path: root/doc/functions.xml

<chapter xmlns="http://docbook.org/ns/docbook"
	 xmlns:xlink="http://www.w3.org/1999/xlink"
	 xml:id="chap-functions">

<title>Functions reference</title>

<para>
  The nixpkgs repository has several utility functions to manipulate Nix expressions.
</para>

<section xml:id="sec-overrides">
  <title>Overriding</title>

  <para>
    Sometimes one wants to override parts of
    <literal>nixpkgs</literal>, e.g. derivation attributes, the results of
    derivations or even the whole package set.
  </para>

  <section xml:id="sec-pkg-override">
    <title>&lt;pkg&gt;.override</title>

    <para>
      The function <varname>override</varname> is usually available for all the
      derivations in the nixpkgs expression (<varname>pkgs</varname>).
    </para>
    <para>
      It is used to override the arguments passed to a function.
    </para>
    <para>
      Example usages:

      <programlisting>pkgs.foo.override { arg1 = val1; arg2 = val2; ... }</programlisting>
      <programlisting>import pkgs.path { overlays = [ (self: super: {
    foo = super.foo.override { barSupport = true ; };
  })]};</programlisting>
      <programlisting>mypkg = pkgs.callPackage ./mypkg.nix {
    mydep = pkgs.mydep.override { ... };
  }</programlisting>
    </para>

    <para>
      In the first example, <varname>pkgs.foo</varname> is the result of a function call
      with some default arguments, usually a derivation.
      Using <varname>pkgs.foo.override</varname> will call the same function with
      the given new arguments.
    </para>

  </section>

  <section xml:id="sec-pkg-overrideAttrs">
    <title>&lt;pkg&gt;.overrideAttrs</title>

    <para>
      The function <varname>overrideAttrs</varname> allows overriding the
      attribute set passed to a <varname>stdenv.mkDerivation</varname> call,
      producing a new derivation based on the original one.
      This function is available on all derivations produced by the
      <varname>stdenv.mkDerivation</varname> function, which is most packages
      in the nixpkgs expression <varname>pkgs</varname>.
    </para>

    <para>
      Example usage:

      <programlisting>helloWithDebug = pkgs.hello.overrideAttrs (oldAttrs: rec {
    separateDebugInfo = true;
  });</programlisting>
    </para>

    <para>
      In the above example, the <varname>separateDebugInfo</varname> attribute is
      overridden to be true, thus building debug info for
      <varname>helloWithDebug</varname>, while all other attributes will be
      retained from the original <varname>hello</varname> package.
    </para>

    <para>
      The argument <varname>oldAttrs</varname> is conventionally used to refer to
      the attr set originally passed to <varname>stdenv.mkDerivation</varname>.
    </para>

    <note>
      <para>
        Note that <varname>separateDebugInfo</varname> is processed only by the
        <varname>stdenv.mkDerivation</varname> function, not the generated, raw
        Nix derivation. Thus, using <varname>overrideDerivation</varname> will
        not work in this case, as it overrides only the attributes of the final
        derivation. It is for this reason that <varname>overrideAttrs</varname>
        should be preferred in (almost) all cases to
        <varname>overrideDerivation</varname>, i.e. to allow using
        <varname>sdenv.mkDerivation</varname> to process input arguments, as well
        as the fact that it is easier to use (you can use the same attribute
        names you see in your Nix code, instead of the ones generated (e.g.
        <varname>buildInputs</varname> vs <varname>nativeBuildInputs</varname>,
        and involves less typing.
      </para>
    </note>

  </section>


  <section xml:id="sec-pkg-overrideDerivation">
    <title>&lt;pkg&gt;.overrideDerivation</title>

    <warning>
      <para>You should prefer <varname>overrideAttrs</varname> in almost all
      cases, see its documentation for the reasons why.
      <varname>overrideDerivation</varname> is not deprecated and will continue
      to work, but is less nice to use and does not have as many abilities as
      <varname>overrideAttrs</varname>.
      </para>
    </warning>

    <warning>
      <para>Do not use this function in Nixpkgs as it evaluates a Derivation
      before modifying it, which breaks package abstraction and removes
      error-checking of function arguments. In addition, this
      evaluation-per-function application incurs a performance penalty,
      which can become a problem if many overrides are used.
      It is only intended for ad-hoc customisation, such as in
      <filename>~/.config/nixpkgs/config.nix</filename>.
    </para>
    </warning>

    <para>
      The function <varname>overrideDerivation</varname> creates a new derivation
      based on an existing one by overriding the original's attributes with
      the attribute set produced by the specified function.
      This function is available on all
      derivations defined using the <varname>makeOverridable</varname> function.
      Most standard derivation-producing functions, such as
      <varname>stdenv.mkDerivation</varname>, are defined using this
      function, which means most packages in the nixpkgs expression,
      <varname>pkgs</varname>, have this function.
    </para>

    <para>
      Example usage:

      <programlisting>mySed = pkgs.gnused.overrideDerivation (oldAttrs: {
    name = "sed-4.2.2-pre";
    src = fetchurl {
      url = ftp://alpha.gnu.org/gnu/sed/sed-4.2.2-pre.tar.bz2;
      sha256 = "11nq06d131y4wmf3drm0yk502d2xc6n5qy82cg88rb9nqd2lj41k";
    };
    patches = [];
  });</programlisting>
    </para>

    <para>
      In the above example, the <varname>name</varname>, <varname>src</varname>,
      and <varname>patches</varname> of the derivation will be overridden, while
      all other attributes will be retained from the original derivation.
    </para>

    <para>
      The argument <varname>oldAttrs</varname> is used to refer to the attribute set of
      the original derivation.
    </para>

    <note>
      <para>
        A package's attributes are evaluated *before* being modified by
        the <varname>overrideDerivation</varname> function.
        For example, the <varname>name</varname> attribute reference
        in <varname>url = "mirror://gnu/hello/${name}.tar.gz";</varname>
        is filled-in *before* the <varname>overrideDerivation</varname> function
        modifies the attribute set. This means that overriding the
        <varname>name</varname> attribute, in this example, *will not* change the
        value of the <varname>url</varname> attribute. Instead, we need to override
        both the <varname>name</varname> *and* <varname>url</varname> attributes.
      </para>
    </note>

  </section>

  <section xml:id="sec-lib-makeOverridable">
    <title>lib.makeOverridable</title>

    <para>
      The function <varname>lib.makeOverridable</varname> is used to make the result
      of a function easily customizable. This utility only makes sense for functions
      that accept an argument set and return an attribute set.
    </para>

    <para>
      Example usage:

      <programlisting>f = { a, b }: { result = a+b; }
  c = lib.makeOverridable f { a = 1; b = 2; }</programlisting>

    </para>

    <para>
      The variable <varname>c</varname> is the value of the <varname>f</varname> function
      applied with some default arguments. Hence the value of <varname>c.result</varname>
      is <literal>3</literal>, in this example.
    </para>

    <para>
      The variable <varname>c</varname> however also has some additional functions, like
      <link linkend="sec-pkg-override">c.override</link> which can be used to
      override the default arguments. In this example the value of
      <varname>(c.override { a = 4; }).result</varname> is 6.
    </para>

  </section>

</section>

<section xml:id="sec-generators">
  <title>Generators</title>

  <para>
    Generators are functions that create file formats from nix
    data structures, e. g. for configuration files.
    There are generators available for: <literal>INI</literal>,
    <literal>JSON</literal> and <literal>YAML</literal>
  </para>

  <para>
    All generators follow a similar call interface: <code>generatorName
    configFunctions data</code>, where <literal>configFunctions</literal> is
    an attrset of user-defined functions that format nested parts of the
    content.
    They each have common defaults, so often they do not need to be set
    manually. An example is <code>mkSectionName ? (name: libStr.escape [ "[" "]"
    ] name)</code> from the <literal>INI</literal> generator. It receives the
    name of a section and sanitizes it. The default
    <literal>mkSectionName</literal> escapes <literal>[</literal> and
    <literal>]</literal> with a backslash.
  </para>

  <para>
    Generators can be fine-tuned to produce exactly the file format required
    by your application/service. One example is an INI-file format which uses
    <literal>: </literal> as separator, the strings
    <literal>"yes"</literal>/<literal>"no"</literal> as boolean values
    and requires all string values to be quoted:
  </para>

<programlisting>
with lib;
let
  customToINI = generators.toINI {
    # specifies how to format a key/value pair
    mkKeyValue = generators.mkKeyValueDefault {
      # specifies the generated string for a subset of nix values
      mkValueString = v:
             if v == true then ''"yes"''
        else if v == false then ''"no"''
        else if isString v then ''"${v}"''
        # and delegats all other values to the default generator
        else generators.mkValueStringDefault {} v;
    } ":";
  };

# the INI file can now be given as plain old nix values
in customToINI {
  main = {
    pushinfo = true;
    autopush = false;
    host = "localhost";
    port = 42;
  };
  mergetool = {
    merge = "diff3";
  };
}
</programlisting>

  <para>This will produce the following INI file as nix string:</para>

<programlisting>
[main]
autopush:"no"
host:"localhost"
port:42
pushinfo:"yes"
str\:ange:"very::strange"

[mergetool]
merge:"diff3"
</programlisting>

  <note><para>Nix store paths can be converted to strings by enclosing a
  derivation attribute like so: <code>"${drv}"</code>.</para></note>

  <para>
    Detailed documentation for each generator can be found in
    <literal>lib/generators.nix</literal>.
  </para>

</section>

<section xml:id="sec-debug">
  <title>Debugging Nix Expressions</title>

  <para>Nix is a unityped, dynamic language, this means every value can
  potentially appear anywhere. Since it is also non-strict, evaluation order
  and what ultimately is evaluated might surprise you. Therefore it is important
  to be able to debug nix expressions.</para>


  <para>In the <literal>lib/debug.nix</literal> file you will find a number of
  functions that help (pretty-)printing values while evaluation is runnnig. You
  can even specify how deep these values should be printed recursively, and
  transform them on the fly. Please consult the docstrings in
  <literal>lib/debug.nix</literal> for usage information.</para>
</section>


<section xml:id="sec-fhs-environments">
  <title>buildFHSUserEnv</title>

  <para>
    <function>buildFHSUserEnv</function> provides a way to build and run
    FHS-compatible lightweight sandboxes. It creates an isolated root with
    bound <filename>/nix/store</filename>, so its footprint in terms of disk
    space needed is quite small. This allows one to run software which is hard or
    unfeasible to patch for NixOS -- 3rd-party source trees with FHS assumptions,
    games distributed as tarballs, software with integrity checking and/or external
    self-updated binaries. It uses Linux namespaces feature to create
    temporary lightweight environments which are destroyed after all child
    processes exit, without root user rights requirement. Accepted arguments are:
  </para>

  <variablelist>
    <varlistentry>
    <term><literal>name</literal></term>

    <listitem><para>Environment name.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>targetPkgs</literal></term>

    <listitem><para>Packages to be installed for the main host's architecture
    (i.e. x86_64 on x86_64 installations). Along with libraries binaries are also
    installed.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>multiPkgs</literal></term>

    <listitem><para>Packages to be installed for all architectures supported by
    a host (i.e. i686 and x86_64 on x86_64 installations). Only libraries are
    installed by default.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>extraBuildCommands</literal></term>

    <listitem><para>Additional commands to be executed for finalizing the
    directory structure.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>extraBuildCommandsMulti</literal></term>

    <listitem><para>Like <literal>extraBuildCommands</literal>, but
    executed only on multilib architectures.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>extraOutputsToInstall</literal></term>

    <listitem><para>Additional derivation outputs to be linked for both
    target and multi-architecture packages.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>extraInstallCommands</literal></term>

    <listitem><para>Additional commands to be executed for finalizing the
    derivation with runner script.</para></listitem>
    </varlistentry>

    <varlistentry>
    <term><literal>runScript</literal></term>

    <listitem><para>A command that would be executed inside the sandbox and
    passed all the command line arguments. It defaults to
    <literal>bash</literal>.</para></listitem>
    </varlistentry>
  </variablelist>

  <para>
    One can create a simple environment using a <literal>shell.nix</literal>
    like that:
  </para>

<programlisting><![CDATA[
{ pkgs ? import <nixpkgs> {} }:

(pkgs.buildFHSUserEnv {
  name = "simple-x11-env";
  targetPkgs = pkgs: (with pkgs;
    [ udev
      alsaLib
    ]) ++ (with pkgs.xorg;
    [ libX11
      libXcursor
      libXrandr
    ]);
  multiPkgs = pkgs: (with pkgs;
    [ udev
      alsaLib
    ]);
  runScript = "bash";
}).env
]]></programlisting>

  <para>
    Running <literal>nix-shell</literal> would then drop you into a shell with
    these libraries and binaries available. You can use this to run
    closed-source applications which expect FHS structure without hassles:
    simply change <literal>runScript</literal> to the application path,
    e.g. <filename>./bin/start.sh</filename> -- relative paths are supported.
  </para>
</section>

<section xml:id="sec-pkgs-dockerTools">
<title>pkgs.dockerTools</title>

<para>
  <varname>pkgs.dockerTools</varname> is a set of functions for creating and
  manipulating Docker images according to the
  <link xlink:href="https://github.com/moby/moby/blob/master/image/spec/v1.2.md#docker-image-specification-v120">
  Docker Image Specification v1.2.0
  </link>. Docker itself is not used to perform any of the operations done by these
  functions.
</para>

<warning>
  <para>
  The <varname>dockerTools</varname> API is unstable and may be subject to
  backwards-incompatible changes in the future.
  </para>
</warning>

<section xml:id="ssec-pkgs-dockerTools-buildImage">
  <title>buildImage</title>

  <para>
  This function is analogous to the <command>docker build</command> command,
  in that can used to build a Docker-compatible repository tarball containing
  a single image with one or multiple layers. As such, the result
  is suitable for being loaded in Docker with <command>docker load</command>.
  </para>

  <para>
  The parameters of <varname>buildImage</varname> with relative example values are
  described below:
  </para>

  <example xml:id='ex-dockerTools-buildImage'><title>Docker build</title>
  <programlisting>
  buildImage {
    name = "redis"; <co xml:id='ex-dockerTools-buildImage-1' />
    tag = "latest"; <co xml:id='ex-dockerTools-buildImage-2' />

    fromImage = someBaseImage; <co xml:id='ex-dockerTools-buildImage-3' />
    fromImageName = null; <co xml:id='ex-dockerTools-buildImage-4' />
    fromImageTag = "latest"; <co xml:id='ex-dockerTools-buildImage-5' />

    contents = pkgs.redis; <co xml:id='ex-dockerTools-buildImage-6' />
    runAsRoot = '' <co xml:id='ex-dockerTools-buildImage-runAsRoot' />
      #!${stdenv.shell}
      mkdir -p /data
    '';

    config = { <co xml:id='ex-dockerTools-buildImage-8' />
      Cmd = [ "/bin/redis-server" ];
      WorkingDir = "/data";
      Volumes = {
        "/data" = {};
      };
    };
  }
  </programlisting>
  </example>

  <para>The above example will build a Docker image <literal>redis/latest</literal>
  from the given base image. Loading and running this image in Docker results in
  <literal>redis-server</literal> being started automatically.
  </para>

  <calloutlist>
  <callout arearefs='ex-dockerTools-buildImage-1'>
    <para>
    <varname>name</varname> specifies the name of the resulting image.
    This is the only required argument for <varname>buildImage</varname>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-2'>
    <para>
    <varname>tag</varname> specifies the tag of the resulting image.
    By default it's <literal>latest</literal>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-3'>
    <para>
    <varname>fromImage</varname> is the repository tarball containing the base image.
    It must be a valid Docker image, such as exported by <command>docker save</command>.
    By default it's <literal>null</literal>, which can be seen as equivalent
    to <literal>FROM scratch</literal> of a <filename>Dockerfile</filename>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-4'>
    <para>
    <varname>fromImageName</varname> can be used to further specify
    the base image within the repository, in case it contains multiple images.
    By default it's <literal>null</literal>, in which case
    <varname>buildImage</varname> will peek the first image available
    in the repository.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-5'>
    <para>
    <varname>fromImageTag</varname> can be used to further specify the tag
    of the base image within the repository, in case an image contains multiple tags.
    By default it's <literal>null</literal>, in which case
    <varname>buildImage</varname> will peek the first tag available for the base image.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-6'>
    <para>
    <varname>contents</varname> is a derivation that will be copied in the new
    layer of the resulting image. This can be similarly seen as
    <command>ADD contents/ /</command> in a <filename>Dockerfile</filename>.
    By default it's <literal>null</literal>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-runAsRoot'>
    <para>
    <varname>runAsRoot</varname> is a bash script that will run as root
    in an environment that overlays the existing layers of the base image with
    the new resulting layer, including the previously copied
    <varname>contents</varname> derivation.
    This can be similarly seen as
    <command>RUN ...</command> in a <filename>Dockerfile</filename>.

    <note>
      <para>
      Using this parameter requires the <literal>kvm</literal>
      device to be available.
      </para>
    </note>
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-buildImage-8'>
    <para>
    <varname>config</varname> is used to specify the configuration of the
    containers that will be started off the built image in Docker.
    The available options are listed in the
    <link xlink:href="https://github.com/moby/moby/blob/master/image/spec/v1.2.md#image-json-field-descriptions">
      Docker Image Specification v1.2.0
    </link>.
    </para>
  </callout>

  </calloutlist>

  <para>
  After the new layer has been created, its closure
  (to which <varname>contents</varname>, <varname>config</varname> and
  <varname>runAsRoot</varname> contribute) will be copied in the layer itself.
  Only new dependencies that are not already in the existing layers will be copied.
  </para>

  <para>
  At the end of the process, only one new single layer will be produced and
  added to the resulting image.
  </para>

  <para>
  The resulting repository will only list the single image
  <varname>image/tag</varname>. In the case of <xref linkend='ex-dockerTools-buildImage'/>
  it would be <varname>redis/latest</varname>.
  </para>

  <para>
  It is possible to inspect the arguments with which an image was built
  using its <varname>buildArgs</varname> attribute.
  </para>



  <note>
  <para>
  If you see errors similar to <literal>getProtocolByName: does not exist (no such protocol name: tcp)</literal>
  you may need to add <literal>pkgs.iana-etc</literal> to <varname>contents</varname>.
  </para>
  </note>

  <note>
  <para>
  If you see errors similar to <literal>Error_Protocol ("certificate has unknown CA",True,UnknownCa)</literal>
  you may need to add <literal>pkgs.cacert</literal> to <varname>contents</varname>.
  </para>
  </note>

</section>

<section xml:id="ssec-pkgs-dockerTools-fetchFromRegistry">
  <title>pullImage</title>

  <para>
  This function is analogous to the <command>docker pull</command> command,
  in that can be used to fetch a Docker image from a Docker registry.
  Currently only registry <literal>v1</literal> is supported.
  By default <link xlink:href="https://hub.docker.com/">Docker Hub</link>
  is used to pull images.
  </para>

  <para>
  Its parameters are described in the example below:
  </para>

  <example xml:id='ex-dockerTools-pullImage'><title>Docker pull</title>
  <programlisting>
  pullImage {
    imageName = "debian"; <co xml:id='ex-dockerTools-pullImage-1' />
    imageTag = "jessie"; <co xml:id='ex-dockerTools-pullImage-2' />
    imageId = null; <co xml:id='ex-dockerTools-pullImage-3' />
    sha256 = "1bhw5hkz6chrnrih0ymjbmn69hyfriza2lr550xyvpdrnbzr4gk2"; <co xml:id='ex-dockerTools-pullImage-4' />

    indexUrl = "https://index.docker.io"; <co xml:id='ex-dockerTools-pullImage-5' />
    registryVersion = "v1";
  }
  </programlisting>
  </example>

  <calloutlist>
  <callout arearefs='ex-dockerTools-pullImage-1'>
    <para>
    <varname>imageName</varname> specifies the name of the image to be downloaded,
    which can also include the registry namespace (e.g. <literal>library/debian</literal>).
    This argument is required.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-pullImage-2'>
    <para>
    <varname>imageTag</varname> specifies the tag of the image to be downloaded.
    By default it's <literal>latest</literal>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-pullImage-3'>
    <para>
    <varname>imageId</varname>, if specified this exact image will be fetched, instead
    of <varname>imageName/imageTag</varname>. However, the resulting repository
    will still be named <varname>imageName/imageTag</varname>.
    By default it's <literal>null</literal>.
    </para>
  </callout>

  <callout arearefs='ex-dockerTools-pullImage-4'>
    <para>
    <varname>sha256</varname> is the checksum of the whole fetched image.
    This argument is required.
    </para>

    <note>
    <para>The checksum is computed on the unpacked directory, not on the final tarball.</para>
    </note>

  </callout>

  <callout arearefs='ex-dockerTools-pullImage-5'>
    <para>
    In the above example the default values are shown for the variables
    <varname>indexUrl</varname> and <varname>registryVersion</varname>.
    Hence by default the Docker.io registry is used to pull the images.
    </para>
  </callout>
  </calloutlist>

</section>

<section xml:id="ssec-pkgs-dockerTools-exportImage">
  <title>exportImage</title>

  <para>
  This function is analogous to the <command>docker export</command> command,
  in that can used to flatten a Docker image that contains multiple layers.
  It is in fact the result of the merge of all the layers of the image.
  As such, the result is suitable for being imported in Docker
  with <command>docker import</command>.
  </para>

  <note>
  <para>
    Using this function requires the <literal>kvm</literal>
    device to be available.
  </para>
  </note>

  <para>
  The parameters of <varname>exportImage</varname> are the following:
  </para>

  <example xml:id='ex-dockerTools-exportImage'><title>Docker export</title>
  <programlisting>
  exportImage {
    fromImage = someLayeredImage;
    fromImageName = null;
    fromImageTag = null;

    name = someLayeredImage.name;
  }
  </programlisting>
  </example>

  <para>
  The parameters relative to the base image have the same synopsis as
  described in <xref linkend='ssec-pkgs-dockerTools-buildImage'/>, except that
  <varname>fromImage</varname> is the only required argument in this case.
  </para>

  <para>
  The <varname>name</varname> argument is the name of the derivation output,
  which defaults to <varname>fromImage.name</varname>.
  </para>
</section>

<section xml:id="ssec-pkgs-dockerTools-shadowSetup">
  <title>shadowSetup</title>

  <para>
  This constant string is a helper for setting up the base files for managing
  users and groups, only if such files don't exist already.
  It is suitable for being used in a
  <varname>runAsRoot</varname> <xref linkend='ex-dockerTools-buildImage-runAsRoot'/> script for cases like
  in the example below:
  </para>

  <example xml:id='ex-dockerTools-shadowSetup'><title>Shadow base files</title>
  <programlisting>
  buildImage {
    name = "shadow-basic";

    runAsRoot = ''
      #!${stdenv.shell}
      ${shadowSetup}
      groupadd -r redis
      useradd -r -g redis redis
      mkdir /data
      chown redis:redis /data
    '';
  }
  </programlisting>
  </example>

  <para>
  Creating base files like <literal>/etc/passwd</literal> or
  <literal>/etc/login.defs</literal> are necessary for shadow-utils to
  manipulate users and groups.
  </para>

</section>

</section>

</chapter>