nixpkgs-suyu/doc/stdenv.xml

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

<title>The Standard Environment</title>


<para>The standard build environment in the Nix Packages collection
provides an environment for building Unix packages that does a lot of
common build tasks automatically.  In fact, for Unix packages that use
the standard <literal>./configure; make; make install</literal> build
interface, you don’t need to write a build script at all; the standard
environment does everything automatically.  If
<literal>stdenv</literal> doesn’t do what you need automatically, you
can easily customise or override the various build phases.</para>


<section xml:id="sec-using-stdenv"><title>Using
<literal>stdenv</literal></title>

<para>To build a package with the standard environment, you use the
function <varname>stdenv.mkDerivation</varname>, instead of the
primitive built-in function <varname>derivation</varname>, e.g.

<programlisting>
stdenv.mkDerivation {
  name = "libfoo-1.2.3";
  src = fetchurl {
    url = http://example.org/libfoo-1.2.3.tar.bz2;
    sha256 = "0x2g1jqygyr5wiwg4ma1nd7w4ydpy82z9gkcv8vh2v8dn3y58v5m";
  };
}</programlisting>

(<varname>stdenv</varname> needs to be in scope, so if you write this
in a separate Nix expression from
<filename>pkgs/all-packages.nix</filename>, you need to pass it as a
function argument.)  Specifying a <varname>name</varname> and a
<varname>src</varname> is the absolute minimum you need to do.  Many
packages have dependencies that are not provided in the standard
environment.  It’s usually sufficient to specify those dependencies in
the <varname>buildInputs</varname> attribute:

<programlisting>
stdenv.mkDerivation {
  name = "libfoo-1.2.3";
  ...
  buildInputs = [libbar perl ncurses];
}</programlisting>

This attribute ensures that the <filename>bin</filename>
subdirectories of these packages appear in the <envar>PATH</envar>
environment variable during the build, that their
<filename>include</filename> subdirectories are searched by the C
compiler, and so on.  (See <xref linkend="ssec-setup-hooks"/> for
details.)</para>

<para>Often it is necessary to override or modify some aspect of the
build.  To make this easier, the standard environment breaks the
package build into a number of <emphasis>phases</emphasis>, all of
which can be overridden or modified individually: unpacking the
sources, applying patches, configuring, building, and installing.
(There are some others; see <xref linkend="sec-stdenv-phases"/>.)
For instance, a package that doesn’t supply a makefile but instead has
to be compiled “manually” could be handled like this:

<programlisting>
stdenv.mkDerivation {
  name = "fnord-4.5";
  ...
  buildPhase = ''
    gcc foo.c -o foo
  '';
  installPhase = ''
    mkdir -p $out/bin
    cp foo $out/bin
  '';
}</programlisting>

(Note the use of <literal>''</literal>-style string literals, which
are very convenient for large multi-line script fragments because they
don’t need escaping of <literal>"</literal> and <literal>\</literal>,
and because indentation is intelligently removed.)</para>

<para>There are many other attributes to customise the build.  These
are listed in <xref linkend="ssec-stdenv-attributes"/>.</para>

<para>While the standard environment provides a generic builder, you
can still supply your own build script:

<programlisting>
stdenv.mkDerivation {
  name = "libfoo-1.2.3";
  ...
  builder = ./builder.sh;
}</programlisting>

where the builder can do anything it wants, but typically starts with

<programlisting>
source $stdenv/setup
</programlisting>

to let <literal>stdenv</literal> set up the environment (e.g., process
the <varname>buildInputs</varname>).  If you want, you can still use
<literal>stdenv</literal>’s generic builder:

<programlisting>
source $stdenv/setup

buildPhase() {
  echo "... this is my custom build phase ..."
  gcc foo.c -o foo
}

installPhase() {
  mkdir -p $out/bin
  cp foo $out/bin
}

genericBuild
</programlisting>

</para>

</section>


<section xml:id="sec-tools-of-stdenv"><title>Tools provided by
<literal>stdenv</literal></title>

<para>The standard environment provides the following packages:

<itemizedlist>

  <listitem><para>The GNU C Compiler, configured with C and C++
  support.</para></listitem>

  <listitem><para>GNU coreutils (contains a few dozen standard Unix
  commands).</para></listitem>

  <listitem><para>GNU findutils (contains
  <command>find</command>).</para></listitem>

  <listitem><para>GNU diffutils (contains <command>diff</command>,
  <command>cmp</command>).</para></listitem>

  <listitem><para>GNU <command>sed</command>.</para></listitem>

  <listitem><para>GNU <command>grep</command>.</para></listitem>

  <listitem><para>GNU <command>awk</command>.</para></listitem>

  <listitem><para>GNU <command>tar</command>.</para></listitem>

  <listitem><para><command>gzip</command>, <command>bzip2</command>
  and <command>xz</command>.</para></listitem>

  <listitem><para>GNU Make.  It has been patched to provide
  <quote>nested</quote> output that can be fed into the
  <command>nix-log2xml</command> command and
  <command>log2html</command> stylesheet to create a structured,
  readable output of the build steps performed by
  Make.</para></listitem>

  <listitem><para>Bash.  This is the shell used for all builders in
  the Nix Packages collection.  Not using <command>/bin/sh</command>
  removes a large source of portability problems.</para></listitem>

  <listitem><para>The <command>patch</command>
  command.</para></listitem>

</itemizedlist>

</para>

<para>On Linux, <literal>stdenv</literal> also includes the
<command>patchelf</command> utility.</para>

</section>


<section xml:id="ssec-stdenv-dependencies"><title>Specifying dependencies</title>

<para>
  As described in the Nix manual, almost any <filename>*.drv</filename> store path in a derivation's attribute set will induce a dependency on that derivation.
  <varname>mkDerivation</varname>, however, takes a few attributes intended to, between them, include all the dependencies of a package.
  This is done both for structure and consistency, but also so that certain other setup can take place.
  For example, certain dependencies need their bin directories added to the <envar>PATH</envar>.
  That is built-in, but other setup is done via a pluggable mechanism that works in conjunction with these dependency attributes.
  See <xref linkend="ssec-setup-hooks"/> for details.
</para>
<para>
  Dependencies can be broken down along three axes: their host and target platforms relative to the new derivation's, and whether they are propagated.
  The platform distinctions are motivated by cross compilation; see <xref linkend="chap-cross"/> for exactly what each platform means.
  <footnote><para>
    The build platform is ignored because it is a mere implementation detail of the package satisfying the dependency:
    As a general programming principle, dependencies are always <emphasis>specified</emphasis> as interfaces, not concrete implementation.
  </para></footnote>
  But even if one is not cross compiling, the platforms imply whether or not the dependency is needed at run-time or build-time, a concept that makes perfect sense outside of cross compilation.
  For now, the run-time/build-time distinction is just a hint for mental clarity, but in the future it perhaps could be enforced.
</para>
<para>
  The extension of <envar>PATH</envar> with dependencies, alluded to above, proceeds according to the relative platforms alone.
  The process is carried out only for dependencies whose host platform matches the new derivation's build platform–i.e. which run on the platform where the new derivation will be built.
  <footnote><para>
    Currently, that means for native builds all dependencies are put on the <envar>PATH</envar>.
    But in the future that may not be the case for sake of matching cross:
    the platforms would be assumed to be unique for native and cross builds alike, so only the <varname>depsBuild*</varname> and <varname>nativeBuildDependencies</varname> dependencies would affect the <envar>PATH</envar>.
  </para></footnote>
  For each dependency <replaceable>dep</replaceable> of those dependencies, <filename><replaceable>dep</replaceable>/bin</filename>, if present, is added to the <envar>PATH</envar> environment variable.
</para>
<para>
  The dependency is propagated when it forces some of its other-transitive (non-immediate) downstream dependencies to also take it on as an immediate dependency.
  Nix itself already takes a package's transitive dependencies into account, but this propagation ensures nixpkgs-specific infrastructure like setup hooks (mentioned above) also are run as if the propagated dependency.
</para>
<para>
  It is important to note dependencies are not necessary propagated as the same sort of dependency that they were before, but rather as the corresponding sort so that the platform rules still line up.
  The exact rules for dependency propagation can be given by assigning each sort of dependency two integers based one how it's host and target platforms are offset from the depending derivation's platforms.
  Those offsets are given are given below in the descriptions of each dependency list attribute.
  Algorithmically, we traverse propagated inputs, accumulating every propagated dep's propagated deps and adjusting them to account for the "shift in perspective" described by the current dep's platform offsets.
  This results in sort a transitive closure of the dependency relation, with the offsets being approximately summed when two dependency links are combined.
  We also prune transitive deps whose combined offsets go out-of-bounds, which can be viewed as a filter over that transitive closure removing dependencies that are blatantly absurd.
</para>
<para>
  We can define the process precisely with <link xlink:href="https://en.wikipedia.org/wiki/Natural_deduction">Natural Deduction</link> using the inference rules.
  This probably seems a bit obtuse, but so is the bash code that actually implements it!
  <footnote><para>
    The <function>findInputs</function> function, currently residing in <filename>pkgs/stdenv/generic/setup.sh</filename>, implements the propagation logic.
  </para></footnote>
  They're confusing in very different ways so...hopefully if something doesn't make sense in one presentation, it does in the other!
  <programlisting>
let mapOffset(h, t, i) = i + (if i &lt;= 0 then h else t - 1)

propagated-dep(h0, t0, A, B)
propagated-dep(h1, t1, B, C)
h0 + h1 in {-1, 0, 1}
h0 + t1 in {-1, 0, 1}
-------------------------------------- Transitive property
propagated-dep(mapOffset(h0, t0, h1),
               mapOffset(h0, t0, t1),
               A, C)</programlisting>
  <programlisting>
let mapOffset(h, t, i) = i + (if i &lt;= 0 then h else t - 1)

dep(h0, _, A, B)
propagated-dep(h1, t1, B, C)
h0 + h1 in {-1, 0, 1}
h0 + t1 in {-1, 0, -1}
-------------------------------------- Take immediate deps' propagated deps
propagated-dep(mapOffset(h0, t0, h1),
               mapOffset(h0, t0, t1),
               A, C)</programlisting>
  <programlisting>
propagated-dep(h, t, A, B)
-------------------------------------- Propagated deps count as deps
dep(h, t, A, B)</programlisting>
  Some explanation of this monstrosity is in order.
  In the common case, the target offset of a dependency is the successor to the target offset: <literal>t = h + 1</literal>.
  That means that:
  <programlisting>
let f(h, t, i) = i + (if i &lt;= 0 then h else t - 1)
let f(h, h + 1, i) = i + (if i &lt;= 0 then h else (h + 1) - 1)
let f(h, h + 1, i) = i + (if i &lt;= 0 then h else h)
let f(h, h + 1, i) = i + h
  </programlisting>
  This is where the "sum-like" comes from above:
  We can just sum all the host offset to get the host offset of the transitive dependency.
  The target offset is the transitive dep is simply the host offset + 1, just as it was with the dependencies composed to make this transitive one;
  it can be ignored as it doesn't add any new information.
</para>
<para>
  Because of the bounds checks, the uncommon cases are <literal>h = t</literal> and <literal>h + 2 = t</literal>.
  In the former case, the motivation for <function>mapOffset</function> is that since its host and target platforms are the same, no transitive dep of it should be able to "discover" an offset greater than its reduced target offsets.
  <function>mapOffset</function> effectively "squashes" all its transitive dependencies' offsets so that none will ever be greater than the target offset of the original <literal>h = t</literal> package.
  In the other case, <literal>h + 1</literal> is skipped over between the host and target offsets.
  Instead of squashing the offsets, we need to "rip" them apart so no transitive dependencies' offset is that one.
</para>
<para>
Overall, the unifying theme here is that propagation shouldn't be introducing transitive dependencies involving platforms the needing package is unaware of.
The offset bounds checking and definition of <function>mapOffset</function> together ensure that this is the case.
Discovering a new offset is discovering a new platform, and since those platforms weren't in the derivation "spec" of the needing package, they cannot be relevant.
From a capability perspective, we can imagine that the host and target platforms of a package are the capabilities a package requires, and the depending package must provide the capability to the dependency.
</para>

<variablelist>
  <title>Variables specifying dependencies</title>

   <varlistentry>
     <term><varname>depsBuildBuild</varname></term>
     <listitem>
       <para>
         A list of dependencies whose host and target platforms are the new derivation's build platform.
         This means a <literal>-1</literal> host and <literal>-1</literal> target offset from the new derivation's platforms.
         They are programs/libraries used at build time that furthermore produce programs/libraries also used at build time.
         If the dependency doesn't care about the target platform (i.e. isn't a compiler or similar tool), put it in <varname>nativeBuildInputs</varname>instead.
         The most common use for this <literal>buildPackages.stdenv.cc</literal>, the default C compiler for this role.
         That example crops up more than one might think in old commonly used C libraries.
       </para>
       <para>
         Since these packages are able to be run at build time, that are always added to the <envar>PATH</envar>, as described above.
         But since these packages are only guaranteed to be able to run then, they shouldn't persist as run-time dependencies.
         This isn't currently enforced, but could be in the future.
       </para>
     </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>nativeBuildInputs</varname></term>
    <listitem>
      <para>
        A list of dependencies whose host platform is the new derivation's build platform, and target platform is the new derivation's host platform.
        This means a <literal>-1</literal> host offset and <literal>0</literal> target offset from the new derivation's platforms.
        They are programs/libraries used at build time that, if they are a compiler or similar tool, produce code to run at run time—i.e. tools used to build the new derivation.
        If the dependency doesn't care about the target platform (i.e. isn't a compiler or similar tool), put it here, rather than in <varname>depsBuildBuild</varname> or <varname>depsBuildTarget</varname>.
        This would be called <varname>depsBuildHost</varname> but for historical continuity.
      </para>
      <para>
         Since these packages are able to be run at build time, that are added to the <envar>PATH</envar>, as described above.
         But since these packages only are guaranteed to be able to run then, they shouldn't persist as run-time dependencies.
         This isn't currently enforced, but could be in the future.
       </para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsBuildTarget</varname></term>
    <listitem>
      <para>
        A list of dependencies whose host platform is the new derivation's build platform, and target platform is the new derivation's target platform.
        This means a <literal>-1</literal> host offset and <literal>1</literal> target offset from the new derivation's platforms.
        They are programs used at build time that produce code to run at run with code produced by the depending package.
        Most commonly, these would tools used to build the runtime or standard library the currently-being-built compiler will inject into any code it compiles.
        In many cases, the currently-being built compiler is itself employed for that task, but when that compiler won't run (i.e. its build and host platform differ) this is not possible.
        Other times, the compiler relies on some other tool, like binutils, that is always built separately so the dependency is unconditional.
      </para>
      <para>
        This is a somewhat confusing dependency to wrap ones head around, and for good reason.
        As the only one where the platform offsets are not adjacent integers, it requires thinking of a bootstrapping stage <emphasis>two</emphasis> away from the current one.
        It and it's use-case go hand in hand and are both considered poor form:
        try not to need this sort dependency, and try not avoid building standard libraries / runtimes in the same derivation as the compiler produces code using them.
        Instead strive to build those like a normal library, using the newly-built compiler just as a normal library would.
        In short, do not use this attribute unless you are packaging a compiler and are sure it is needed.
     </para>
     <para>
       Since these packages are able to be run at build time, that are added to the <envar>PATH</envar>, as described above.
       But since these packages only are guaranteed to be able to run then, they shouldn't persist as run-time dependencies.
       This isn't currently enforced, but could be in the future.
     </para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsHostHost</varname></term>
    <listitem><para>
      A list of dependencies whose host and target platforms match the new derivation's host platform.
      This means a both <literal>0</literal> host offset and <literal>0</literal> target offset from the new derivation's host platform.
      These are packages used at run-time to generate code also used at run-time.
      In practice, that would usually be tools used by compilers for metaprogramming/macro systems, or libraries used by the macros/metaprogramming code itself.
      It's always preferable to use a <varname>depsBuildBuild</varname> dependency in the derivation being built than a <varname>depsHostHost</varname> on the tool doing the building for this purpose.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>buildInputs</varname></term>
    <listitem>
      <para>
        A list of dependencies whose host platform and target platform match the new derivation's.
        This means a <literal>0</literal> host offset and <literal>1</literal> target offset from the new derivation's host platform.
        This would be called <varname>depsHostTarget</varname> but for historical continuity.
        If the dependency doesn't care about the target platform (i.e. isn't a compiler or similar tool), put it here, rather than in <varname>depsBuildBuild</varname>.
      </para>
      <para>
        These often are programs/libraries used by the new derivation at <emphasis>run</emphasis>-time, but that isn't always the case.
        For example, the machine code in a statically linked library is only used at run time, but the derivation containing the library is only needed at build time.
        Even in the dynamic case, the library may also be needed at build time to appease the linker.
      </para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsTargetTarget</varname></term>
    <listitem><para>
      A list of dependencies whose host platform matches the new derivation's target platform.
      This means a <literal>1</literal> offset from the new derivation's platforms.
      These are packages that run on the target platform, e.g. the standard library or run-time deps of standard library that a compiler insists on knowing about.
      It's poor form in almost all cases for a package to depend on another from a future stage [future stage corresponding to positive offset].
      Do not use this attribute unless you are packaging a compiler and are sure it is needed.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsBuildBuildPropagated</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>depsBuildBuild</varname>.
      This perhaps never ought to be used, but it is included for consistency [see below for the others].
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>propagatedNativeBuildInputs</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>nativeBuildInputs</varname>.
      This would be called <varname>depsBuildHostPropagated</varname> but for historical continuity.
      For example, if package <varname>Y</varname> has <literal>propagatedNativeBuildInputs = [X]</literal>, and package <varname>Z</varname> has <literal>buildInputs = [Y]</literal>, then package <varname>Z</varname> will be built as if it included package <varname>X</varname> in its <varname>nativeBuildInputs</varname>.
      If instead, package <varname>Z</varname> has <literal>nativeBuildInputs = [Y]</literal>, then <varname>Z</varname> will be built as if it included <varname>X</varname> in the <varname>depsBuildBuild</varname> of package <varname>Z</varname>, because of the sum of the two <literal>-1</literal> host offsets.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsBuildTargetPropagated</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>depsBuildTarget</varname>.
      This is prefixed for the same reason of alerting potential users.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsHostHostPropagated</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>depsHostHost</varname>.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>propagatedBuildInputs</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>buildInputs</varname>.
      This would be called <varname>depsHostTargetPropagated</varname> but for historical continuity.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>depsTargetTarget</varname></term>
    <listitem><para>
      The propagated equivalent of <varname>depsTargetTarget</varname>.
      This is prefixed for the same reason of alerting potential users.
    </para></listitem>
  </varlistentry>

</variablelist>

</section>


<section xml:id="ssec-stdenv-attributes"><title>Attributes</title>

<variablelist>
  <title>Variables affecting <literal>stdenv</literal>
  initialisation</title>

  <varlistentry>
    <term><varname>NIX_DEBUG</varname></term>
    <listitem><para>
      A natural number indicating how much information to log.
      If set to 1 or higher, <literal>stdenv</literal> will print moderate debug information during the build.
      In particular, the <command>gcc</command> and <command>ld</command> wrapper scripts will print out the complete command line passed to the wrapped tools.
      If set to 6 or higher, the <literal>stdenv</literal> setup script will be run with <literal>set -x</literal> tracing.
      If set to 7 or higher, the <command>gcc</command> and <command>ld</command> wrapper scripts will also be run with <literal>set -x</literal> tracing.
    </para></listitem>
  </varlistentry>

</variablelist>

<variablelist>
  <title>Variables affecting build properties</title>

  <varlistentry>
    <term><varname>enableParallelBuilding</varname></term>
    <listitem>
      <para>If set to <literal>true</literal>, <literal>stdenv</literal> will
      pass specific flags to <literal>make</literal> and other build tools to
      enable parallel building with up to <literal>build-cores</literal>
      workers.</para>

      <para>Unless set to <literal>false</literal>, some build systems with good
      support for parallel building including <literal>cmake</literal>,
      <literal>meson</literal>, and <literal>qmake</literal> will set it to
      <literal>true</literal>.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preferLocalBuild</varname></term>
    <listitem><para>If set, specifies that the package is so lightweight
    in terms of build operations (e.g. write a text file from a Nix string
    to the store) that there's no need to look for it in binary caches --
    it's faster to just build it locally. It also tells Hydra and other
    facilities that this package doesn't need to be exported in binary
    caches (noone would use it, after all).</para></listitem>
  </varlistentry>

</variablelist>

<variablelist>
  <title>Special variables</title>

  <varlistentry>
    <term><varname>passthru</varname></term>
    <listitem><para>This is an attribute set which can be filled with arbitrary
    values. For example:

<programlisting>
passthru = {
  foo = "bar";
  baz = {
    value1 = 4;
    value2 = 5;
  };
}
</programlisting>

</para>

    <para>Values inside it are not passed to the builder, so you can change
    them without triggering a rebuild. However, they can be accessed outside of a
    derivation directly, as if they were set inside a derivation itself, e.g.
    <literal>hello.baz.value1</literal>. We don't specify any usage or
    schema of <literal>passthru</literal> - it is meant for values that would be
    useful outside the derivation in other parts of a Nix expression (e.g. in other
    derivations). An example would be to convey some specific dependency of your
    derivation which contains a program with plugins support.  Later, others who
    make derivations with plugins can use passed-through dependency to ensure that
    their plugin would be binary-compatible with built program.</para></listitem>
  </varlistentry>

</variablelist>

</section>


<section xml:id="sec-stdenv-phases"><title>Phases</title>

<para>The generic builder has a number of <emphasis>phases</emphasis>.
Package builds are split into phases to make it easier to override
specific parts of the build (e.g., unpacking the sources or installing
the binaries).  Furthermore, it allows a nicer presentation of build
logs in the Nix build farm.</para>

<para>Each phase can be overridden in its entirety either by setting
the environment variable
<varname><replaceable>name</replaceable>Phase</varname> to a string
containing some shell commands to be executed, or by redefining the
shell function
<varname><replaceable>name</replaceable>Phase</varname>.  The former
is convenient to override a phase from the derivation, while the
latter is convenient from a build script.

However, typically one only wants to <emphasis>add</emphasis> some
commands to a phase, e.g. by defining <literal>postInstall</literal>
or <literal>preFixup</literal>, as skipping some of the default actions
may have unexpected consequences.
</para>


<section xml:id="ssec-controlling-phases"><title>Controlling
phases</title>

<para>There are a number of variables that control what phases are
executed and in what order:

<variablelist>
  <title>Variables affecting phase control</title>

  <varlistentry>
    <term><varname>phases</varname></term>
    <listitem>
      <para>Specifies the phases.  You can change the order in which
      phases are executed, or add new phases, by setting this
      variable.  If it’s not set, the default value is used, which is
      <literal>$prePhases unpackPhase patchPhase $preConfigurePhases
      configurePhase $preBuildPhases buildPhase checkPhase
      $preInstallPhases installPhase fixupPhase $preDistPhases
      distPhase $postPhases</literal>.
      </para>

      <para>Usually, if you just want to add a few phases, it’s more
      convenient to set one of the variables below (such as
      <varname>preInstallPhases</varname>), as you then don’t specify
      all the normal phases.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>prePhases</varname></term>
    <listitem>
      <para>Additional phases executed before any of the default phases.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preConfigurePhases</varname></term>
    <listitem>
      <para>Additional phases executed just before the configure phase.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preBuildPhases</varname></term>
    <listitem>
      <para>Additional phases executed just before the build phase.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preInstallPhases</varname></term>
    <listitem>
      <para>Additional phases executed just before the install phase.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preFixupPhases</varname></term>
    <listitem>
      <para>Additional phases executed just before the fixup phase.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preDistPhases</varname></term>
    <listitem>
      <para>Additional phases executed just before the distribution phase.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postPhases</varname></term>
    <listitem>
      <para>Additional phases executed after any of the default
      phases.</para>
    </listitem>
  </varlistentry>

</variablelist>

</para>

</section>


<section xml:id="ssec-unpack-phase"><title>The unpack phase</title>

<para>The unpack phase is responsible for unpacking the source code of
the package.  The default implementation of
<function>unpackPhase</function> unpacks the source files listed in
the <envar>src</envar> environment variable to the current directory.
It supports the following files by default:

<variablelist>

  <varlistentry>
    <term>Tar files</term>
    <listitem><para>These can optionally be compressed using
    <command>gzip</command> (<filename>.tar.gz</filename>,
    <filename>.tgz</filename> or <filename>.tar.Z</filename>),
    <command>bzip2</command> (<filename>.tar.bz2</filename> or
    <filename>.tbz2</filename>) or <command>xz</command>
    (<filename>.tar.xz</filename> or
    <filename>.tar.lzma</filename>).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Zip files</term>
    <listitem><para>Zip files are unpacked using
    <command>unzip</command>.  However, <command>unzip</command> is
    not in the standard environment, so you should add it to
    <varname>buildInputs</varname> yourself.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Directories in the Nix store</term>
    <listitem><para>These are simply copied to the current directory.
    The hash part of the file name is stripped,
    e.g. <filename>/nix/store/1wydxgby13cz...-my-sources</filename>
    would be copied to
    <filename>my-sources</filename>.</para></listitem>
  </varlistentry>

</variablelist>

Additional file types can be supported by setting the
<varname>unpackCmd</varname> variable (see below).</para>

<para></para>

<variablelist>
  <title>Variables controlling the unpack phase</title>

  <varlistentry>
    <term><varname>srcs</varname> / <varname>src</varname></term>
    <listitem><para>The list of source files or directories to be
    unpacked or copied.  One of these must be set.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>sourceRoot</varname></term>
    <listitem><para>After running <function>unpackPhase</function>,
    the generic builder changes the current directory to the directory
    created by unpacking the sources.  If there are multiple source
    directories, you should set <varname>sourceRoot</varname> to the
    name of the intended directory.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>setSourceRoot</varname></term>
    <listitem><para>Alternatively to setting
    <varname>sourceRoot</varname>, you can set
    <varname>setSourceRoot</varname> to a shell command to be
    evaluated by the unpack phase after the sources have been
    unpacked.  This command must set
    <varname>sourceRoot</varname>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preUnpack</varname></term>
    <listitem><para>Hook executed at the start of the unpack
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postUnpack</varname></term>
    <listitem><para>Hook executed at the end of the unpack
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontMakeSourcesWritable</varname></term>
    <listitem><para>If set to <literal>1</literal>, the unpacked
    sources are <emphasis>not</emphasis> made
    writable.  By default, they are made writable to prevent problems
    with read-only sources.  For example, copied store directories
    would be read-only without this.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>unpackCmd</varname></term>
    <listitem><para>The unpack phase evaluates the string
    <literal>$unpackCmd</literal> for any unrecognised file.  The path
    to the current source file is contained in the
    <varname>curSrc</varname> variable.</para></listitem>
  </varlistentry>

</variablelist>

</section>


<section xml:id="ssec-patch-phase"><title>The patch phase</title>

<para>The patch phase applies the list of patches defined in the
<varname>patches</varname> variable.</para>

<variablelist>
  <title>Variables controlling the patch phase</title>

  <varlistentry>
    <term><varname>patches</varname></term>
    <listitem><para>The list of patches.  They must be in the format
    accepted by the <command>patch</command> command, and may
    optionally be compressed using <command>gzip</command>
    (<filename>.gz</filename>), <command>bzip2</command>
    (<filename>.bz2</filename>) or <command>xz</command>
    (<filename>.xz</filename>).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>patchFlags</varname></term>
    <listitem><para>Flags to be passed to <command>patch</command>.
    If not set, the argument <option>-p1</option> is used, which
    causes the leading directory component to be stripped from the
    file names in each patch.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>prePatch</varname></term>
    <listitem><para>Hook executed at the start of the patch
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postPatch</varname></term>
    <listitem><para>Hook executed at the end of the patch
    phase.</para></listitem>
  </varlistentry>

</variablelist>

</section>


<section xml:id="ssec-configure-phase"><title>The configure phase</title>

<para>The configure phase prepares the source tree for building.  The
default <function>configurePhase</function> runs
<filename>./configure</filename> (typically an Autoconf-generated
script) if it exists.</para>

<variablelist>
  <title>Variables controlling the configure phase</title>

  <varlistentry>
    <term><varname>configureScript</varname></term>
    <listitem><para>The name of the configure script.  It defaults to
    <filename>./configure</filename> if it exists; otherwise, the
    configure phase is skipped.  This can actually be a command (like
    <literal>perl ./Configure.pl</literal>).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>configureFlags</varname></term>
    <listitem><para>A list of strings passed as additional arguments to the
    configure script.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>configureFlagsArray</varname></term>
    <listitem><para>A shell array containing additional arguments
    passed to the configure script.  You must use this instead of
    <varname>configureFlags</varname> if the arguments contain
    spaces.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontAddPrefix</varname></term>
    <listitem><para>By default, the flag
    <literal>--prefix=$prefix</literal> is added to the configure
    flags.  If this is undesirable, set this variable to
    true.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>prefix</varname></term>
    <listitem><para>The prefix under which the package must be
    installed, passed via the <option>--prefix</option> option to the
    configure script.  It defaults to
    <option>$out</option>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontAddDisableDepTrack</varname></term>
    <listitem><para>By default, the flag
    <literal>--disable-dependency-tracking</literal> is added to the
    configure flags to speed up Automake-based builds.  If this is
    undesirable, set this variable to true.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontFixLibtool</varname></term>
    <listitem><para>By default, the configure phase applies some
    special hackery to all files called <filename>ltmain.sh</filename>
    before running the configure script in order to improve the purity
    of Libtool-based packages<footnote><para>It clears the
    <varname>sys_lib_<replaceable>*</replaceable>search_path</varname>
    variables in the Libtool script to prevent Libtool from using
    libraries in <filename>/usr/lib</filename> and
    such.</para></footnote>.  If this is undesirable, set this
    variable to true.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontDisableStatic</varname></term>
    <listitem><para>By default, when the configure script has
    <option>--enable-static</option>, the option
    <option>--disable-static</option> is added to the configure flags.</para>
    <para>If this is undesirable, set this variable to
    true.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>configurePlatforms</varname></term>
    <listitem><para>
      By default, when cross compiling, the configure script has <option>--build=...</option> and <option>--host=...</option> passed.
      Packages can instead pass <literal>[ "build" "host" "target" ]</literal> or a subset to control exactly which platform flags are passed.
      Compilers and other tools should use this to also pass the target platform, for example.
      Note eventually these will be passed when in native builds too, to improve determinism: build-time guessing, as is done today, is a risk of impurity.
    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preConfigure</varname></term>
    <listitem><para>Hook executed at the start of the configure
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postConfigure</varname></term>
    <listitem><para>Hook executed at the end of the configure
    phase.</para></listitem>
  </varlistentry>

</variablelist>


</section>


<section xml:id="build-phase"><title>The build phase</title>

<para>The build phase is responsible for actually building the package
(e.g. compiling it).  The default <function>buildPhase</function>
simply calls <command>make</command> if a file named
<filename>Makefile</filename>, <filename>makefile</filename> or
<filename>GNUmakefile</filename> exists in the current directory (or
the <varname>makefile</varname> is explicitly set); otherwise it does
nothing.</para>

<variablelist>
  <title>Variables controlling the build phase</title>

  <varlistentry>
    <term><varname>dontBuild</varname></term>
    <listitem><para>Set to true to skip the build phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>makefile</varname></term>
    <listitem><para>The file name of the Makefile.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>makeFlags</varname></term>
    <listitem><para>A list of strings passed as additional flags to
    <command>make</command>.  These flags are also used by the default
    install and check phase.  For setting make flags specific to the
    build phase, use <varname>buildFlags</varname> (see below).

<programlisting>
makeFlags = [ "PREFIX=$(out)" ];
</programlisting>

    <note><para>The flags are quoted in bash, but environment variables can
    be specified by using the make syntax.</para></note></para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>makeFlagsArray</varname></term>
    <listitem><para>A shell array containing additional arguments
    passed to <command>make</command>.  You must use this instead of
    <varname>makeFlags</varname> if the arguments contain
    spaces, e.g.

<programlisting>
makeFlagsArray=(CFLAGS="-O0 -g" LDFLAGS="-lfoo -lbar")
</programlisting>

    Note that shell arrays cannot be passed through environment
    variables, so you cannot set <varname>makeFlagsArray</varname> in
    a derivation attribute (because those are passed through
    environment variables): you have to define them in shell
    code.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>buildFlags</varname> / <varname>buildFlagsArray</varname></term>
    <listitem><para>A list of strings passed as additional flags to
    <command>make</command>.  Like <varname>makeFlags</varname> and
    <varname>makeFlagsArray</varname>, but only used by the build
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preBuild</varname></term>
    <listitem><para>Hook executed at the start of the build
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postBuild</varname></term>
    <listitem><para>Hook executed at the end of the build
    phase.</para></listitem>
  </varlistentry>

</variablelist>


<para>
You can set flags for <command>make</command> through the
<varname>makeFlags</varname> variable.</para>

<para>Before and after running <command>make</command>, the hooks
<varname>preBuild</varname> and <varname>postBuild</varname> are
called, respectively.</para>

</section>


<section xml:id="ssec-check-phase"><title>The check phase</title>

<para>The check phase checks whether the package was built correctly
by running its test suite.  The default
<function>checkPhase</function> calls <command>make check</command>,
but only if the <varname>doCheck</varname> variable is enabled.</para>

<variablelist>
  <title>Variables controlling the check phase</title>

  <varlistentry>
    <term><varname>doCheck</varname></term>
    <listitem><para>If set to a non-empty string, the check phase is
    executed, otherwise it is skipped (default).  Thus you should set

    <programlisting>
doCheck = true;</programlisting>

    in the derivation to enable checks.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>makeFlags</varname> /
    <varname>makeFlagsArray</varname> /
    <varname>makefile</varname></term>
    <listitem><para>See the build phase for details.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>checkTarget</varname></term>
    <listitem><para>The make target that runs the tests.  Defaults to
    <literal>check</literal>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>checkFlags</varname> / <varname>checkFlagsArray</varname></term>
    <listitem><para>A list of strings passed as additional flags to
    <command>make</command>.  Like <varname>makeFlags</varname> and
    <varname>makeFlagsArray</varname>, but only used by the check
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preCheck</varname></term>
    <listitem><para>Hook executed at the start of the check
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postCheck</varname></term>
    <listitem><para>Hook executed at the end of the check
    phase.</para></listitem>
  </varlistentry>

</variablelist>


</section>


<section xml:id="ssec-install-phase"><title>The install phase</title>

<para>The install phase is responsible for installing the package in
the Nix store under <envar>out</envar>.  The default
<function>installPhase</function> creates the directory
<literal>$out</literal> and calls <command>make
install</command>.</para>

<variablelist>
  <title>Variables controlling the install phase</title>

  <varlistentry>
    <term><varname>makeFlags</varname> /
    <varname>makeFlagsArray</varname> /
    <varname>makefile</varname></term>
    <listitem><para>See the build phase for details.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>installTargets</varname></term>
    <listitem><para>The make targets that perform the installation.
    Defaults to <literal>install</literal>.  Example:

<programlisting>
installTargets = "install-bin install-doc";</programlisting>

    </para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>installFlags</varname> / <varname>installFlagsArray</varname></term>
    <listitem><para>A list of strings passed as additional flags to
    <command>make</command>.  Like <varname>makeFlags</varname> and
    <varname>makeFlagsArray</varname>, but only used by the install
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preInstall</varname></term>
    <listitem><para>Hook executed at the start of the install
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postInstall</varname></term>
    <listitem><para>Hook executed at the end of the install
    phase.</para></listitem>
  </varlistentry>

</variablelist>


</section>


<section xml:id="ssec-fixup-phase"><title>The fixup phase</title>

<para>The fixup phase performs some (Nix-specific) post-processing
actions on the files installed under <filename>$out</filename> by the
install phase.  The default <function>fixupPhase</function> does the
following:

<itemizedlist>

  <listitem><para>It moves the <filename>man/</filename>,
  <filename>doc/</filename> and <filename>info/</filename>
  subdirectories of <envar>$out</envar> to
  <filename>share/</filename>.</para></listitem>

  <listitem><para>It strips libraries and executables of debug
  information.</para></listitem>

  <listitem><para>On Linux, it applies the <command>patchelf</command>
  command to ELF executables and libraries to remove unused
  directories from the <literal>RPATH</literal> in order to prevent
  unnecessary runtime dependencies.</para></listitem>

  <listitem><para>It rewrites the interpreter paths of shell scripts
  to paths found in <envar>PATH</envar>.  E.g.,
  <filename>/usr/bin/perl</filename> will be rewritten to
  <filename>/nix/store/<replaceable>some-perl</replaceable>/bin/perl</filename>
  found in <envar>PATH</envar>.</para></listitem>

</itemizedlist>

</para>

<variablelist>
  <title>Variables controlling the fixup phase</title>

  <varlistentry>
    <term><varname>dontStrip</varname></term>
    <listitem><para>If set, libraries and executables are not
    stripped.  By default, they are.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontMoveSbin</varname></term>
    <listitem><para>If set, files in <filename>$out/sbin</filename> are not moved
    to <filename>$out/bin</filename>. By default, they are.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>stripAllList</varname></term>
    <listitem><para>List of directories to search for libraries and
    executables from which <emphasis>all</emphasis> symbols should be
    stripped.  By default, it’s empty.  Stripping all symbols is
    risky, since it may remove not just debug symbols but also ELF
    information necessary for normal execution.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>stripAllFlags</varname></term>
    <listitem><para>Flags passed to the <command>strip</command>
    command applied to the files in the directories listed in
    <varname>stripAllList</varname>.  Defaults to <option>-s</option>
    (i.e. <option>--strip-all</option>).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>stripDebugList</varname></term>
    <listitem><para>List of directories to search for libraries and
    executables from which only debugging-related symbols should be
    stripped.  It defaults to <literal>lib bin
    sbin</literal>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>stripDebugFlags</varname></term>
    <listitem><para>Flags passed to the <command>strip</command>
    command applied to the files in the directories listed in
    <varname>stripDebugList</varname>.  Defaults to
    <option>-S</option>
    (i.e. <option>--strip-debug</option>).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontPatchELF</varname></term>
    <listitem><para>If set, the <command>patchelf</command> command is
    not used to remove unnecessary <literal>RPATH</literal> entries.
    Only applies to Linux.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontPatchShebangs</varname></term>
    <listitem><para>If set, scripts starting with
    <literal>#!</literal> do not have their interpreter paths
    rewritten to paths in the Nix store.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>forceShare</varname></term>
    <listitem><para>The list of directories that must be moved from
    <filename>$out</filename> to <filename>$out/share</filename>.
    Defaults to <literal>man doc info</literal>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>setupHook</varname></term>
    <listitem><para>A package can export a <link
    linkend="ssec-setup-hooks">setup hook</link> by setting this
    variable.  The setup hook, if defined, is copied to
    <filename>$out/nix-support/setup-hook</filename>.  Environment
    variables are then substituted in it using <function
    linkend="fun-substituteAll">substituteAll</function>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preFixup</varname></term>
    <listitem><para>Hook executed at the start of the fixup
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postFixup</varname></term>
    <listitem><para>Hook executed at the end of the fixup
    phase.</para></listitem>
  </varlistentry>

  <varlistentry xml:id="stdenv-separateDebugInfo">
    <term><varname>separateDebugInfo</varname></term>
    <listitem><para>If set to <literal>true</literal>, the standard
    environment will enable debug information in C/C++ builds. After
    installation, the debug information will be separated from the
    executables and stored in the output named
    <literal>debug</literal>. (This output is enabled automatically;
    you don’t need to set the <varname>outputs</varname> attribute
    explicitly.) To be precise, the debug information is stored in
    <filename><replaceable>debug</replaceable>/lib/debug/.build-id/<replaceable>XX</replaceable>/<replaceable>YYYY…</replaceable></filename>,
    where <replaceable>XXYYYY…</replaceable> is the <replaceable>build
    ID</replaceable> of the binary — a SHA-1 hash of the contents of
    the binary. Debuggers like GDB use the build ID to look up the
    separated debug information.</para>

    <para>For example, with GDB, you can add

<programlisting>
set debug-file-directory ~/.nix-profile/lib/debug
</programlisting>

    to <filename>~/.gdbinit</filename>. GDB will then be able to find
    debug information installed via <literal>nix-env
    -i</literal>.</para>

    </listitem>
  </varlistentry>

</variablelist>

</section>

<section xml:id="ssec-installCheck-phase"><title>The installCheck phase</title>

<para>The installCheck phase checks whether the package was installed
correctly by running its test suite against the installed directories.
The default <function>installCheck</function> calls <command>make
installcheck</command>.</para>

<variablelist>
  <title>Variables controlling the installCheck phase</title>

  <varlistentry>
    <term><varname>doInstallCheck</varname></term>
    <listitem><para>If set to a non-empty string, the installCheck phase is
    executed, otherwise it is skipped (default).  Thus you should set

    <programlisting>doInstallCheck = true;</programlisting>

    in the derivation to enable install checks.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preInstallCheck</varname></term>
    <listitem><para>Hook executed at the start of the installCheck
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postInstallCheck</varname></term>
    <listitem><para>Hook executed at the end of the installCheck
    phase.</para></listitem>
  </varlistentry>

</variablelist>

</section>

<section xml:id="ssec-distribution-phase"><title>The distribution
phase</title>

<para>The distribution phase is intended to produce a source
distribution of the package.  The default
<function>distPhase</function> first calls <command>make
dist</command>, then it copies the resulting source tarballs to
<filename>$out/tarballs/</filename>.  This phase is only executed if
the attribute <varname>doDist</varname> is set.</para>

<variablelist>
  <title>Variables controlling the distribution phase</title>

  <varlistentry>
    <term><varname>distTarget</varname></term>
    <listitem><para>The make target that produces the distribution.
    Defaults to <literal>dist</literal>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>distFlags</varname> / <varname>distFlagsArray</varname></term>
    <listitem><para>Additional flags passed to
    <command>make</command>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>tarballs</varname></term>
    <listitem><para>The names of the source distribution files to be
    copied to <filename>$out/tarballs/</filename>.  It can contain
    shell wildcards.  The default is
    <filename>*.tar.gz</filename>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>dontCopyDist</varname></term>
    <listitem><para>If set, no files are copied to
    <filename>$out/tarballs/</filename>.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>preDist</varname></term>
    <listitem><para>Hook executed at the start of the distribution
    phase.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>postDist</varname></term>
    <listitem><para>Hook executed at the end of the distribution
    phase.</para></listitem>
  </varlistentry>

</variablelist>


</section>


</section>


<section xml:id="ssec-stdenv-functions"><title>Shell functions</title>

<para>The standard environment provides a number of useful
functions.</para>

<variablelist>


  <varlistentry xml:id='fun-makeWrapper'>
    <term><function>makeWrapper</function>
    <replaceable>executable</replaceable>
    <replaceable>wrapperfile</replaceable>
    <replaceable>args</replaceable></term>
    <listitem><para>Constructs a wrapper for a program with various
    possible arguments. For example:

<programlisting>
# adds `FOOBAR=baz` to `$out/bin/foo`’s environment
makeWrapper $out/bin/foo $wrapperfile --set FOOBAR baz

# prefixes the binary paths of `hello` and `git`
# Be advised that paths often should be patched in directly
# (via string replacements or in `configurePhase`).
makeWrapper $out/bin/foo $wrapperfile --prefix PATH : ${lib.makeBinPath [ hello git ]}
</programlisting>

    There’s many more kinds of arguments, they are documented in
    <literal>nixpkgs/pkgs/build-support/setup-hooks/make-wrapper.sh</literal>.</para>

    <para><literal>wrapProgram</literal> is a convenience function you probably
    want to use most of the time.</para>

    </listitem>
  </varlistentry>


  <varlistentry xml:id='fun-substitute'>
    <term><function>substitute</function>
    <replaceable>infile</replaceable>
    <replaceable>outfile</replaceable>
    <replaceable>subs</replaceable></term>

    <listitem>
      <para>Performs string substitution on the contents of
      <replaceable>infile</replaceable>, writing the result to
      <replaceable>outfile</replaceable>.  The substitutions in
      <replaceable>subs</replaceable> are of the following form:

        <variablelist>
          <varlistentry>
            <term><option>--replace</option>
            <replaceable>s1</replaceable>
            <replaceable>s2</replaceable></term>
            <listitem><para>Replace every occurrence of the string
            <replaceable>s1</replaceable> by
            <replaceable>s2</replaceable>.</para></listitem>
          </varlistentry>

          <varlistentry>
            <term><option>--subst-var</option>
            <replaceable>varName</replaceable></term>
            <listitem><para>Replace every occurrence of
            <literal>@<replaceable>varName</replaceable>@</literal> by
            the contents of the environment variable
            <replaceable>varName</replaceable>.  This is useful for
            generating files from templates, using
            <literal>@<replaceable>...</replaceable>@</literal> in the
            template as placeholders.</para></listitem>
          </varlistentry>

          <varlistentry>
            <term><option>--subst-var-by</option>
            <replaceable>varName</replaceable>
            <replaceable>s</replaceable></term>
            <listitem><para>Replace every occurrence of
            <literal>@<replaceable>varName</replaceable>@</literal> by
            the string <replaceable>s</replaceable>.</para></listitem>
          </varlistentry>

        </variablelist>

      </para>

      <para>Example:

<programlisting>
substitute ./foo.in ./foo.out \
    --replace /usr/bin/bar $bar/bin/bar \
    --replace "a string containing spaces" "some other text" \
    --subst-var someVar
</programlisting>

      </para>

      <para><function>substitute</function> is implemented using the
      <command
      xlink:href="http://replace.richardlloyd.org.uk/">replace</command>
      command.  Unlike with the <command>sed</command> command, you
      don’t have to worry about escaping special characters.  It
      supports performing substitutions on binary files (such as
      executables), though there you’ll probably want to make sure
      that the replacement string is as long as the replaced
      string.</para>

    </listitem>
  </varlistentry>


  <varlistentry xml:id='fun-substituteInPlace'>
    <term><function>substituteInPlace</function>
    <replaceable>file</replaceable>
    <replaceable>subs</replaceable></term>
    <listitem><para>Like <function>substitute</function>, but performs
    the substitutions in place on the file
    <replaceable>file</replaceable>.</para></listitem>
  </varlistentry>


  <varlistentry xml:id='fun-substituteAll'>
    <term><function>substituteAll</function>
    <replaceable>infile</replaceable>
    <replaceable>outfile</replaceable></term>
    <listitem><para>Replaces every occurrence of
    <literal>@<replaceable>varName</replaceable>@</literal>, where
    <replaceable>varName</replaceable> is any environment variable, in
    <replaceable>infile</replaceable>, writing the result to
    <replaceable>outfile</replaceable>.  For instance, if
    <replaceable>infile</replaceable> has the contents

<programlisting>
#! @bash@/bin/sh
PATH=@coreutils@/bin
echo @foo@
</programlisting>

    and the environment contains
    <literal>bash=/nix/store/bmwp0q28cf21...-bash-3.2-p39</literal>
    and
    <literal>coreutils=/nix/store/68afga4khv0w...-coreutils-6.12</literal>,
    but does not contain the variable <varname>foo</varname>, then the
    output will be

<programlisting>
#! /nix/store/bmwp0q28cf21...-bash-3.2-p39/bin/sh
PATH=/nix/store/68afga4khv0w...-coreutils-6.12/bin
echo @foo@
</programlisting>

    That is, no substitution is performed for undefined variables.</para>

    <para>Environment variables that start with an uppercase letter or an
    underscore are filtered out,
    to prevent global variables (like <literal>HOME</literal>) or private
    variables (like <literal>__ETC_PROFILE_DONE</literal>) from accidentally
    getting substituted.
    The variables also have to be valid bash “names”, as
    defined in the bash manpage (alphanumeric or <literal>_</literal>,
    must not start with a number).</para>
  </listitem>
  </varlistentry>


  <varlistentry xml:id='fun-substituteAllInPlace'>
    <term><function>substituteAllInPlace</function>
    <replaceable>file</replaceable></term>
    <listitem><para>Like <function>substituteAll</function>, but performs
    the substitutions in place on the file
    <replaceable>file</replaceable>.</para></listitem>
  </varlistentry>


  <varlistentry xml:id='fun-stripHash'>
    <term><function>stripHash</function>
    <replaceable>path</replaceable></term>
    <listitem><para>Strips the directory and hash part of a store
    path, outputting the name part to <literal>stdout</literal>.
    For example:

<programlisting>
# prints coreutils-8.24
stripHash "/nix/store/9s9r019176g7cvn2nvcw41gsp862y6b4-coreutils-8.24"
</programlisting>

    If you wish to store the result in another variable, then the
    following idiom may be useful:

<programlisting>
name="/nix/store/9s9r019176g7cvn2nvcw41gsp862y6b4-coreutils-8.24"
someVar=$(stripHash $name)
</programlisting>

    </para></listitem>
  </varlistentry>


  <varlistentry xml:id='fun-wrapProgram'>
    <term><function>wrapProgram</function>
    <replaceable>executable</replaceable>
    <replaceable>makeWrapperArgs</replaceable></term>
    <listitem><para>Convenience function for <literal>makeWrapper</literal>
    that automatically creates a sane wrapper file

    It takes all the same arguments as <literal>makeWrapper</literal>,
    except for <literal>--argv0</literal>.</para>

    <para>It cannot be applied multiple times, since it will overwrite the wrapper
    file.</para>
    </listitem>
  </varlistentry>


</variablelist>

</section>


<section xml:id="ssec-setup-hooks"><title>Package setup hooks</title>

<para>The following packages provide a setup hook:

<variablelist>

  <varlistentry>
    <term>Bintools Wrapper</term>
    <listitem>
      <para>
        Bintools Wrapper wraps the binary utilities for a bunch of miscellaneous purposes.
        These are GNU Binutils when targetting Linux, and a mix of cctools and GNU binutils for Darwin.
        [The "Bintools" name is supposed to be a compromise between "Binutils" and "cctools" not denoting any specific implementation.]
        Specifically, the underlying bintools package, and a C standard library (glibc or Darwin's libSystem, just for the dynamic loader) are all fed in, and dependency finding, hardening (see below), and purity checks for each are handled by Bintools Wrapper.
        Packages typically depend on CC Wrapper, which in turn (at run time) depends on Bintools Wrapper.
      </para>
      <para>
        Bintools Wrapper was only just recently split off from CC Wrapper, so the division of labor is still being worked out.
        For example, it shouldn't care about about the C standard library, but just take a derivation with the dynamic loader (which happens to be the glibc on linux).
        Dependency finding however is a task both wrappers will continue to need to share, and probably the most important to understand.
        It is currently accomplished by collecting directories of host-platform dependencies (i.e. <varname>buildInputs</varname> and <varname>nativeBuildInputs</varname>) in environment variables.
        Bintools Wrapper's setup hook causes any <filename>lib</filename> and <filename>lib64</filename> subdirectories to be added to <envar>NIX_LDFLAGS</envar>.
        Since CC Wrapper and Bintools Wrapper use the same strategy, most of the Bintools Wrapper code is sparsely commented and refers to CC Wrapper.
        But CC Wrapper's code, by contrast, has quite lengthy comments.
        Bintools Wrapper merely cites those, rather than repeating them, to avoid falling out of sync.
      </para>
      <para>
        A final task of the setup hook is defining a number of standard environment variables to tell build systems which executables full-fill which purpose.
        They are defined to just be the base name of the tools, under the assumption that Bintools Wrapper's binaries will be on the path.
        Firstly, this helps poorly-written packages, e.g. ones that look for just <command>gcc</command> when <envar>CC</envar> isn't defined yet <command>clang</command> is to be used.
        Secondly, this helps packages not get confused when cross-compiling, in which case multiple Bintools Wrappers may simultaneously be in use.
        <footnote><para>
          Each wrapper targets a single platform, so if binaries for multiple platforms are needed, the underlying binaries must be wrapped multiple times.
          As this is a property of the wrapper itself, the multiple wrappings are needed whether or not the same underlying binaries can target multiple platforms.
        </para></footnote>
        <envar>BUILD_</envar>- and <envar>TARGET_</envar>-prefixed versions of the normal environment variable are defined for the additional Bintools Wrappers, properly disambiguating them.
      </para>
      <para>
        A problem with this final task is that Bintools Wrapper is honest and defines <envar>LD</envar> as <command>ld</command>.
        Most packages, however, firstly use the C compiler for linking, secondly use <envar>LD</envar> anyways, defining it as the C compiler, and thirdly, only so define <envar>LD</envar> when it is undefined as a fallback.
        This triple-threat means Bintools Wrapper will break those packages, as LD is already defined as the actual linker which the package won't override yet doesn't want to use.
        The workaround is to define, just for the problematic package, <envar>LD</envar> as the C compiler.
        A good way to do this would be <command>preConfigure = "LD=$CC"</command>.
      </para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term>CC Wrapper</term>
    <listitem>
      <para>
        CC Wrapper wraps a C toolchain for a bunch of miscellaneous purposes.
        Specifically, a C compiler (GCC or Clang), wrapped binary tools, and a C standard library (glibc or Darwin's libSystem, just for the dynamic loader) are all fed in, and dependency finding, hardening (see below), and purity checks for each are handled by CC Wrapper.
        Packages typically depend on CC Wrapper, which in turn (at run time) depends on Bintools Wrapper.
      </para>
      <para>
        Dependency finding is undoubtedly the main task of CC Wrapper.
        This works just like Bintools Wrapper, except that any <filename>include</filename> subdirectory of any relevant dependency is added to <envar>NIX_CFLAGS_COMPILE</envar>.
        The setup hook itself contains some lengthy comments describing the exact convoluted mechanism by which this is accomplished.
      </para>
      <para>
        CC Wrapper also like Bintools Wrapper defines standard environment variables with the names of the tools it wraps, for the same reasons described above.
        Importantly, while it includes a <command>cc</command> symlink to the c compiler for portability, the <envar>CC</envar> will be defined using the compiler's "real name" (i.e. <command>gcc</command> or <command>clang</command>).
        This helps lousy build systems that inspect on the name of the compiler rather than run it.
      </para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term>Perl</term>
    <listitem><para>Adds the <filename>lib/site_perl</filename> subdirectory
    of each build input to the <envar>PERL5LIB</envar>
    environment variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Python</term>
    <listitem><para>Adds the
    <filename>lib/${python.libPrefix}/site-packages</filename> subdirectory of
    each build input to the <envar>PYTHONPATH</envar> environment
    variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>pkg-config</term>
    <listitem><para>Adds the <filename>lib/pkgconfig</filename> and
    <filename>share/pkgconfig</filename> subdirectories of each
    build input to the <envar>PKG_CONFIG_PATH</envar> environment
    variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Automake</term>
    <listitem><para>Adds the <filename>share/aclocal</filename>
    subdirectory of each build input to the <envar>ACLOCAL_PATH</envar>
    environment variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Autoconf</term>
    <listitem><para>The <varname>autoreconfHook</varname> derivation adds
    <varname>autoreconfPhase</varname>, which runs autoreconf, libtoolize and
    automake, essentially preparing the configure script in autotools-based
    builds.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>libxml2</term>
    <listitem><para>Adds every file named
    <filename>catalog.xml</filename> found under the
    <filename>xml/dtd</filename> and <filename>xml/xsl</filename>
    subdirectories of each build input to the
    <envar>XML_CATALOG_FILES</envar> environment
    variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>teTeX / TeX Live</term>
    <listitem><para>Adds the <filename>share/texmf-nix</filename>
    subdirectory of each build input to the <envar>TEXINPUTS</envar>
    environment variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>Qt 4</term>
    <listitem><para>Sets the <envar>QTDIR</envar> environment variable
    to Qt’s path.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>gdk-pixbuf</term>
    <listitem><para>Exports <envar>GDK_PIXBUF_MODULE_FILE</envar>
    environment variable the the builder. Add librsvg package
    to <varname>buildInputs</varname> to get svg support.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>GHC</term>
    <listitem><para>Creates a temporary package database and registers
    every Haskell build input in it (TODO: how?).</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>GStreamer</term>
    <listitem><para>Adds the
    GStreamer plugins subdirectory of
    each build input to the <envar>GST_PLUGIN_SYSTEM_PATH_1_0</envar> or
    <envar>GST_PLUGIN_SYSTEM_PATH</envar> environment variable.</para></listitem>
  </varlistentry>

  <varlistentry>
    <term>paxctl</term>
    <listitem><para>Defines the <varname>paxmark</varname> helper for
    setting per-executable PaX flags on Linux (where it is available by
    default; on all other platforms, <varname>paxmark</varname> is a no-op).
    For example, to disable secure memory protections on the executable
    <replaceable>foo</replaceable>:
    <programlisting>
      postFixup = ''
        paxmark m $out/bin/<replaceable>foo</replaceable>
      '';
    </programlisting>
    The <literal>m</literal> flag is the most common flag and is typically
    required for applications that employ JIT compilation or otherwise need to
    execute code generated at run-time.  Disabling PaX protections should be
    considered a last resort: if possible, problematic features should be
    disabled or patched to work with PaX.</para></listitem>
  </varlistentry>

</variablelist>

</para>

</section>


<section xml:id="sec-purity-in-nixpkgs"><title>Purity in Nixpkgs</title>

<para>[measures taken to prevent dependencies on packages outside the
store, and what you can do to prevent them]</para>

<para>GCC doesn't search in locations such as
<filename>/usr/include</filename>.  In fact, attempts to add such
directories through the <option>-I</option> flag are filtered out.
Likewise, the linker (from GNU binutils) doesn't search in standard
locations such as <filename>/usr/lib</filename>.  Programs built on
Linux are linked against a GNU C Library that likewise doesn't search
in the default system locations.</para>

</section>

<section xml:id="sec-hardening-in-nixpkgs"><title>Hardening in Nixpkgs</title>

<para>There are flags available to harden packages at compile or link-time.
These can be toggled using the <varname>stdenv.mkDerivation</varname> parameters
<varname>hardeningDisable</varname> and <varname>hardeningEnable</varname>.
</para>

<para>
Both parameters take a list of flags as strings. The special
<varname>"all"</varname> flag can be passed to <varname>hardeningDisable</varname>
to turn off all hardening. These flags can also be used as environment variables
for testing or development purposes.
</para>

<para>The following flags are enabled by default and might require disabling with
<varname>hardeningDisable</varname> if the program to package is incompatible.
</para>

<variablelist>

  <varlistentry>
    <term><varname>format</varname></term>
    <listitem><para>Adds the <option>-Wformat -Wformat-security
    -Werror=format-security</option> compiler options. At present,
    this warns about calls to <varname>printf</varname> and
    <varname>scanf</varname> functions where the format string is
    not a string literal and there are no format arguments, as in
    <literal>printf(foo);</literal>. This may be a security hole
    if the format string came from untrusted input and contains
    <literal>%n</literal>.</para>

    <para>This needs to be turned off or fixed for errors similar to:</para>

    <programlisting>
/tmp/nix-build-zynaddsubfx-2.5.2.drv-0/zynaddsubfx-2.5.2/src/UI/guimain.cpp:571:28: error: format not a string literal and no format arguments [-Werror=format-security]
         printf(help_message);
                            ^
cc1plus: some warnings being treated as errors
    </programlisting></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>stackprotector</varname></term>
    <listitem>
    <para>Adds the <option>-fstack-protector-strong
    --param ssp-buffer-size=4</option>
    compiler options. This adds safety checks against stack overwrites
    rendering many potential code injection attacks into aborting situations.
    In the best case this turns code injection vulnerabilities into denial
    of service or into non-issues (depending on the application).</para>

    <para>This needs to be turned off or fixed for errors similar to:</para>

    <programlisting>
bin/blib.a(bios_console.o): In function `bios_handle_cup':
/tmp/nix-build-ipxe-20141124-5cbdc41.drv-0/ipxe-5cbdc41/src/arch/i386/firmware/pcbios/bios_console.c:86: undefined reference to `__stack_chk_fail'
    </programlisting></listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>fortify</varname></term>
    <listitem>
    <para>Adds the <option>-O2 -D_FORTIFY_SOURCE=2</option> compiler
    options. During code generation the compiler knows a great deal of
    information about buffer sizes (where possible), and attempts to replace
    insecure unlimited length buffer function calls with length-limited ones.
    This is especially useful for old, crufty code. Additionally, format
    strings in writable memory that contain '%n' are blocked. If an application
    depends on such a format string, it will need to be worked around.
    </para>

    <para>Additionally, some warnings are enabled which might trigger build
    failures if compiler warnings are treated as errors in the package build.
    In this case, set <option>NIX_CFLAGS_COMPILE</option> to
    <option>-Wno-error=warning-type</option>.</para>

    <para>This needs to be turned off or fixed for errors similar to:</para>

    <programlisting>
malloc.c:404:15: error: return type is an incomplete type
malloc.c:410:19: error: storage size of 'ms' isn't known
    </programlisting>
    <programlisting>
strdup.h:22:1: error: expected identifier or '(' before '__extension__'
    </programlisting>
    <programlisting>
strsep.c:65:23: error: register name not specified for 'delim'
    </programlisting>
    <programlisting>
installwatch.c:3751:5: error: conflicting types for '__open_2'
    </programlisting>
    <programlisting>
fcntl2.h:50:4: error: call to '__open_missing_mode' declared with attribute error: open with O_CREAT or O_TMPFILE in second argument needs 3 arguments
    </programlisting>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>pic</varname></term>
    <listitem>
    <para>Adds the <option>-fPIC</option> compiler options. This options adds
    support for position independent code in shared libraries and thus making
    ASLR possible.</para>
    <para>Most notably, the Linux kernel, kernel modules and other code
    not running in an operating system environment like boot loaders won't
    build with PIC enabled. The compiler will is most cases complain that
    PIC is not supported for a specific build.
    </para>

    <para>This needs to be turned off or fixed for assembler errors similar to:</para>

    <programlisting>
ccbLfRgg.s: Assembler messages:
ccbLfRgg.s:33: Error: missing or invalid displacement expression `private_key_len@GOTOFF'
    </programlisting>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>strictoverflow</varname></term>
    <listitem>
    <para>Signed integer overflow is undefined behaviour according to the C
    standard. If it happens, it is an error in the program as it should check
    for overflow before it can happen, not afterwards. GCC provides built-in
    functions to perform arithmetic with overflow checking, which are correct
    and faster than any custom implementation. As a workaround, the option
    <option>-fno-strict-overflow</option> makes gcc behave as if signed
    integer overflows were defined.
    </para>

    <para>This flag should not trigger any build or runtime errors.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>relro</varname></term>
    <listitem>
    <para>Adds the <option>-z relro</option> linker option. During program
    load, several ELF memory sections need to be written to by the linker,
    but can be turned read-only before turning over control to the program.
    This prevents some GOT (and .dtors) overwrite attacks, but at least the
    part of the GOT used by the dynamic linker (.got.plt) is still vulnerable.
    </para>

    <para>This flag can break dynamic shared object loading. For instance, the
    module systems of Xorg and OpenCV are incompatible with this flag. In almost
    all cases the <varname>bindnow</varname> flag must also be disabled and
    incompatible programs typically fail with similar errors at runtime.</para>
    </listitem>
  </varlistentry>

  <varlistentry>
    <term><varname>bindnow</varname></term>
    <listitem>
    <para>Adds the <option>-z bindnow</option> linker option. During program
    load, all dynamic symbols are resolved, allowing for the complete GOT to
    be marked read-only (due to <varname>relro</varname>). This prevents GOT
    overwrite attacks. For very large applications, this can incur some
    performance loss during initial load while symbols are resolved, but this
    shouldn't be an issue for daemons.
    </para>

    <para>This flag can break dynamic shared object loading. For instance, the
    module systems of Xorg and PHP are incompatible with this flag. Programs
    incompatible with this flag often fail at runtime due to missing symbols,
    like:</para>

    <programlisting>
intel_drv.so: undefined symbol: vgaHWFreeHWRec
    </programlisting>
    </listitem>
  </varlistentry>

</variablelist>

<para>The following flags are disabled by default and should be enabled
with <varname>hardeningEnable</varname> for packages that take untrusted
input like network services.
</para>

<variablelist>

  <varlistentry>
    <term><varname>pie</varname></term>
    <listitem>
    <para>Adds the <option>-fPIE</option> compiler and <option>-pie</option>
    linker options. Position Independent Executables are needed to take
    advantage of Address Space Layout Randomization, supported by modern
    kernel versions. While ASLR can already be enforced for data areas in
    the stack and heap (brk and mmap), the code areas must be compiled as
    position-independent. Shared libraries already do this with the
    <varname>pic</varname> flag, so they gain ASLR automatically, but binary
    .text regions need to be build with <varname>pie</varname> to gain ASLR.
    When this happens, ROP attacks are much harder since there are no static
    locations to bounce off of during a memory corruption attack.
    </para>
    </listitem>
  </varlistentry>

</variablelist>

<para>For more in-depth information on these hardening flags and hardening in
general, refer to the
<link xlink:href="https://wiki.debian.org/Hardening">Debian Wiki</link>,
<link xlink:href="https://wiki.ubuntu.com/Security/Features">Ubuntu Wiki</link>,
<link xlink:href="https://wiki.gentoo.org/wiki/Project:Hardened">Gentoo Wiki</link>,
and the <link xlink:href="https://wiki.archlinux.org/index.php/DeveloperWiki:Security">
Arch Wiki</link>.
</para>

</section>

</chapter>