and then we will see what to do next. Do you think we go the wrong way? Thank you, Llu?s. 2010/2/14 Eelco Dolstra : > Llus Batlle wrote: > >> Author: viric >> Date: 2010-02-13 16:27:50 +0000 (Sat, 13 Feb 2010) >> New Revision: 19980 >> >> You can view the changes in this commit at: >> ? ?https://svn.nixos.org/viewvc/nix?rev=19980&view=rev >> >> Added: >> ? ?nix/branches/noaterm/ >> >> Log: >> Creating a branch to try to get a nix without aterm. > > Can you give some background on this? ?What is the reason for getting rid of it? > > -- > Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ > From S.vanderBurg at tudelft.nl Sun Feb 14 23:00:58 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Sun, 14 Feb 2010 23:00:58 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric -nix/branches References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl><4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> Message-ID: <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> So, what are you guys going to do? Implementing a clean room implementation of ATerm with the same interfaces or forking ATerm? And if so, why should this ATerm version developed inside the Nix repository? -----Oorspronkelijk bericht----- Van: nix-dev-bounces at cs.uu.nl namens Llu?s Batlle Verzonden: zo 14-2-2010 17:28 Aan: Eelco Dolstra - EWI CC: nix-dev at cs.uu.nl Onderwerp: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric -nix/branches Hello Eelco, I had the feeling that it becomes hard for us to follow the aterm development. You mentioned about they not accepting your fixes, we have newer aterm versions provoking segfaults around, either in i686 and x86_64, ... And so we are meanwhile stuck to an old aterm library. As the aterm functionality is not that big, and most of its benefits rely on its efficiency (correct me if I understood wrong), I thought that we could have a "slower-but-reliable" implementation of that. Pierron also got cought by the idea of getting rid of aterm, and being already two, we decided to start on something. At least, making a branch. Pierron has been thinking on all this the weekend, and he even thinks that a new implementation can beat the performance of aterm (I assume, "easily"). I don't know his ideas concrete, maybe whenever he can commit something, we can keep on talking on this. I'm pushing him to get first something api-compatible with aterm, and then we will see what to do next. Do you think we go the wrong way? Thank you, Llu?s. 2010/2/14 Eelco Dolstra : > Llus Batlle wrote: > >> Author: viric >> Date: 2010-02-13 16:27:50 +0000 (Sat, 13 Feb 2010) >> New Revision: 19980 >> >> You can view the changes in this commit at: >> ? ?https://svn.nixos.org/viewvc/nix?rev=19980&view=rev >> >> Added: >> ? ?nix/branches/noaterm/ >> >> Log: >> Creating a branch to try to get a nix without aterm. > > Can you give some background on this? ?What is the reason for getting rid of it? > > -- > Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ > _______________________________________________ nix-dev mailing list nix-dev at cs.uu.nl https://mail.cs.uu.nl/mailman/listinfo/nix-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100214/b7a5eebe/attachment.html From nicolas.b.pierron at gmail.com Sun Feb 14 23:58:51 2010 From: nicolas.b.pierron at gmail.com (Nicolas Pierron) Date: Sun, 14 Feb 2010 23:58:51 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric -nix/branches In-Reply-To: <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> Message-ID: <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> 2010/2/14 Sander van der Burg - EWI : > So, what are you guys going to do? Implementing a clean room implementation > of ATerm with the same interfaces or forking ATerm? And if so, why should > this ATerm version developed inside the Nix repository? Nix does not need ATerm, Nix need maximal sharing to implements maximal laziness. The garbage collection is almost useless because all terms are kept alive in the ATermMap which is used to cache reduce operations. So the garbage collection is just a way to lose some evaluation time. Some manipulations are done to convert data to structure which are easy to manipulate and to structure which are easy to store. These conversions are costly, why don't we just use one data format? What I want to do is to keep the minimal amount of features required by Nix to optimize the code in consequence. For sure optimizing it will break the ATerm API. And as I've experienced a lot of difficulties to debug the ATerm library (due to all the crap that you don't want to understand). We don't have access to their repository. We don't see them interacting with the code. I never got answer from their issue tracker. We can't follow the last version. We need a patch to make it works. They depends on the assembly language. The library is not thread safe. So I consider it as a dead dependency which is not well suited for our usage. If I succeed to make it as I think of it, I will probably take it out of Nix to use it in my work place too. > -----Oorspronkelijk bericht----- > Van: nix-dev-bounces at cs.uu.nl namens Llu?s Batlle > Verzonden: zo 14-2-2010 17:28 > Aan: Eelco Dolstra - EWI > CC: nix-dev at cs.uu.nl > Onderwerp: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric > -nix/branches > > Hello Eelco, > > I had the feeling that it becomes hard for us to follow the aterm > development. You mentioned about they not accepting your fixes, we > have newer aterm versions provoking segfaults around, either in i686 > and x86_64, ... And so we are meanwhile stuck to an old aterm library. > > As the aterm functionality is not that big, and most of its benefits > rely on its efficiency (correct me if I understood wrong), I thought > that we could have a "slower-but-reliable" implementation of that. > Pierron also got cought by the idea of getting rid of aterm, and being > already two, we decided to start on something. At least, making a > branch. > > Pierron has been thinking on all this the weekend, and he even thinks > that a new implementation can beat the performance of aterm (I assume, > "easily"). I don't know his ideas concrete, maybe whenever he can > commit something, we can keep on talking on this. > > I'm pushing him to get first something api-compatible with aterm, and > then we will see what to do next. > > Do you think we go the wrong way? > > Thank you, > Llu?s. > > 2010/2/14 Eelco Dolstra : >> Llus Batlle wrote: >> >>> Author: viric >>> Date: 2010-02-13 16:27:50 +0000 (Sat, 13 Feb 2010) >>> New Revision: 19980 >>> >>> You can view the changes in this commit at: >>> ? ?https://svn.nixos.org/viewvc/nix?rev=19980&view=rev >>> >>> Added: >>> ? ?nix/branches/noaterm/ >>> >>> Log: >>> Creating a branch to try to get a nix without aterm. >> >> Can you give some background on this? ?What is the reason for getting rid >> of it? >> >> -- >> Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ >> > _______________________________________________ > nix-dev mailing list > nix-dev at cs.uu.nl > https://mail.cs.uu.nl/mailman/listinfo/nix-dev > > > _______________________________________________ > nix-dev mailing list > nix-dev at cs.uu.nl > https://mail.cs.uu.nl/mailman/listinfo/nix-dev > > -- Nicolas Pierron http://www.linkedin.com/in/nicolasbpierron - http://nbp.name/ Lars Rasmussen (Google IO 2009) - Icland is an icland. From e.dolstra at tudelft.nl Mon Feb 15 09:57:47 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Mon, 15 Feb 2010 09:57:47 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 20000 - sandervanderburg - in nixpkgs/trunk/pkgs/desktops/kde-4.4: . plasma-addons In-Reply-To: <201002142304.o1EN4uUL002534@proliant.st.ewi.tudelft.nl> References: <201002142304.o1EN4uUL002534@proliant.st.ewi.tudelft.nl> Message-ID: <4B790C8B.3040300@tudelft.nl> Hi, Sander van der Burg wrote: > Author: sandervanderburg > Date: 2010-02-14 23:04:56 +0000 (Sun, 14 Feb 2010) > New Revision: 20000 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=20000&view=rev > > Modified: > nixpkgs/trunk/pkgs/desktops/kde-4.4/default.nix > nixpkgs/trunk/pkgs/desktops/kde-4.4/plasma-addons/default.nix > > Log: > Added attica to buildInputs of kdeplasma-addons-4.4.0 Revision 20000! Woohoo! :-) -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From S.vanderBurg at tudelft.nl Mon Feb 15 11:59:01 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Mon, 15 Feb 2010 11:59:01 +0100 Subject: [Nix-dev] RE: [Nix-commits] SVN commit: nix - 20000 - sandervanderburg - in nixpkgs/trunk/pkgs/desktops/kde-4.4: . plasma-addons References: <201002142304.o1EN4uUL002534@proliant.st.ewi.tudelft.nl> <4B790C8B.3040300@tudelft.nl> Message-ID: <8A23E531C514ED43B1B469040352E7BA12947F@SRV503.tudelft.net> Indeed. I'm sooo glad that no. 20000 is mine!!!!! gnagna. I'm also making progress with KDE 4.4 now, so I think I have it working in a couple of days.... -----Original Message----- From: Eelco Dolstra [mailto:e.dolstra at tudelft.nl] Sent: Mon 2/15/2010 9:57 AM To: nix-dev at cs.uu.nl; Sander van der Burg - EWI Subject: Re: [Nix-commits] SVN commit: nix - 20000 - sandervanderburg - in nixpkgs/trunk/pkgs/desktops/kde-4.4: . plasma-addons Hi, Sander van der Burg wrote: > Author: sandervanderburg > Date: 2010-02-14 23:04:56 +0000 (Sun, 14 Feb 2010) > New Revision: 20000 > > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=20000&view=rev > > Modified: > nixpkgs/trunk/pkgs/desktops/kde-4.4/default.nix > nixpkgs/trunk/pkgs/desktops/kde-4.4/plasma-addons/default.nix > > Log: > Added attica to buildInputs of kdeplasma-addons-4.4.0 Revision 20000! Woohoo! :-) -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100215/87eb7d95/attachment.html From e.dolstra at tudelft.nl Tue Feb 16 01:13:47 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Tue, 16 Feb 2010 01:13:47 +0100 Subject: [Nix-dev] hydra and the plus sign In-Reply-To: <45219fb01002130320y2cd62fa0o84ea718f07a4f94a@mail.gmail.com> References: <45219fb01002130320y2cd62fa0o84ea718f07a4f94a@mail.gmail.com> Message-ID: <4B79E33B.50205@tudelft.nl> Hi, Llu?s Batlle wrote: > There is some translation to URL escape codes where it should not be, I imagine. No, this was required in Hydra because some store paths contain the `?' character, which has a special meaning in URIs. For completeness I also escaped `+' and `='. > I have to go without manifests, I guess, until someone fixes this? I've just fixed this in nix-prefetch-url (r20036). -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From viriketo at gmail.com Tue Feb 16 08:55:47 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Tue, 16 Feb 2010 08:55:47 +0100 Subject: [Nix-dev] hydra and the plus sign In-Reply-To: <4B79E33B.50205@tudelft.nl> References: <45219fb01002130320y2cd62fa0o84ea718f07a4f94a@mail.gmail.com> <4B79E33B.50205@tudelft.nl> Message-ID: <45219fb01002152355hfea58aeoe5c841c9e1864bed@mail.gmail.com> Hello, 2010/2/16 Eelco Dolstra : > Hi, > > Llu?s Batlle wrote: > >> There is some translation to URL escape codes where it should not be, I imagine. > > No, this was required in Hydra because some store paths contain the `?' > character, which has a special meaning in URIs. ?For completeness I also escaped > `+' and `='. > >> I have to go without manifests, I guess, until someone fixes this? > > I've just fixed this in nix-prefetch-url (r20036). Where does nix-prefetch-url come into the game, when in a nix-build some closure is downloaded from hydra? Has nix-prefetch-url any other role besides the user running it directly to get the hash of a file? Regards, Llu?s. From e.dolstra at tudelft.nl Tue Feb 16 11:24:38 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Tue, 16 Feb 2010 11:24:38 +0100 Subject: [Nix-dev] hydra and the plus sign In-Reply-To: <45219fb01002152355hfea58aeoe5c841c9e1864bed@mail.gmail.com> References: <45219fb01002130320y2cd62fa0o84ea718f07a4f94a@mail.gmail.com> <4B79E33B.50205@tudelft.nl> <45219fb01002152355hfea58aeoe5c841c9e1864bed@mail.gmail.com> Message-ID: <4B7A7266.4060109@tudelft.nl> Hi, Llu?s Batlle wrote: > 2010/2/16 Eelco Dolstra : >> Hi, >> >> Llu?s Batlle wrote: >> >>> There is some translation to URL escape codes where it should not be, I imagine. >> No, this was required in Hydra because some store paths contain the `?' >> character, which has a special meaning in URIs. For completeness I also escaped >> `+' and `='. >> >>> I have to go without manifests, I guess, until someone fixes this? >> I've just fixed this in nix-prefetch-url (r20036). > Where does nix-prefetch-url come into the game, when in a nix-build > some closure is downloaded from hydra? It's called by download-using-manifests (i.e. the substituter that downloads binaries). -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From S.vanderBurg at tudelft.nl Tue Feb 16 12:06:52 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Tue, 16 Feb 2010 12:06:52 +0100 Subject: [Nix-dev] KDE 4.4 is working more or less Message-ID: <8A23E531C514ED43B1B469040352E7BA129482@SRV503.tudelft.net> Good news, I've managed to get all the KDE 4.4.0 packages working and KDE 4.4.0 also seems to boot and work properly in NixOS when changing kde4 = kde44. I recommend not making KDE 4.4.0 the default KDE4 version yet. Although it is declared "stable", there are still minor issues popping up (for instance very recently a hotfix is released for Soprano, which solves a very critical bug). Moreover, some other minor issues are there, such as a development version of Phonon that we used (it seems that it requires a Phonon release which is not released as stable yet). I think we have to wait at least for KDE 4.4.1 when we want to make this the default KDE version. Many thanks go to Lluis in helping me getting KDE 4.4.x supported and fixed. Best, Sander -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100216/c48a47cf/attachment.html From viriketo at gmail.com Tue Feb 16 12:16:58 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Tue, 16 Feb 2010 12:16:58 +0100 Subject: [Nix-dev] KDE 4.4 is working more or less In-Reply-To: <8A23E531C514ED43B1B469040352E7BA129482@SRV503.tudelft.net> References: <8A23E531C514ED43B1B469040352E7BA129482@SRV503.tudelft.net> Message-ID: <45219fb01002160316w6db92e47l2fe7cf3facf26c7@mail.gmail.com> Great! Don't miss the 'updatekde.sh' script I added to kde43. You can use it for kde44 too (I think!). If not, feel free to fix it. :) Thank you! 2010/2/16 Sander van der Burg - EWI : > Good news, > > I've managed to get all the KDE 4.4.0 packages working and KDE 4.4.0 also > seems to boot and work properly in NixOS when changing kde4 = kde44. > > I recommend not making KDE 4.4.0 the default KDE4 version yet. Although it > is declared "stable", there are still minor issues popping up (for instance > very recently a hotfix is released for Soprano, which solves a very critical > bug). Moreover, some other minor issues are there, such as a development > version of Phonon that we used (it seems that it requires a Phonon release > which is not released as stable yet). > > I think we have to wait at least for KDE 4.4.1 when we want to make this the > default KDE version. > > Many thanks go to Lluis in helping me getting KDE 4.4.x supported and fixed. > > Best, > > Sander > > _______________________________________________ > nix-dev mailing list > nix-dev at cs.uu.nl > https://mail.cs.uu.nl/mailman/listinfo/nix-dev > > From e.dolstra at tudelft.nl Wed Feb 17 10:22:49 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Wed, 17 Feb 2010 10:22:49 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric - nix/branches In-Reply-To: <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> Message-ID: <4B7BB569.7000400@tudelft.nl> Hi, Llu?s Batlle wrote: > I had the feeling that it becomes hard for us to follow the aterm > development. Well, there hardly is any upstream development. So there is not much to follow :-) The only problem is the occasional 64-bit or optimization bug triggered by new GCC releases. Reimplementing aterm-like functionality isn't entirely trivial. For instance, you need a garbage collector (though we could just use the Boehm GC). It might be better to figure out: - Why some people have weird segfaults (I haven't seen them lately myself). - How to reduce memory consumption of the Nix expression evaluator. -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From e.dolstra at tudelft.nl Wed Feb 17 10:38:38 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Wed, 17 Feb 2010 10:38:38 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric -nix/branches In-Reply-To: <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> Message-ID: <4B7BB91E.8050407@tudelft.nl> Hi, Nicolas Pierron wrote: > 2010/2/14 Sander van der Burg - EWI : >> So, what are you guys going to do? Implementing a clean room implementation >> of ATerm with the same interfaces or forking ATerm? And if so, why should >> this ATerm version developed inside the Nix repository? > > Nix does not need ATerm, Nix need maximal sharing to implements > maximal laziness. The garbage collection is almost useless because > all terms are kept alive in the ATermMap which is used to cache reduce > operations. So the garbage collection is just a way to lose some > evaluation time. Not having a garbage collector would doom us forever to unbounded memory use. In the current implementation we just have to improve the evaluation strategy (i.e. not cache every evaluation result) to reduce memory use. > Some manipulations are done to convert data to structure which are > easy to manipulate and to structure which are easy to store. These > conversions are costly, why don't we just use one data format? I'm not sure what you mean. What manipulations and structures? > And as I've experienced a lot of difficulties to debug the ATerm > library (due to all the crap that you don't want to understand). We > don't have access to their repository. We do (well, Hydra does). We even produce tarballs from the ATerm repo: http://hydra.nixos.org/job/meta-environment/aterm-trunk/tarball > They depends on the assembly language. Where? > The library is not thread safe. Having a thread-safe aterm-library would be great. Then we could make nix-env evaluation multi-core :-) -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From simons at cryp.to Thu Feb 18 12:39:07 2010 From: simons at cryp.to (Peter Simons) Date: Thu, 18 Feb 2010 12:39:07 +0100 Subject: [Nix-dev] Revision 19868 breaks checkinstall on RedHat host systems Message-ID: <873a0ydb8k.fsf@write-only.cryp.to> Hi, Revision 19868 introduced a patch to fix compilation of checkinstall on Glibc 2.11, but unfortunately, it breaks compilation of the package on older version -- such as glibc 2.5, which is used on Red Hat host systems. Does anyone have a suggestion how to remedy that problem? Take care, Peter From ludo at gnu.org Thu Feb 18 14:13:22 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 18 Feb 2010 14:13:22 +0100 Subject: [Nix-dev] Re: Revision 19868 breaks checkinstall on RedHat host systems References: <873a0ydb8k.fsf@write-only.cryp.to> Message-ID: <87ljeqit59.fsf@gnu.org> Hi Peter, Peter Simons writes: > Revision 19868 introduced a patch to fix compilation of checkinstall on > Glibc 2.11, but unfortunately, it breaks compilation of the package on > older version -- such as glibc 2.5, which is used on Red Hat host > systems. Note that glibc 2.5 is very old. > Does anyone have a suggestion how to remedy that problem? Perhaps you could make the patch conditional on whether ?stdenv? contains a ?glibc? attribute? Thanks, Ludo?. From simons at cryp.to Thu Feb 18 19:01:51 2010 From: simons at cryp.to (Peter Simons) Date: Thu, 18 Feb 2010 19:01:51 +0100 Subject: [Nix-dev] Re: Revision 19868 breaks checkinstall on RedHat host systems References: <873a0ydb8k.fsf@write-only.cryp.to> <87ljeqit59.fsf@gnu.org> Message-ID: <87wryactio.fsf@write-only.cryp.to> Hi Ludovic, >> Revision 19868 introduced a patch to fix compilation of checkinstall on >> Glibc 2.11, but unfortunately, it breaks compilation of the package on >> older version -- such as glibc 2.5, which is used on Red Hat host >> systems. > > Note that glibc 2.5 is very old. no kidding. That's why it works on systems that have a Red Hat kernel. Unfortunately, using glibc 2.5 is the only way to get a working Nix installation on Red Hat Linux. The issue has been discussed on this list a couple of times. >> Does anyone have a suggestion how to remedy that problem? > > Perhaps you could make the patch conditional on whether ?stdenv? > contains a ?glibc? attribute? That won't help, unfortunately, because the glibc 2.5 systems have a stdenv.glibc attribute, too. It's just that it's version 2.5. Of course, it's possible to use some other suitable mechanism to determine whether that bogus scandir patch can be applied or not, but I guess my favorite solution would be for someone to fix that patch so that it just doesn't break checkinstall. Then we can apply it unconditionally. An obvious candidate for that job might be the person who committed that patch in the first place, maybe? *cough* *cough*. Take care, Peter From rohan.hart17 at gmail.com Thu Feb 18 20:09:53 2010 From: rohan.hart17 at gmail.com (Rohan Hart) Date: Fri, 19 Feb 2010 08:09:53 +1300 Subject: [Nix-dev] doc on removing unused dependencies appears to be incorrect in Nix user's guide Message-ID: The guide (version 0.15pre19948) says that openssl = if sslSupport then openssl else null; will remove the dependency on openssl. This appears not to work although something like this will maybeOpenssl = if sslSupport then openssl else null; buildInputs = [ maybeOpenssl ] Better yet, document use of the standard library function buildInputs = [ ... ] # mandatory dependencies ++ stdenv.lib.optional sslSupport openssl cheers Rohan From ludo at gnu.org Thu Feb 18 20:30:14 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 18 Feb 2010 20:30:14 +0100 Subject: [Nix-dev] Re: Revision 19868 breaks checkinstall on RedHat host systems References: <873a0ydb8k.fsf@write-only.cryp.to> <87ljeqit59.fsf@gnu.org> <87wryactio.fsf@write-only.cryp.to> Message-ID: <87ljeqgx4p.fsf@gnu.org> Hi Peter, Peter Simons writes: > >> Does anyone have a suggestion how to remedy that problem? > > > > Perhaps you could make the patch conditional on whether ?stdenv? > > contains a ?glibc? attribute? > > That won't help, unfortunately, because the glibc 2.5 systems have a > stdenv.glibc attribute, too. It's just that it's version 2.5. Oh, right, I thought you were using ?stdenvNative?. > Of course, it's possible to use some other suitable mechanism to > determine whether that bogus scandir patch can be applied or not, but I > guess my favorite solution would be for someone to fix that patch so > that it just doesn't break checkinstall. Then we can apply it > unconditionally. An obvious candidate for that job might be the person > who committed that patch in the first place, maybe? *cough* *cough*. :-) Ludo?. From S.vanderBurg at tudelft.nl Sun Feb 21 18:10:29 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Sun, 21 Feb 2010 18:10:29 +0100 Subject: [Nix-dev] Specifying licenses on Nix packages Message-ID: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> Hi Nix developers, Lately, Eelco and I are thinking about some licensing issues we've discussed before and we're trying to implement a solution in dealing with these issues. One of the issues we have is that for most packages licenses aren't specified or that the license identifiers aren't well defined. The first thing we have to do is getting consensus about what license identifiers we have and what their meaning is. To solve this. I propose that we define an expression called licenses.nix, which contains all these identifiers and maps them on a description e.g: [ { identifier = "BSD-original"; description = "Original BSD license with advertising clause"; } { identifier = "BSD-2"; description = "2-clause BSD license"; } { identifier = "BSD-3"; description = "3-clause BSD license"; } { identifier = "GPLv2"; description = "GNU General Public License version 2 only"; } { identifier = "GPLv2+"; description = "GNU General Public License version 2 or later"; } { identifier = "GPLv2+classpath"; description = "GNU General Public License version2 or later, with classpath exception"; } { identifier = "GPLv2+OSS"; description = "GNU General Public License version2 or later, with OSI approved licenses exception"; } { identifier = "GPLv3+"; description = "GNU General Public License version 3 or later"; } { identifier = "LGPLv2"; description = "GNU Library General Public License version 2 only"; } { identifier = "LGPLv2.1"; description = "GNU Lesser General Public License version 2.1 only"; } { identifier = "MPL1.1"; description = "Mozilla Public License 1.1"; } ... ] This expression can be used a reference for license identifiers. I noticed that a lot of packages in our repository have no license attribute yet, or the license specification isn't very precise (e.g. some packages have "GPL" as a license, but we don't know whether this means GPLv2 or GPLv3 or maybe GPLv2 or later). Therefore we need to fix those meta attributes and we have to make sure they match license identifiers that we know (i.e. which are defined in licenses.nix). I found some inspiration for license identifiers here: http://fedoraproject.org/wiki/Licensing Policies --------- We can use these license identifiers later on to see whether a package fits in a certain policy. We could define several policies, such as: FSF = [ "BSD-3" "GPLv2" "GPLv3" ... ]; # All licenses that conform to the free software definition defined by the free software foundation OSI = [ "BSD-3" "GPLv2" "GPLv3" ... ]; # All OSI approved licenses DFSG = OSI; # All licenses that conform to the Debian Free Software guidelines freely_distributable = FSF ++ OSI ++ [ "corefonts" ]; # Software which is freely distributable, but not always free as in speech or maybe other custom definitions such as FSF_with_firmware and use a tool that checks of all the installed packages meet a given policy. Disjunctive licenses --------------------- Another issue is that some packages aren't available under a single license, but more disjunctive licenses are applicable (packages such as Mozilla Firefox, Mono, Qt etc.). In such cases the meta.license attribute has to be a list of strings instead of a string. Examples: Mozilla Firefox is triple-licensed under the GPL version 2 or later, LPGL version 2.1 or later or the Mozilla Public License version 1.1: meta.license = [ "GPLv2+" "LGPL2.1+" "MPL1.1" ]; Qt 4.5 is licensed under the GPL version 2 or later (with an exception of linking from every OSI approved license), LGPL version 2.1 or later or a proprietary license: meta.license = [ "GPLv2+OSS" "LGPL2.1+" "proprietary" ]; Collections ----------- Some packages don't really have one single license, but are essentially collections of several components released under multiple licenses. Mono is an example of this, which includes a compiler dual licensed under the X11 license and GPLv2, various other tools released under the GPLv2, runtime libraries released under the LGPLv2.1 and class libraries released under the X11 licence. In such cases we need to specify that the package is a collection of subcomponents, each having its own meta attributes, e.g.: meta.subcomponents = { compiler = { license = [ "X11" "GPLv2" ]; description = "Mono compiler"; }; tools = { license = "GPLv2"; description = "Mono tools"; }; runtimeLibs = { license = "LGPLv2.1"; description = "Mono runtime libraries"; }; classLibs = { license = "X11"; description = "Mono class libraries"; }; }; Maybe when we don't want to explicitly specify everything for a subcomponent the attribute meta.license will automatically apply to all subcomponents, except when it's overridden in meta.subcomponents.*.license. What do you guys think about these notations and conventions? I know this will take some work/effort in implementing it, but I think this is worth the effort and by using this approach we can build a mechanism that automatically verifies/builds a system configuration that conforms to a policy a user wants (e.g. if a user only wants to use free software, we can verify this and so on). Best, Sander -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100221/78955679/attachment.html From marco-oweber at gmx.de Sun Feb 21 20:28:43 2010 From: marco-oweber at gmx.de (Marc Weber) Date: Sun, 21 Feb 2010 20:28:43 +0100 Subject: [Nix-dev] Specifying licenses on Nix packages In-Reply-To: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> Message-ID: <1266780330-sup-8713@nixos> > What do you guys think about these notations and conventions? I know this will take some work/effort in implementing it, but I think this is worth the effort and by using this approach we can build a mechanism that automatically verifies/builds a system configuration that conforms to a policy a user wants (e.g. if a user only wants to use free software, we can verify this and so on). It sounds sound. I'm not missing anything. Sometimes I don't have time to do a lot of research about which license is used. In fact when seeing "GLP" It think it's fine and stop thinking about it. Do you want to require these settings? If so will this lead to two repositiories eg nixpkgs-approved-licenses and nixpkgs-dev or nixpkgs-unofficial ? Go on with your proposal. I just raise the question whether this policy should be enforced to all packages. Marc Weber From viriketo at gmail.com Sun Feb 21 22:58:44 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sun, 21 Feb 2010 22:58:44 +0100 Subject: [Nix-dev] Specifying licenses on Nix packages In-Reply-To: <1266780330-sup-8713@nixos> References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> <1266780330-sup-8713@nixos> Message-ID: <45219fb01002211358g64f4330cs81f130eafed9c902@mail.gmail.com> 2010/2/21 Marc Weber : > Go on with your proposal. I just raise the question whether this > policy should be enforced to all packages. I think we should try to get as close as possible to proper statement of the licenses. That can become an issue, if nixpkgs/NixOS is ever to be used in production, part of a product, where it has to be distributed. Any help from 'the distribution' about the license management should be appreciated. I try to take the time to look for the licenses. And if I don't understand them enough, at least I try to write a comment around the meta.license. From viriketo at gmail.com Mon Feb 22 09:54:55 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Mon, 22 Feb 2010 09:54:55 +0100 Subject: [Nix-dev] qt4 derivation Message-ID: <45219fb01002220054t472c583dm1d4784402bbc5d39@mail.gmail.com> Hello, why we are keeping the 'qt4' derivation attribute to qt 4.4? From viriketo at gmail.com Mon Feb 22 10:00:37 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Mon, 22 Feb 2010 10:00:37 +0100 Subject: [Nix-dev] bug on nix-env version parsing? Message-ID: <45219fb01002220100x7d67b587v3f74b1210a166c2a@mail.gmail.com> Hello, I have this full reproducible: $ nix-env -qc '*' | grep thund thunderbird-3.0 = 3.0 $ nix-env -u --leq --dry-run '*' upgrading `thunderbird-3.0' to `thunderbird-2.0.0.22' ? From marco-oweber at gmx.de Mon Feb 22 11:29:13 2010 From: marco-oweber at gmx.de (Marc Weber) Date: Mon, 22 Feb 2010 11:29:13 +0100 Subject: [Nix-dev] bug on nix-env version parsing? In-Reply-To: <45219fb01002220100x7d67b587v3f74b1210a166c2a@mail.gmail.com> References: <45219fb01002220100x7d67b587v3f74b1210a166c2a@mail.gmail.com> Message-ID: <1266834501-sup-628@nixos> Excerpts from Llu?s Batlle's message of Mon Feb 22 10:00:37 +0100 2010: > I have this full reproducible: > $ nix-env -qc '*' | grep thund > thunderbird-3.0 = 3.0 > $ nix-env -u --leq --dry-run '*' > upgrading `thunderbird-3.0' to `thunderbird-2.0.0.22' thunderbird2 = import ../applications/networking/mailreaders/thunderbird/2.x.nix { thunderbird3 = lowPrio (import ../applications/networking/mailreaders/thunderbird/3.x.nix { ^^^^^^^^ version comparison is fine: builtins.compareVersions "thunderbird-3.0" "thunderbird-2.0.0.22" results in Int(1) Marc Weber From S.vanderBurg at tudelft.nl Mon Feb 22 11:40:21 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Mon, 22 Feb 2010 11:40:21 +0100 Subject: [Nix-dev] qt4 derivation References: <45219fb01002220054t472c583dm1d4784402bbc5d39@mail.gmail.com> Message-ID: <8A23E531C514ED43B1B469040352E7BA129484@SRV503.tudelft.net> Good question. I initially created two derivations, since KDE 4.2.x was not compatible with qt-4.5.x or higher. But we don't maintain KDE 4.2.x anymore, so I think we can make qt-4.5.x the default now (and maybe qt-4.6.x when KDE 4.4.x is stable) -----Original Message----- From: nix-dev-bounces at cs.uu.nl on behalf of Llu?s Batlle Sent: Mon 2/22/2010 9:54 AM To: nix-dev Subject: [Nix-dev] qt4 derivation Hello, why we are keeping the 'qt4' derivation attribute to qt 4.4? _______________________________________________ nix-dev mailing list nix-dev at cs.uu.nl https://mail.cs.uu.nl/mailman/listinfo/nix-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100222/04f738e1/attachment.html From viriketo at gmail.com Mon Feb 22 11:57:44 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Mon, 22 Feb 2010 11:57:44 +0100 Subject: [Nix-dev] bug on nix-env version parsing? In-Reply-To: <1266834501-sup-628@nixos> References: <45219fb01002220100x7d67b587v3f74b1210a166c2a@mail.gmail.com> <1266834501-sup-628@nixos> Message-ID: <45219fb01002220257m751fda86xd9881ebe83997b46@mail.gmail.com> 2010/2/22 Marc Weber : > Excerpts from Llu?s Batlle's message of Mon Feb 22 10:00:37 +0100 2010: >> I have this full reproducible: >> $ nix-env -qc '*' ?| grep thund >> thunderbird-3.0 ? ? ? ? ? ? ? ? ? ?= 3.0 >> $ nix-env -u --leq --dry-run '*' >> upgrading `thunderbird-3.0' to `thunderbird-2.0.0.22' > > ?thunderbird2 = import ../applications/networking/mailreaders/thunderbird/2.x.nix { > ?thunderbird3 = lowPrio (import ../applications/networking/mailreaders/thunderbird/3.x.nix { > ? ? ? ? ? ? ? ?^^^^^^^^ > > version comparison is fine: > > builtins.compareVersions "thunderbird-3.0" "thunderbird-2.0.0.22" > > results in Int(1) Ah, I understand! -u takes into account the priorities before the version checking. And -qc only compares the version. For this specific case, I think I will remove the lowPrio, because thunderbird 2 is not building, if I understand correctly. Regards, Llu?s. From viriketo at gmail.com Mon Feb 22 12:41:32 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Mon, 22 Feb 2010 12:41:32 +0100 Subject: [Nix-dev] qt4 derivation In-Reply-To: <8A23E531C514ED43B1B469040352E7BA129486@SRV503.tudelft.net> References: <45219fb01002220054t472c583dm1d4784402bbc5d39@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA129484@SRV503.tudelft.net> <45219fb01002220247x77a89f48mc62bafc9283de2f4@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA129486@SRV503.tudelft.net> Message-ID: <45219fb01002220341s17aa543cqb6083d98c97b90af@mail.gmail.com> 2010/2/22 Sander van der Burg - EWI : > I haven't checked whether there are problems using qt-4.6.x for KDE 4.3.x. > Maybe it will work fine, but I don't know. I do know that for KDE 4.2.x this > was a problem. > > However, if you decide to make qt-4.6.x the default for KDE 4.3.x you also > have to look into PyQt4, since that also has a dependency on qt-4.5.x and I > don't know if the older PyQt compiles against qt-4.6.x or if the newer PyQt4 > works with KDE 4.3.x. > (I believe the old PyQt4 won't compile against qt-4.6.x) This is enough for me to not to write qt4=qt46; :) I may try the update whenever I can test the change a bit. (I forgot to cc to nix-dev in some letters) From ludo at gnu.org Mon Feb 22 14:28:23 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Mon, 22 Feb 2010 14:28:23 +0100 Subject: [Nix-dev] Old glibc versions In-Reply-To: <201002191540.o1JFechJ030503@proliant.st.ewi.tudelft.nl> (Eelco Dolstra's message of "Fri, 19 Feb 2010 15:40:38 +0000") References: <201002191540.o1JFechJ030503@proliant.st.ewi.tudelft.nl> Message-ID: <87iq9p1jt4.fsf@gnu.org> Hello, Eelco Dolstra writes: > You can view the changes in this commit at: > https://svn.nixos.org/viewvc/nix?rev=20130&view=rev [...] > * Removed glibc 2.10. Since some people are still using 2.5, wouldn?t it make sense to keep such recent versions around, just in case? Also, I think having older glibc versions can be useful for portability/testing purposes. Thanks, Ludo?. From S.vanderBurg at tudelft.nl Mon Feb 22 16:17:39 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Mon, 22 Feb 2010 16:17:39 +0100 Subject: [Nix-dev] Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> <1266780330-sup-8713@nixos> Message-ID: <8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net> I think there is no need in creating a separate branch, since the proposal does not give us conflicts. Eelco developed a prototype tool that can automatically retrieve the license attributes of a closure. We just have to use this tool to see which licenses are missing or incorrectly specified and use this data to correct them. What I need now is consensus about license identifiers. If this consensus is ok to use, then I'll start modifying the incorrect license attributes when I encounter them. The only thing I don't know yet is what Hydra does when the meta.license attribute is a list of strings instead of a list. Maybe Rob or Eelco can clarify this a bit. Moreover, the prototype tool does not yet deal with this either but I don't think it is hard to modify this. -----Original Message----- From: nix-dev-bounces at cs.uu.nl on behalf of Marc Weber Sent: Sun 2/21/2010 8:28 PM To: nix-dev Subject: Re: [Nix-dev] Specifying licenses on Nix packages > What do you guys think about these notations and conventions? I know this will take some work/effort in implementing it, but I think this is worth the effort and by using this approach we can build a mechanism that automatically verifies/builds a system configuration that conforms to a policy a user wants (e.g. if a user only wants to use free software, we can verify this and so on). It sounds sound. I'm not missing anything. Sometimes I don't have time to do a lot of research about which license is used. In fact when seeing "GLP" It think it's fine and stop thinking about it. Do you want to require these settings? If so will this lead to two repositiories eg nixpkgs-approved-licenses and nixpkgs-dev or nixpkgs-unofficial ? Go on with your proposal. I just raise the question whether this policy should be enforced to all packages. Marc Weber _______________________________________________ nix-dev mailing list nix-dev at cs.uu.nl https://mail.cs.uu.nl/mailman/listinfo/nix-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100222/46fdd445/attachment.html From ludo at gnu.org Mon Feb 22 16:24:46 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Mon, 22 Feb 2010 16:24:46 +0100 Subject: [Nix-dev] Re: Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> <1266780330-sup-8713@nixos> <8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net> Message-ID: <87vddpxphd.fsf@gnu.org> Hello Sander, What are you replying to? "Sander van der Burg - EWI" writes: > What I need now is consensus about license identifiers. If this > consensus is ok to use, then I'll start modifying the incorrect > license attributes when I encounter them. Someone added a link to this page on the wiki: . The nice thing is that it contains short names for many licenses, so it could serve as a reference. > The only thing I don't know yet is what Hydra does when the > meta.license attribute is a list of strings instead of a list. Sometimes one wants an ?and? (e.g., GnuTLS is LGPLv2+ but the ?gnutls-cli? program is GPLv3+), sometimes an ?or? (software that is dual-licensed). And that?s for the simplest cases. :-) Thanks, Ludo?. From S.vanderBurg at tudelft.nl Mon Feb 22 16:34:58 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Mon, 22 Feb 2010 16:34:58 +0100 Subject: [Nix-dev] Re: Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net><1266780330-sup-8713@nixos><8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net> <87vddpxphd.fsf@gnu.org> Message-ID: <8A23E531C514ED43B1B469040352E7BA129488@SRV503.tudelft.net> The semantics of a list of strings in my proposal is always disjunctive, so e.g.: meta.license = [ "GPLv2+" "LGPLv2.1+" ]; means that the work is dual-licensed under the GPL version 2 or higher, or the LGPL version 2.1 or higher For the gnutls example you gave, we should specify the licenses like this: meta.subcomponents = { lib = { description = "GNU TLS library"; license = "LGPLv3+"; }; cli = { description = "GNU TLS Command-Line interface"; license = "GPLv3+"; }; }; Which means that the package consists of two subcomponents. The CLI is available under the GPL version 3 or higher and the library under the LGPL version 3 or higher. or maybe like this, if we are too lazy to identify all subcomponents and their licenses: meta.license = "GPLv3+"; meta.subcomponents = { lib = { description = "GNU TLS library"; license = "LGPLv3+"; }; }; Which means that the whole package is available under the GPL version 3 or higher, except for the library subcomponent which is available under the LGPL version 3 or higher. (we even can have disjunctive licenses on subcomponents with this approach) What do you think? -----Original Message----- From: nix-dev-bounces at cs.uu.nl on behalf of Ludovic Court?s Sent: Mon 2/22/2010 4:24 PM To: nix-dev at cs.uu.nl Subject: [Nix-dev] Re: Specifying licenses on Nix packages Hello Sander, What are you replying to? "Sander van der Burg - EWI" writes: > What I need now is consensus about license identifiers. If this > consensus is ok to use, then I'll start modifying the incorrect > license attributes when I encounter them. Someone added a link to this page on the wiki: . The nice thing is that it contains short names for many licenses, so it could serve as a reference. > The only thing I don't know yet is what Hydra does when the > meta.license attribute is a list of strings instead of a list. Sometimes one wants an 'and' (e.g., GnuTLS is LGPLv2+ but the 'gnutls-cli' program is GPLv3+), sometimes an 'or' (software that is dual-licensed). And that's for the simplest cases. :-) Thanks, Ludo'. _______________________________________________ nix-dev mailing list nix-dev at cs.uu.nl https://mail.cs.uu.nl/mailman/listinfo/nix-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100222/5344925c/attachment.html From ludo at gnu.org Mon Feb 22 16:53:57 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Mon, 22 Feb 2010 16:53:57 +0100 Subject: [Nix-dev] Re: Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> <1266780330-sup-8713@nixos> <8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net> <87vddpxphd.fsf@gnu.org> Message-ID: <87wry5uuzu.fsf@gnu.org> Hi again, :-) ludo at gnu.org (Ludovic Court?s) writes: > What are you replying to? For some reason, your initial message didn?t make it through Gmane (!). For those who missed it, here it is: http://mail.cs.uu.nl/pipermail/nix-dev/2010-February/003911.html . We quickly discussed it at FOSDEM, but I?d like to summarize my position. If find the ?license calculus project? you have in mind interesting, but I?m skeptical about its practicality. The reason is that this whole idea assumes that legal texts written in natural languages, subject to human interpretation, can be reasoned about in a mathematical way. I think this assumption doesn?t hold, unfortunately. Even if the interpretation of law and licenses were unambiguous, stable in time, context-independent, etc., other practical issues would make it very hard to automatically reason on software licensing. A few examples: * What is ?the license? of OpenOffice.org? * What is ?the license? of Teeworlds? (It?s home-made.) * What is ?the license? of GNU Guile once you?ve typed ?(use-modules (ice-9 readline))?? (libguile is LGPLv2+, but that module is GPLv3+ as it dlopens GNU Readline.) * Is the Open Font License (OFL) FSF-free? (See http://thread.gmane.org/gmane.linux.distributions.gnu-linux-libre/406 for a discussion.) * Can you write a predicate that tells whether a given library is a ?System Library? according to the GPL? ... As a Free Software supporter, I?m taking the straightforward approach of removing software that doesn?t correspond to my policy: http://repo.or.cz/w/nixpkgs-libre.git . Join us now, share the software! :-) Thanks, Ludo?. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100222/a1fa0f11/attachment.bin From S.vanderBurg at tudelft.nl Mon Feb 22 17:30:02 2010 From: S.vanderBurg at tudelft.nl (Sander van der Burg - EWI) Date: Mon, 22 Feb 2010 17:30:02 +0100 Subject: [Nix-dev] Re: Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net><1266780330-sup-8713@nixos><8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net><87vddpxphd.fsf@gnu.org> <87wry5uuzu.fsf@gnu.org> Message-ID: <8A23E531C514ED43B1B469040352E7BA12948A@SRV503.tudelft.net> My goal is not to develop a system that could verify everything and see whether all the licenses are used properly or not. I know this is only possible when a license is defined in a formal language which we can use to verify a system, which is not the case in practice. What I do want to solve is that we currently have no notion at all about what licenses are applicable to certain components and what eventual problems/issues there are. The advantages/benefits I see of using these notations and conventions is: - We have a notion of what licenses there are specified or what licenses there are missing on packages in the nix repository. Currently, we don't have any notion. - We could immediately see when we install a package whether or not it matches a policy, such as installing a non-free package or a package with a "problematic/vague" license. e.g. a user could decide not installing packages that are (partially) non-free (e.g. a package available under a proprietary license) - We have some (but no complete) insight why a composition is possibly non-free. - We could decide to disable certain builds on the Hydra that we don't want to distribute based on some custom defined constraints I know that there are a lot of cases that we cannot solve, but by having a convention for specifying licenses we at least have some insight/notion in these problems, which is better than not having it all IMHO. As a start we need to specify these license more concise I think, which I proposed in my initial message. From these conventions we can derive other possible use-cases. I know this will not solve all our problems/issues, but it will give us more insight in the licensing structure of Nixpkgs and also offers us more options in dealing with them. -----Original Message----- From: nix-dev-bounces at cs.uu.nl on behalf of Ludovic Court?s Sent: Mon 2/22/2010 4:53 PM To: nix-dev at cs.uu.nl Subject: [Nix-dev] Re: Specifying licenses on Nix packages Hi again, :-) ludo at gnu.org (Ludovic Court?s) writes: > What are you replying to? For some reason, your initial message didn't make it through Gmane (!). For those who missed it, here it is: http://mail.cs.uu.nl/pipermail/nix-dev/2010-February/003911.html . We quickly discussed it at FOSDEM, but I'd like to summarize my position. If find the "license calculus project" you have in mind interesting, but I'm skeptical about its practicality. The reason is that this whole idea assumes that legal texts written in natural languages, subject to human interpretation, can be reasoned about in a mathematical way. I think this assumption doesn't hold, unfortunately. Even if the interpretation of law and licenses were unambiguous, stable in time, context-independent, etc., other practical issues would make it very hard to automatically reason on software licensing. A few examples: * What is "the license" of OpenOffice.org? * What is "the license" of Teeworlds? (It's home-made.) * What is "the license" of GNU Guile once you've typed '(use-modules (ice-9 readline))'? (libguile is LGPLv2+, but that module is GPLv3+ as it dlopens GNU Readline.) * Is the Open Font License (OFL) FSF-free? (See http://thread.gmane.org/gmane.linux.distributions.gnu-linux-libre/406 for a discussion.) * Can you write a predicate that tells whether a given library is a "System Library" according to the GPL? ... As a Free Software supporter, I'm taking the straightforward approach of removing software that doesn't correspond to my policy: http://repo.or.cz/w/nixpkgs-libre.git . Join us now, share the software! :-) Thanks, Ludo'. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100222/eb0b98c0/attachment.html From ludo at gnu.org Tue Feb 23 14:07:18 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Tue, 23 Feb 2010 14:07:18 +0100 Subject: [Nix-dev] Re: Specifying licenses on Nix packages References: <8A23E531C514ED43B1B469040352E7BA129483@SRV503.tudelft.net> <1266780330-sup-8713@nixos> <8A23E531C514ED43B1B469040352E7BA129487@SRV503.tudelft.net> Message-ID: <87ljekksmx.fsf@gnu.org> Hey, BTW, for the ?Related Work? section: ?gnulib-tool? implements a simplistic ?license calculus?: http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=gnulib-tool;h=ba41b6180b5b9742ec9c6b402f91b1a6fe02eaf4;hb=HEAD#l3393 Fun! Thanks, Ludo?. From wahjava.ml at gmail.com Tue Feb 23 17:32:21 2010 From: wahjava.ml at gmail.com (Ashish SHUKLA) Date: Tue, 23 Feb 2010 22:02:21 +0530 Subject: [Nix-dev] Some newbie doubts Message-ID: <86hbp7c3qi.fsf@chateau.d.if> Hi everyone, I used NixOS for few days some months ago, and I liked it. Now, I wanted to switch to it, but I'm having few doubts about the NixOS, and how I can get it to fit my usage needs. At the moment, all of my disks are GPT partitioned, and I use GRUB2 as the bootloader. Last time, I used NixOS, there was no GRUB2 support and one of my disk had old MBR style partitions, so I used GRUB support with my non-GPT disk. So my first question is whether I'm able to use GRUB2 with NixOS, in the same way as it works with GRUB. I searched for GRUB2 support in NixOS on Google, found that it is now possible, but I just wanted to confirm this, as I've found no such mention in the documentation yet. The way I'm going to use is that in my current GRUB2 configuration, I'll add a menu for NixOS, which will load GRUB2 configuration from the partition where I installed NixOS. I also routinely build Emacs (and some other apps) from the git/subversion repositories, and I prefer their installation to be integrated/register with underlying package management system. So, I'm wondering if there is any support in Nix expressions to be able to build packages using VCS (version control system) repositories as the sources instead of some tarballs. Following is how this happens currently during package building process: 1. Check for the presence of repository directory. 2. If repository directory is present, then update the source code using VCS tool (.i.e. git pull, svn update, etc.), else checkout copy from the repository (.i.e. git clone, svn co, etc.). 3. After checkout/updating process, export the source code from repository to a directory (.i.e. git archive, svn export, etc.) 4. Build the source code. 5. Install it in a destination directory. 6. Build a package using the contents of the destination directory. So, I'm wondering if Nix expression library has support for this kind of scenario, or not ? And also I noticed the layout of NixOS website has changed recently. I'm not able to locate list of packages currently available in the stable channel. Can anyone point me to some link which contains the list of packages available in the current 'nixpkgs' collection ? Thanks in advance, Ashish SHUKLA -- They who can give up essential liberty to obtain a little temporary safety, deserve neither liberty nor safety. -- Benjamin Franklin, Memoirs of the life and writings of Benjamin Franklin -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100223/f31bc0c0/attachment.bin From ludo at gnu.org Tue Feb 23 17:55:03 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Tue, 23 Feb 2010 17:55:03 +0100 Subject: [Nix-dev] Re: Some newbie doubts References: <86hbp7c3qi.fsf@chateau.d.if> Message-ID: <87ljejki3c.fsf@gnu.org> Hello, wahjava.ml at gmail.com (Ashish SHUKLA) writes: > At the moment, all of my disks are GPT partitioned, and I use GRUB2 as the > bootloader. NixOS supports GRUB 2 now: just set ?boot.loader.grub.version? = 2. > 1. Check for the presence of repository directory. > 2. If repository directory is present, then update the source code using VCS > tool (.i.e. git pull, svn update, etc.), else checkout copy from the > repository (.i.e. git clone, svn co, etc.). The Nix approach is to start anew to maximize reproducibility, i.e., always use ?git clone?. This is what the ?fetchgit? Nix function does. For instance, you could regularly update the ?emacsSnapshot? package, updating the Git/Bzr revision identifier and corresponding hash every once in a while, for the benefit of all Nixpkgs users. :-) Now, Marc Weber implemented a tool to support the kind of workflow you have in mind, but I don?t know the details. Marc? > And also I noticed the layout of NixOS website has changed recently. I'm not > able to locate list of packages currently available in the stable channel. Can > anyone point me to some link which contains the list of packages available in > the current 'nixpkgs' collection ? Try this: http://hydra.nixos.org/jobset/nixpkgs/trunk . Hydra is the system that continuously builds Nixpkgs. Thanks, Ludo?. From wahjava.ml at gmail.com Wed Feb 24 10:48:38 2010 From: wahjava.ml at gmail.com (Ashish SHUKLA) Date: Wed, 24 Feb 2010 15:18:38 +0530 Subject: [Nix-dev] Re: Some newbie doubts In-Reply-To: <87ljejki3c.fsf@gnu.org> ("Ludovic =?utf-8?Q?Court=C3=A8s=22'?= =?utf-8?Q?s?= message of "Tue, 23 Feb 2010 17:55:03 +0100") References: <86hbp7c3qi.fsf@chateau.d.if> <87ljejki3c.fsf@gnu.org> Message-ID: <86d3zvc6bt.fsf@chateau.d.if> Ludovic Court?s writes: > Hello, Hi, > wahjava.ml at gmail.com (Ashish SHUKLA) > writes: >> At the moment, all of my disks are GPT partitioned, and I use GRUB2 as the >> bootloader. > NixOS supports GRUB 2 now: just set ?boot.loader.grub.version? = 2. This works, though grub.cfg doesn't contain any line like: 'set root=(hd0,5)'. But anyways, I'm trying to understand underlying concepts of Nix and its expressions. Thanks for the help. Ashish -- They who can give up essential liberty to obtain a little temporary safety, deserve neither liberty nor safety. -- Benjamin Franklin, Memoirs of the life and writings of Benjamin Franklin -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100224/987fbdd0/attachment.bin From viriketo at gmail.com Wed Feb 24 23:39:32 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Wed, 24 Feb 2010 23:39:32 +0100 Subject: [Nix-dev] hydra last checked nixpkgs trunk at 14:37 Message-ID: <45219fb01002241439g3e659203r7c4de5cc80160d23@mail.gmail.com> I tried to find a problem in nixpkgs' release.nix, and I could not. An hydra problem, or a nixpkgs problem? Regards, Llu?s. From e.dolstra at tudelft.nl Thu Feb 25 11:30:52 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Thu, 25 Feb 2010 11:30:52 +0100 Subject: [Nix-dev] Re: [Nix-commits] SVN commit: nix - 19980 - viric -nix/branches In-Reply-To: <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> Message-ID: <4B86515C.1000805@tudelft.nl> Hi, Nicolas Pierron wrote: > 2010/2/14 Sander van der Burg - EWI : >> So, what are you guys going to do? Implementing a clean room implementation >> of ATerm with the same interfaces or forking ATerm? And if so, why should >> this ATerm version developed inside the Nix repository? > > Nix does not need ATerm, Nix need maximal sharing to implements > maximal laziness. Having given this some more thought, I think that creating another term library to replace ATerm won't solve our problems at all (and it's rather reinventing the wheel). This is because those problems are caused not by the ATerm library but by the evaluation strategy of the Nix expression evaluator. This evaluation strategy is very simple, since it's entirely based on substituting terms in other terms. For instance, a beta redex ((x: E1) E2) is rewritten to (E1 [x := E2]) (i.e. every occurrence of `x' in E1 is replaced by E2). Such a strategy is normally very slow because it leads to work duplication: if `x' is used multiple times, then E2 will be evaluated multiple times. However, thanks to the normal form cache, we can efficiently see whether we've already evaluated E2. This is very simple to implement (just a few lines of code), but it doesn't scale because the normal form cache prevents garbage collection. Using a different term implementation doesn't fix this: - The normal form cache will still grow without bounds. - You still spend lots of time doing substitutions. (According to Valgrind, Nix spends up to 50% of its time in the substitute() function, and 30% or so of the allocations originate there.) A more conventional way to interpret a functional language is to have an explicit environment mapping variables in scope to their values. I.e., when you evaluate ((x: E1) E2), you add [x := E2] to the environment and evaluate E1. If `x' is needed, you fetch its value E2 from the environment, evaluate it to its normal form E2', and update the environment with [x := E2'] so that you don't need to evaluate `x' again. Or you compile to some byte-code representation where variables point to closures in the heap that are overwritten by their normal forms when entered. I'm sort of leaning towards that approach (compile to Parrot or Guile or whatever). -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From ludo at gnu.org Thu Feb 25 14:38:55 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 25 Feb 2010 14:38:55 +0100 Subject: [Nix-dev] Compiling Nix code References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> Message-ID: <87y6ih77v4.fsf_-_@gnu.org> Hi, Eelco Dolstra writes: > Or you compile to some byte-code representation > where variables point to closures in the heap that are overwritten by their > normal forms when entered. I'm sort of leaning towards that approach (compile > to Parrot or Guile or whatever). Of course, I?m somewhat biased as to which solution you should choose ;-), but here?s how it could work with Guile: - There?d be a compiler front-end for the Nix language. It would compile to the ?Tree-IL? language, which is roughly residual Scheme code after macro expansion [0]. - The lexer and parser themselves can be translated from Flex/Bison to SILex [1] and ?parse-lalr? [2]. - Since Tree-IL is call-by-value, parameters in a function call would need to wrapped in a R5RS ?delay? form or rather a SRFI-45 ?lazy? form [2]. If need be, promises can be made more efficient by adding special VM support. - Guile bindings for Nix? libstore, etc., would have to be implemented so that the compiled Nix code can use them. (There?s a C API to libguile and a new ?dynamic FFI?, which allows the implementation of bindings for C code without writing a single line of C.) - Attribute sets could be implemented on top of Bagwell?s vlist-based hash lists, a purely functional dictionary type [4]. As a bonus Nix programs would be able to use other Guile libraries. Imagine yourself writing a multi-threaded, Unicode-capable, GTK+ or networking app in Nix! :-) Thanks, Ludo?. [0] http://www.gnu.org/software/guile/docs/master/guile.html/Tree_002dIL.html [1] http://www.iro.umontreal.ca/~dube/ (works for Guile.) [2] http://code.google.com/p/lalr-scm/ (available in Guile.) [3] http://srfi.schemers.org/srfi-45/srfi-45.html (not yet available in Guile.) [4] http://hydra.nixos.org/build/278519/download/2/guile.html/VHashes.html -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 197 bytes Desc: not available Url : http://lists.science.uu.nl/pipermail/nix-dev/attachments/20100225/71d28d74/attachment.bin From e.dolstra at tudelft.nl Thu Feb 25 15:25:48 2010 From: e.dolstra at tudelft.nl (Eelco Dolstra) Date: Thu, 25 Feb 2010 15:25:48 +0100 Subject: [Nix-dev] Compiling Nix code In-Reply-To: <87y6ih77v4.fsf_-_@gnu.org> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> Message-ID: <4B86886C.6020303@tudelft.nl> Hi, Ludovic Court?s wrote: > - The lexer and parser themselves can be translated from Flex/Bison to > SILex [1] and ?parse-lalr? [2]. Why would it be necessary to change the parser? (Apart of course from changing the actions in the Bison parser to emit Tree-IL.) -- Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/ From 7c6f434c at mail.ru Thu Feb 25 15:39:06 2010 From: 7c6f434c at mail.ru (Michael Raskin) Date: Thu, 25 Feb 2010 17:39:06 +0300 Subject: [Nix-dev] Compiling Nix code In-Reply-To: <87y6ih77v4.fsf_-_@gnu.org> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> Message-ID: <4B868B8A.7050007@mail.ru> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/25/2010 04:38 PM, Ludovic Court?s wrote: >> Or you compile to some byte-code representation >> where variables point to closures in the heap that are overwritten by their >> normal forms when entered. I'm sort of leaning towards that approach (compile >> to Parrot or Guile or whatever). > > Of course, I?m somewhat biased as to which solution you should choose > ;-), but here?s how it could work with Guile: > > - There?d be a compiler front-end for the Nix language. It would > compile to the ?Tree-IL? language, which is roughly residual Scheme > code after macro expansion [0]. Now what is the benefit of not compiling into Scheme code for an actual compiler? > - The lexer and parser themselves can be translated from Flex/Bison to > SILex [1] and ?parse-lalr? [2]. Or just used as is in a shared library. Less changes - less bugs in transition to hunt. > - Since Tree-IL is call-by-value, parameters in a function call would > need to wrapped in a R5RS ?delay? form or rather a SRFI-45 ?lazy? > form [2]. If need be, promises can be made more efficient by adding > special VM support. Fortunately, that can be done in any Scheme and in many functional languages. > - Attribute sets could be implemented on top of Bagwell?s vlist-based > hash lists, a purely functional dictionary type [4]. Why not use whatever hashtable performs best? Obviously, Nix code will not be able > As a bonus Nix programs would be able to use other Guile libraries. > Imagine yourself writing a multi-threaded, Unicode-capable, GTK+ or Unicode-capable? Please make Unicode-capable version of Guile the default in Nixpkgs. (string-length "?") -> 2. Yes, guile_1_9 just fails on this test in a UTF-8 locale. mzscheme and ikarus handle this correctly - just in case anyone cares. Among Lisps sbcl, clisp and ecl also handle this as they should. > networking app in Nix! :-) Great. There goes reproducibility. I thought about adding new possibilities to Nix to enable loading shared libraries long ago, but never found a way to check that it will not add too much randomness into evaluation. Why would I want to write a GUI application in Nix?? -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJLhouJAAoJEE6tnN0aWvw3RQIH/Azk1gjCljBIgZiY7HBS7juv gFOaOhjoHLxWnAfLoAjUycBndkEUMBziDaPXwPQHlGydBVXVkEP/TxxOWzgl5Oly 4weT3fYB2LxRyhzi/i/tgRqzHGMUK1jmu9VhR8SkP9MhjiPrlRfcbcT5kYHeE+pB oRfy678D3lG58aGp7C0QGSjACmwJ/oM7059osEB05RdMOFObxCRjZoBl4kL2pt79 zPKIewQ4LsTxTry2DGerDE0rM7zxuIyoa4LcxuAIB3YfNqq3+3P9HTPkES/4T2ig QHp4nqEVmL5I57NpGmnLb6WvYrsKTmkFd0psHDd7ErWuDKZVU5zumvn7xMFVH0E= =qaeg -----END PGP SIGNATURE----- From ludo at gnu.org Thu Feb 25 15:40:29 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 25 Feb 2010 15:40:29 +0100 Subject: [Nix-dev] Re: Compiling Nix code References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> <4B86886C.6020303@tudelft.nl> Message-ID: <87sk8p750i.fsf@gnu.org> Hi, Eelco Dolstra writes: > Ludovic Court?s wrote: > >> - The lexer and parser themselves can be translated from Flex/Bison to >> SILex [1] and ?parse-lalr? [2]. > > Why would it be necessary to change the parser? (Apart of course from changing > the actions in the Bison parser to emit Tree-IL.) Tree-IL has a Scheme API, so I was considering a front-end written in Scheme. Thanks, Ludo?. From ludo at gnu.org Thu Feb 25 15:46:24 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 25 Feb 2010 15:46:24 +0100 Subject: [Nix-dev] Re: Compiling Nix code References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> <4B868B8A.7050007@mail.ru> Message-ID: <87mxyx74qn.fsf@gnu.org> Hi, Michael Raskin <7c6f434c at mail.ru> writes: > Unicode-capable? Please make Unicode-capable version of Guile the > default in Nixpkgs. (string-length "?") -> 2. It works for me: --8<---------------cut here---------------start------------->8--- scheme@(guile-user)> (setlocale LC_ALL "") $3 = "LC_CTYPE=en_US.UTF-8;LC_NUMERIC=en_US.UTF-8;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=fr_FR.UTF-8;LC_NAME=en_US.UTF-8;LC_ADDRESS=en_US.UTF-8;LC_TELEPHONE=en_US.UTF-8;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=en_US.UTF-8" scheme@(guile-user)> (string-length "?") $4 = 1 --8<---------------cut here---------------end--------------->8--- Please direct bug reports to ?bug-guile at gnu.org? and gratuitous bashing to /dev/null. Thanks, Ludo?. From 7c6f434c at mail.ru Thu Feb 25 16:04:23 2010 From: 7c6f434c at mail.ru (Michael Raskin) Date: Thu, 25 Feb 2010 18:04:23 +0300 Subject: [Nix-dev] Re: Compiling Nix code In-Reply-To: <87mxyx74qn.fsf@gnu.org> References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> <4B868B8A.7050007@mail.ru> <87mxyx74qn.fsf@gnu.org> Message-ID: <4B869177.60004@mail.ru> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/25/2010 05:46 PM, Ludovic Court?s wrote: >> Unicode-capable? Please make Unicode-capable version of Guile the >> default in Nixpkgs. (string-length "?") -> 2. > > It works for me: > > --8<---------------cut here---------------start------------->8--- > scheme@(guile-user)> (setlocale LC_ALL "") > $3 = "LC_CTYPE=en_US.UTF-8;LC_NUMERIC=en_US.UTF-8;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=fr_FR.UTF-8;LC_NAME=en_US.UTF-8;LC_ADDRESS=en_US.UTF-8;LC_TELEPHONE=en_US.UTF-8;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=en_US.UTF-8" > scheme@(guile-user)> (string-length "?") > $4 = 1 > --8<---------------cut here---------------end--------------->8--- > > Please direct bug reports to ?bug-guile at gnu.org? and gratuitous bashing > to /dev/null. So I have to explicitly set locale inside Guile for it to take the locale from environment? OK, in this case it is not exactly bug. I did a completely out-of-the-box experiment. Thanks for the hint, I'll add that line to .guile just in case. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJLhpF2AAoJEE6tnN0aWvw3eqYIAKnz/ewq1S0r8tKNFxSQH1+G x0oIVDkqrTJddSiM2nlto2DvwkATrfxJHL58bv8RxYR1l+XW1fFbVGGXv/WgY6XH TZYpaAHR4ZQDAj9FqO12RwyWG7llatCZ2sZO1RCh+N1HIMy8AJjocyTc3Ht5Er7u PaGOMfunIw7h9jOKnj1yU53EUiqiOuQLt7B8cQqoKeZ0n5GiUPzrgfbNoOlSVMRa qpxBuuye8WqE/Q3VwPJ+4rV674PqROty20gPGF+ZdQlyRNDmomwchMucyD24k4/X yeRk1PC3B+nPFcEM5EMBzyLagF1g9L6otOkiU8/8Nx7f89i1btIwrJLHh5eSZVk= =g1q7 -----END PGP SIGNATURE----- From ludo at gnu.org Thu Feb 25 16:20:16 2010 From: ludo at gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Date: Thu, 25 Feb 2010 16:20:16 +0100 Subject: [Nix-dev] Re: Compiling Nix code In-Reply-To: <4B869177.60004@mail.ru> (Michael Raskin's message of "Thu, 25 Feb 2010 18:04:23 +0300") References: <201002131627.o1DGRpi9002875@proliant.st.ewi.tudelft.nl> <4B7822C8.5080107@tudelft.nl> <45219fb01002140828q2c3c8c01ib06ab124e7b85663@mail.gmail.com> <8A23E531C514ED43B1B469040352E7BA12947E@SRV503.tudelft.net> <787b29b91002141458u62f1e694j16bbda93ee25856b@mail.gmail.com> <4B86515C.1000805@tudelft.nl> <87y6ih77v4.fsf_-_@gnu.org> <4B868B8A.7050007@mail.ru> <87mxyx74qn.fsf@gnu.org> <4B869177.60004@mail.ru> Message-ID: <87fx4p7367.fsf@gnu.org> Hi, Michael Raskin <7c6f434c at mail.ru> writes: > On 02/25/2010 05:46 PM, Ludovic Court?s wrote: >>> Unicode-capable? Please make Unicode-capable version of Guile the >>> default in Nixpkgs. (string-length "?") -> 2. >> >> It works for me: >> >> --8<---------------cut here---------------start------------->8--- >> scheme@(guile-user)> (setlocale LC_ALL "") >> $3 = "LC_CTYPE=en_US.UTF-8;LC_NUMERIC=en_US.UTF-8;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=en_US.UTF-8;LC_MESSAGES=en_US.UTF-8;LC_PAPER=fr_FR.UTF-8;LC_NAME=en_US.UTF-8;LC_ADDRESS=en_US.UTF-8;LC_TELEPHONE=en_US.UTF-8;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=en_US.UTF-8" >> scheme@(guile-user)> (string-length "?") >> $4 = 1 >> --8<---------------cut here---------------end--------------->8--- >> >> Please direct bug reports to ?bug-guile at gnu.org? and gratuitous bashing >> to /dev/null. > > So I have to explicitly set locale inside Guile for it to take the > locale from environment? Yes, like in C on POSIX systems. Thanks, Ludo?. From viriketo at gmail.com Sat Feb 27 16:46:16 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sat, 27 Feb 2010 16:46:16 +0100 Subject: [Nix-dev] Understanding nix evaluation Message-ID: <45219fb01002270746l65ea67afh343f991e2cf91acd@mail.gmail.com> Hello, maybe the question goes mostly to Eelco. I still have not understood why the evaluation fails if we remove the cache in evalExpr(). Isn't the aterm garbage collection working well? Are we using Aterms in such a way that the garbage collection cannot work? Regards, Llu?s. From viriketo at gmail.com Sat Feb 27 17:29:54 2010 From: viriketo at gmail.com (=?UTF-8?Q?Llu=C3=ADs_Batlle?=) Date: Sat, 27 Feb 2010 17:29:54 +0100 Subject: [Nix-dev] Re: Understanding nix evaluation In-Reply-To: <45219fb01002270746l65ea67afh343f991e2cf91acd@mail.gmail.com> References: <45219fb01002270746l65ea67afh343f991e2cf91acd@mail.gmail.com> Message-ID: <45219fb01002270829w4f61e980kfd8c0e5f4cbce46e@mail.gmail.com> Nicolas pointed me to the Maximal Laziness paper, which answered my question. What I now don't understand, is how to make nix use less memory, still keeping the maximal laziness property. 2010/2/27 Llu?s Batlle : > Hello, > > maybe the question goes mostly to Eelco. > > I still have not understood why the evaluation fails if we remove the > cache in evalExpr(). Isn't the aterm garbage collection working well? > Are we using Aterms in such a way that the garbage collection cannot > work? > > Regards, > Llu?s. >