New syntax for GADT-style record declarations, and associated refactoring

[ghc-hetmet.git] / docs / users_guide / separate_compilation.xml

<?xml version="1.0" encoding="iso-8859-1"?>
  <sect1 id="separate-compilation">
    <title>Filenames and separate compilation</title>

    <indexterm><primary>separate compilation</primary></indexterm>
    <indexterm><primary>recompilation checker</primary></indexterm>
    <indexterm><primary>make and recompilation</primary></indexterm>

    <para>This section describes what files GHC expects to find, what
    files it creates, where these files are stored, and what options
    affect this behaviour.</para>

    <para>Note that this section is written with
    <firstterm>hierarchical modules</firstterm> in mind (see <xref
    linkend="hierarchical-modules"/>); hierarchical modules are an
    extension to Haskell 98 which extends the lexical syntax of
    module names to include a dot &lsquo;.&rsquo;.  Non-hierarchical
    modules are thus a special case in which none of the module names
    contain dots.</para>

    <para>Pathname conventions vary from system to system.  In
    particular, the directory separator is
    &lsquo;<literal>/</literal>&rsquo; on Unix systems and
    &lsquo;<literal>\</literal>&rsquo; on Windows systems.  In the
    sections that follow, we shall consistently use
    &lsquo;<literal>/</literal>&rsquo; as the directory separator;
    substitute this for the appropriate character for your
    system.</para>

    <sect2 id="source-files">
      <title>Haskell source files</title>

      <indexterm><primary>filenames</primary></indexterm>

      <para>Each Haskell source module should be placed in a file on
      its own.</para>

      <para>Usually, the file should be named after the module name,
      replacing dots in the module name by directory separators.  For
      example, on a Unix system, the module <literal>A.B.C</literal>
      should be placed in the file <literal>A/B/C.hs</literal>,
      relative to some base directory.  If the module is not going to
      be imported by another module (<literal>Main</literal>, for
      example), then you are free to use any filename for it.</para>

      <indexterm><primary>unicode</primary></indexterm>

      <para> GHC assumes that source files are
      ASCII<indexterm><primary>ASCII</primary></indexterm> or
      UTF-8<indexterm><primary>UTF-8</primary></indexterm> only, other
      encodings<indexterm><primary>encoding</primary></indexterm> are
      not recognised.  However, invalid UTF-8 sequences will be
      ignored in comments, so it is possible to use other encodings
      such as
      Latin-1<indexterm><primary>Latin-1</primary></indexterm>, as
      long as the non-comment source code is ASCII only.</para>
    </sect2>

    <sect2 id="output-files">
      <title>Output files</title>

      <indexterm><primary>interface files</primary></indexterm>
      <indexterm><primary><literal>.hi</literal> files</primary></indexterm>
      <indexterm><primary>object files</primary></indexterm>
      <indexterm><primary><literal>.o</literal> files</primary></indexterm>

      <para>When asked to compile a source file, GHC normally
      generates two files: an <firstterm>object file</firstterm>, and
      an <firstterm>interface file</firstterm>. </para>

      <para>The object file, which normally ends in a
      <literal>.o</literal> suffix, contains the compiled code for the
      module.</para>

      <para>The interface file,
      which normally ends in a <literal>.hi</literal> suffix, contains
      the information that GHC needs in order to compile further
      modules that depend on this module.  It contains things like the
      types of exported functions, definitions of data types, and so
      on.  It is stored in a binary format, so don't try to read one;
      use the <option>--show-iface</option> option instead (see <xref
      linkend="hi-options"/>).</para>

      <para>You should think of the object file and the interface file as a
      pair, since the interface file is in a sense a compiler-readable
      description of the contents of the object file.  If the
      interface file and object file get out of sync for any reason,
      then the compiler may end up making assumptions about the object
      file that aren't true; trouble will almost certainly follow.
      For this reason, we recommend keeping object files and interface
      files in the same place (GHC does this by default, but it is
      possible to override the defaults as we'll explain
      shortly).</para>

      <para>Every module has a <emphasis>module name</emphasis>
      defined in its source code (<literal>module A.B.C where
      ...</literal>).</para>

      <para>The name of the object file generated by GHC is derived
      according to the following rules, where
      <replaceable>osuf</replaceable> is the object-file suffix (this
      can be changed with the <option>-osuf</option> option).</para>

      <itemizedlist>
        <listitem>
          <para>If there is no <option>-odir</option> option (the
          default), then the object filename is derived from the
          source filename (ignoring the module name) by replacing the
          suffix with <replaceable>osuf</replaceable>.</para>
        </listitem>
        <listitem>
          <para>If
          <option>-odir</option>&nbsp;<replaceable>dir</replaceable>
          has been specified, then the object filename is
          <replaceable>dir</replaceable>/<replaceable>mod</replaceable>.<replaceable>osuf</replaceable>,
          where <replaceable>mod</replaceable> is the module name with
          dots replaced by slashes.  GHC will silently create the necessary directory 
          structure underneath <replaceable>dir</replaceable>, if it does not
          already exist.</para> 
        </listitem>
      </itemizedlist>

      <para>The name of the interface file is derived using the same
      rules, except that the suffix is
      <replaceable>hisuf</replaceable> (<literal>.hi</literal> by
      default) instead of <replaceable>osuf</replaceable>, and the
      relevant options are <option>-hidir</option> and
      <option>-hisuf</option> instead of <option>-odir</option> and
      <option>-osuf</option> respectively.</para>

      <para>For example, if GHC compiles the module
      <literal>A.B.C</literal> in the file
      <filename>src/A/B/C.hs</filename>, with no
      <literal>-odir</literal> or <literal>-hidir</literal> flags, the
      interface file will be put in <literal>src/A/B/C.hi</literal>
      and the object file in <literal>src/A/B/C.o</literal>.</para>

      <para>For any module that is imported, GHC requires that the
      name of the module in the import statement exactly matches the
      name of the module in the interface file (or source file) found
      using the strategy specified in <xref linkend="search-path"/>.
      This means that for most modules, the source file name should
      match the module name.</para>

      <para>However, note that it is reasonable to have a module
      <literal>Main</literal> in a file named
      <filename>foo.hs</filename>, but this only works because GHC
      never needs to search for the interface for module
      <literal>Main</literal> (because it is never imported).  It is
      therefore possible to have several <literal>Main</literal>
      modules in separate source files in the same directory, and GHC
      will not get confused.</para>

      <para>In batch compilation mode, the name of the object file can
      also be overridden using the <option>-o</option> option, and the
      name of the interface file can be specified directly using the
      <option>-ohi</option> option.</para>
    </sect2>

    <sect2 id="search-path">
      <title>The search path</title>

      <indexterm><primary>search path</primary>
      </indexterm>
      <indexterm><primary>interface files, finding them</primary></indexterm>
      <indexterm><primary>finding interface files</primary></indexterm>

      <para>In your program, you import a module
      <literal>Foo</literal> by saying <literal>import Foo</literal>.
      In <option>--make</option> mode or GHCi, GHC will look for a
      source file for <literal>Foo</literal> and arrange to compile it
      first.  Without <option>--make</option>, GHC will look for the
      interface file for <literal>Foo</literal>, which should have
      been created by an earlier compilation of
      <literal>Foo</literal>.  GHC uses the same strategy in each of
      these cases for finding the appropriate file.</para>

      <para>This strategy is as follows: GHC keeps a list of
      directories called the <firstterm>search path</firstterm>.  For
      each of these directories, it tries appending
      <replaceable>basename</replaceable><literal>.</literal><replaceable>extension</replaceable>
      to the directory, and checks whether the file exists.  The value
      of <replaceable>basename</replaceable> is the module name with
      dots replaced by the directory separator ('/' or '\', depending
      on the system), and <replaceable>extension</replaceable> is a
      source extension (<literal>hs</literal>, <literal>lhs</literal>)
      if we are in <option>--make</option> mode or GHCi, or
      <replaceable>hisuf</replaceable> otherwise.</para>

      <para>For example, suppose the search path contains directories
      <literal>d1</literal>, <literal>d2</literal>, and
      <literal>d3</literal>, and we are in <literal>--make</literal>
      mode looking for the source file for a module
      <literal>A.B.C</literal>.  GHC will look in
      <literal>d1/A/B/C.hs</literal>, <literal>d1/A/B/C.lhs</literal>,
      <literal>d2/A/B/C.hs</literal>, and so on.</para>

      <para>The search path by default contains a single directory:
      <quote>.</quote> (i.e. the current directory).  The following
      options can be used to add to or change the contents of the
      search path:</para>

      <variablelist>
        <varlistentry>
          <term><option>-i<replaceable>dirs</replaceable></option></term>
          <listitem>
            <para><indexterm><primary><option>-i<replaceable>dirs</replaceable></option>
            </primary></indexterm>This flag appends a colon-separated
            list of <filename>dirs</filename> to the search path.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term><option>-i</option></term>
          <listitem>
            <para>resets the search path back to nothing.</para>
          </listitem>
        </varlistentry>
      </variablelist>

      <para>This isn't the whole story: GHC also looks for modules in
      pre-compiled libraries, known as packages.  See the section on
      packages (<xref linkend="packages"/>) for details.</para>
    </sect2>

    <sect2 id="options-output">
      <title>Redirecting the compilation output(s)</title>

      <indexterm><primary>output-directing options</primary></indexterm>
      <indexterm><primary>redirecting compilation output</primary></indexterm>

      <variablelist>
        <varlistentry>
          <term>
            <option>-o</option> <replaceable>file</replaceable>
            <indexterm><primary><option>-o</option></primary></indexterm>
          </term>
          <listitem>
            <para>GHC's compiled output normally goes into a
            <filename>.hc</filename>, <filename>.o</filename>, etc.,
            file, depending on the last-run compilation phase.  The
            option <option>-o <replaceable>file</replaceable></option>
            re-directs the output of that last-run phase to
            <replaceable>file</replaceable>.</para>

            <para>Note: this &ldquo;feature&rdquo; can be
            counterintuitive: <command>ghc -C -o foo.o
            foo.hs</command> will put the intermediate C code in the
            file <filename>foo.o</filename>, name
            notwithstanding!</para>

            <para>This option is most often used when creating an
            executable file, to set the filename of the executable.
            For example:
<screen>   ghc -o prog --make Main</screen>

            will compile the program starting with module
            <literal>Main</literal>  and put the executable in the
            file <literal>prog</literal>.</para>

            <para>Note: on Windows, if the result is an executable
            file, the extension "<filename>.exe</filename>" is added
            if the specified filename does not already have an
            extension.  Thus
<programlisting>
   ghc -o foo Main.hs
</programlisting>
          will compile and link the module
          <filename>Main.hs</filename>, and put the resulting
          executable in <filename>foo.exe</filename> (not
          <filename>foo</filename>).</para>

          <para>If you use <command>ghc --make</command> and you don't
          use the <option>-o</option>, the name GHC will choose
          for the executable will be based on the name of the file
          containing the module <literal>Main</literal>. 
          Note that with GHC the <literal>Main</literal> module doesn't
          have to be put in file <filename>Main.hs</filename>.
          Thus both
<programlisting>
   ghc --make Prog
</programlisting>
          and
<programlisting>
   ghc --make Prog.hs
</programlisting>
          will produce <filename>Prog</filename> (or
          <filename>Prog.exe</filename> if you are on Windows).</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-odir</option> <replaceable>dir</replaceable>
            <indexterm><primary><option>-odir</option></primary></indexterm>
          </term>
          <listitem>
            <para>Redirects object files to directory
            <replaceable>dir</replaceable>.  For example:</para>

<screen>
$ ghc -c parse/Foo.hs parse/Bar.hs gurgle/Bumble.hs -odir `uname -m`
</screen>

            <para>The object files, <filename>Foo.o</filename>,
            <filename>Bar.o</filename>, and
            <filename>Bumble.o</filename> would be put into a
            subdirectory named after the architecture of the executing
            machine (<filename>x86</filename>,
            <filename>mips</filename>, etc).</para>

            <para>Note that the <option>-odir</option> option does
            <emphasis>not</emphasis> affect where the interface files
            are put; use the <option>-hidir</option> option for that.
            In the above example, they would still be put in
            <filename>parse/Foo.hi</filename>,
            <filename>parse/Bar.hi</filename>, and
            <filename>gurgle/Bumble.hi</filename>.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-ohi</option>  <replaceable>file</replaceable>
            <indexterm><primary><option>-ohi</option></primary></indexterm>
          </term>
          <listitem>
            <para>The interface output may be directed to another file
            <filename>bar2/Wurble.iface</filename> with the option
            <option>-ohi bar2/Wurble.iface</option> (not
            recommended).</para>

            <para>WARNING: if you redirect the interface file
            somewhere that GHC can't find it, then the recompilation
            checker may get confused (at the least, you won't get any
            recompilation avoidance).  We recommend using a
            combination of <option>-hidir</option> and
            <option>-hisuf</option> options instead, if
            possible.</para>

            <para>To avoid generating an interface at all, you could
            use this option to redirect the interface into the bit
            bucket: <literal>-ohi /dev/null</literal>, for
            example.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-hidir</option>  <replaceable>dir</replaceable>
            <indexterm><primary><option>-hidir</option></primary></indexterm>
          </term>
          <listitem>
            <para>Redirects all generated interface files into
            <replaceable>dir</replaceable>, instead of the
            default.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-stubdir</option>  <replaceable>dir</replaceable>
            <indexterm><primary><option>-stubdir</option></primary></indexterm>
          </term>
          <listitem>
            <para>Redirects all generated FFI stub files into
            <replaceable>dir</replaceable>.  Stub files are generated when the
            Haskell source contains a <literal>foreign export</literal> or
            <literal>foreign import "&amp;wrapper"</literal> declaration (see <xref
              linkend="foreign-export-ghc" />).  The <option>-stubdir</option>
              option behaves in exactly the same way as <option>-odir</option>
              and <option>-hidir</option> with respect to hierarchical
            modules.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-outputdir</option> <replaceable>dir</replaceable>
            <indexterm><primary><option>-outputdir</option></primary></indexterm>
          </term>
          <listitem>
            <para>The <option>-outputdir</option> option is shorthand for
              the combination
              of <option>-odir</option>, <option>-hidir</option>,
              and <option>-stubdir</option>.
            </para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-osuf</option> <replaceable>suffix</replaceable>
            <indexterm><primary><option>-osuf</option></primary></indexterm>
          </term>
          <term>
            <option>-hisuf</option> <replaceable>suffix</replaceable>
            <indexterm><primary><option>-hisuf</option></primary></indexterm>
          </term>
          <term>
            <option>-hcsuf</option> <replaceable>suffix</replaceable>
            <indexterm><primary><option>-hcsuf</option></primary></indexterm>
          </term>
          <listitem>
            <para>The <option>-osuf</option>
            <replaceable>suffix</replaceable> will change the
            <literal>.o</literal> file suffix for object files to
            whatever you specify.  We use this when compiling
            libraries, so that objects for the profiling versions of
            the libraries don't clobber the normal ones.</para>

            <para>Similarly, the <option>-hisuf</option>
            <replaceable>suffix</replaceable> will change the
            <literal>.hi</literal> file suffix for non-system
            interface files (see <xref linkend="hi-options"/>).</para>

            <para>Finally, the option <option>-hcsuf</option>
            <replaceable>suffix</replaceable> will change the
            <literal>.hc</literal> file suffix for compiler-generated
            intermediate C files.</para>

            <para>The <option>-hisuf</option>/<option>-osuf</option>
            game is particularly useful if you want to compile a
            program both with and without profiling, in the same
            directory.  You can say:
            <screen>
              ghc ...</screen>
            to get the ordinary version, and
            <screen>
              ghc ... -osuf prof.o -hisuf prof.hi -prof -auto-all</screen>
            to get the profiled version.</para>
          </listitem>
        </varlistentry>
      </variablelist>
    </sect2>
  
    <sect2 id="keeping-intermediates">
      <title>Keeping Intermediate Files</title>
      <indexterm><primary>intermediate files, saving</primary>
      </indexterm>
      <indexterm><primary><literal>.hc</literal> files, saving</primary>
      </indexterm>
      <indexterm><primary><literal>.s</literal> files, saving</primary>
      </indexterm>

      <para>The following options are useful for keeping certain
      intermediate files around, when normally GHC would throw these
      away after compilation:</para>

      <variablelist>
        <varlistentry>
          <term>
            <option>-keep-hc-file</option>,
            <option>-keep-hc-files</option>
            <indexterm><primary><option>-keep-hc-file</option></primary></indexterm>
            <indexterm><primary><option>-keep-hc-files</option></primary></indexterm>
          </term>
          <listitem>
            <para>Keep intermediate <literal>.hc</literal> files when
            doing <literal>.hs</literal>-to-<literal>.o</literal>
            compilations via C (NOTE: <literal>.hc</literal> files
            aren't generated when using the native code generator, you
            may need to use <option>-fvia-C</option> to force them
            to be produced).</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-keep-s-file</option>,
            <option>-keep-s-files</option>
            <indexterm><primary><option>-keep-s-file</option></primary></indexterm>
            <indexterm><primary><option>-keep-s-files</option></primary></indexterm>
          </term>
          <listitem>
            <para>Keep intermediate <literal>.s</literal> files.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-keep-raw-s-file</option>,
            <option>-keep-raw-s-files</option>
            <indexterm><primary><option>-keep-raw-s-file</option></primary></indexterm>
            <indexterm><primary><option>-keep-raw-s-files</option></primary></indexterm>
          </term>
          <listitem>
            <para>Keep intermediate <literal>.raw-s</literal> files.
            These are the direct output from the C compiler, before
            GHC does &ldquo;assembly mangling&rdquo; to produce the
            <literal>.s</literal> file.  Again, these are not produced
            when using the native code generator.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-keep-tmp-files</option>
            <indexterm><primary><option>-keep-tmp-files</option></primary></indexterm>
            <indexterm><primary>temporary files</primary><secondary>keeping</secondary></indexterm>
          </term>
          <listitem>
            <para>Instructs the GHC driver not to delete any of its
            temporary files, which it normally keeps in
            <literal>/tmp</literal> (or possibly elsewhere; see <xref
            linkend="temp-files"/>).  Running GHC with
            <option>-v</option> will show you what temporary files
            were generated along the way.</para>
          </listitem>
        </varlistentry>
      </variablelist>
    </sect2>

    <sect2 id="temp-files">
      <title>Redirecting temporary files</title>

      <indexterm>
        <primary>temporary files</primary>
        <secondary>redirecting</secondary>
      </indexterm>

      <variablelist>
        <varlistentry>
          <term>
            <option>-tmpdir</option>
            <indexterm><primary><option>-tmpdir</option></primary></indexterm>
          </term>
          <listitem>
            <para>If you have trouble because of running out of space
            in <filename>/tmp</filename> (or wherever your
            installation thinks temporary files should go), you may
            use the <option>-tmpdir
            &lt;dir&gt;</option><indexterm><primary>-tmpdir
            &lt;dir&gt; option</primary></indexterm> option to specify
            an alternate directory.  For example, <option>-tmpdir
            .</option> says to put temporary files in the current
            working directory.</para>

            <para>Alternatively, use your <constant>TMPDIR</constant>
            environment variable.<indexterm><primary>TMPDIR
            environment variable</primary></indexterm> Set it to the
            name of the directory where temporary files should be put.
            GCC and other programs will honour the
            <constant>TMPDIR</constant> variable as well.</para>

            <para>Even better idea: Set the
            <constant>DEFAULT_TMPDIR</constant> make variable when
            building GHC, and never worry about
            <constant>TMPDIR</constant> again. (see the build
            documentation).</para>
          </listitem>
        </varlistentry>
      </variablelist>
    </sect2>

    <sect2 id="hi-options">
      <title>Other options related to interface files</title>
      <indexterm><primary>interface files, options</primary></indexterm>

      <variablelist>
        <varlistentry>
          <term>
            <option>-ddump-hi</option>
            <indexterm><primary><option>-ddump-hi</option></primary></indexterm>
          </term>
          <listitem>
            <para>Dumps the new interface to standard output.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-ddump-hi-diffs</option>
            <indexterm><primary><option>-ddump-hi-diffs</option></primary></indexterm>
          </term>
          <listitem>
            <para>The compiler does not overwrite an existing
            <filename>.hi</filename> interface file if the new one is
            the same as the old one; this is friendly to
            <command>make</command>.  When an interface does change,
            it is often enlightening to be informed.  The
            <option>-ddump-hi-diffs</option> option will make GHC
            report the differences between the old and
            new <filename>.hi</filename> files.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>-ddump-minimal-imports</option>
            <indexterm><primary><option>-ddump-minimal-imports</option></primary></indexterm>
          </term>
          <listitem>
            <para>Dump to the file "M.imports" (where M is the module
            being compiled) a "minimal" set of import declarations.
            You can safely replace all the import declarations in
            "M.hs" with those found in "M.imports".  Why would you
            want to do that?  Because the "minimal" imports (a) import
            everything explicitly, by name, and (b) import nothing
            that is not required.  It can be quite painful to maintain
            this property by hand, so this flag is intended to reduce
            the labour.</para>
          </listitem>
        </varlistentry>

        <varlistentry>
          <term>
            <option>--show-iface</option> <replaceable>file</replaceable>
            <indexterm><primary><option>--show-iface</option></primary></indexterm>
          </term>
          <listitem>
            <para>where <replaceable>file</replaceable> is the name of
            an interface file, dumps the contents of that interface in
            a human-readable (ish) format. See <xref linkend="modes"/>.</para>
          </listitem>
        </varlistentry>
      </variablelist>
    </sect2>

    <sect2 id="recomp">
      <title>The recompilation checker</title>

      <indexterm><primary>recompilation checker</primary></indexterm>

      <variablelist>
        <varlistentry>
          <term>
            <option>-fforce-recomp</option>
            <indexterm><primary><option>-fforce-recomp</option></primary></indexterm>
            <indexterm><primary><option>-fno-force-recomp</option></primary></indexterm>
          </term>
          <listitem>
            <para>Turn off recompilation checking (which is on by
            default).  Recompilation checking normally stops
            compilation early, leaving an existing
            <filename>.o</filename> file in place, if it can be
            determined that the module does not need to be
            recompiled.</para>
          </listitem>
        </varlistentry>
      </variablelist>

      <para>In the olden days, GHC compared the newly-generated
      <filename>.hi</filename> file with the previous version; if they
      were identical, it left the old one alone and didn't change its
      modification date.  In consequence, importers of a module with
      an unchanged output <filename>.hi</filename> file were not
      recompiled.</para>

      <para>This doesn't work any more.  Suppose module
      <literal>C</literal> imports module <literal>B</literal>, and
      <literal>B</literal> imports module <literal>A</literal>.  So
      changes to module <literal>A</literal> might require module
      <literal>C</literal> to be recompiled, and hence when
      <filename>A.hi</filename> changes we should check whether
      <literal>C</literal> should be recompiled.  However, the
      dependencies of <literal>C</literal> will only list
      <literal>B.hi</literal>, not <literal>A.hi</literal>, and some
      changes to <literal>A</literal> (changing the definition of a
      function that appears in an inlining of a function exported by
      <literal>B</literal>, say) may conceivably not change
      <filename>B.hi</filename> one jot.  So now&hellip;</para>

      <para>GHC calculates a fingerprint (in fact an MD5 hash) of each
      interface file, and of each declaration within the interface
      file.  It also keeps in every interface file a list of the
      fingerprints of everything it used when it last compiled the
      file.  If the source file's modification date is earlier than
      the <filename>.o</filename> file's date (i.e. the source hasn't
      changed since the file was last compiled), and the recompilation
      checking is on, GHC will be clever.  It compares the fingerprints
      on the things it needs this time with the fingerprints
      on the things it needed last time (gleaned from the
      interface file of the module being compiled); if they are all
      the same it stops compiling early in the process saying
      &ldquo;Compilation IS NOT required&rdquo;.  What a beautiful
      sight!</para>

      <para>You can read
      about <ulink url="http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/RecompilationAvoidance">how
          all this works</ulink> in the GHC commentary.</para>

    </sect2>

    <sect2 id="mutual-recursion">
      <title>How to compile mutually recursive modules</title>

      <indexterm><primary>module system, recursion</primary></indexterm>
      <indexterm><primary>recursion, between modules</primary></indexterm>

      <para>GHC supports the compilation of mutually recursive modules.
      This section explains how.</para>

      <para>Every cycle in the module import graph must be broken by a <filename>hs-boot</filename> file.
      Suppose that modules <filename>A.hs</filename> and <filename>B.hs</filename> are Haskell source files, 
      thus:
<programlisting>
module A where
    import B( TB(..) )
    
    newtype TA = MkTA Int
    
    f :: TB -&#62; TA
    f (MkTB x) = MkTA x

module B where
    import {-# SOURCE #-} A( TA(..) )
    
    data TB = MkTB !Int
    
    g :: TA -&#62; TB
    g (MkTA x) = MkTB x
</programlisting>
<indexterm><primary><literal>hs-boot</literal>
      files</primary></indexterm> <indexterm><primary>importing,
      <literal>hi-boot</literal> files</primary></indexterm>
Here <filename>A</filename> imports <filename>B</filename>, but <filename>B</filename> imports
<filename>A</filename> with a <literal>{-# SOURCE #-}</literal> pragma, which breaks the
circular dependency.  Every loop in the module import graph must be broken by a <literal>{-# SOURCE #-}</literal> import;
or, equivalently, the module import graph must be acyclic if <literal>{-# SOURCE #-}</literal> imports are ignored.
</para>
<para>For every module <filename>A.hs</filename> that is <literal>{-# SOURCE #-}</literal>-imported
in this way there must exist a source file <literal>A.hs-boot</literal>.  This file contains an abbreviated
version of <filename>A.hs</filename>, thus:
<programlisting>
module A where
    newtype TA = MkTA Int
</programlisting>
</para>
<para>To compile these three files, issue the following commands:
<programlisting>
  ghc -c A.hs-boot    -- Produces A.hi-boot, A.o-boot
  ghc -c B.hs         -- Consumes A.hi-boot, produces B.hi, B.o
  ghc -c A.hs         -- Consumes B.hi, produces A.hi, A.o
  ghc -o foo A.o B.o  -- Linking the program
</programlisting>
</para>
<para>There are several points to note here:
<itemizedlist>
<listitem>
  <para>The file <filename>A.hs-boot</filename> is a programmer-written source file.
  It must live in the same directory as its parent source file <filename>A.hs</filename>.
  Currently, if you use a literate source file <filename>A.lhs</filename> you must
  also use a literate boot file, <filename>A.lhs-boot</filename>; and vice versa.
  </para></listitem>

<listitem><para>
  A <filename>hs-boot</filename> file is compiled by GHC, just like a <filename>hs</filename> file:
<programlisting>
  ghc -c A.hs-boot
</programlisting>
When a hs-boot file <filename>A.hs-boot</filename> 
   is compiled, it is checked for scope and type errors.
   When its parent module <filename>A.hs</filename> is compiled, the two are compared, and
   an error is reported if the two are inconsistent.
   </para></listitem>
   
        <listitem>
          <para> Just as compiling <filename>A.hs</filename> produces an
            interface file <filename>A.hi</filename>, and an object file
            <filename>A.o</filename>, so compiling
            <filename>A.hs-boot</filename> produces an interface file
            <filename>A.hi-boot</filename>, and an pseudo-object file
            <filename>A.o-boot</filename>: </para>

          <itemizedlist>
            <listitem>
              <para>The pseudo-object file <filename>A.o-boot</filename> is
                empty (don't link it!), but it is very useful when using a
                Makefile, to record when the <filename>A.hi-boot</filename> was
                last brought up to date (see <xref
                  linkend="using-make"/>).</para>
            </listitem>

            <listitem>
              <para>The <filename>hi-boot</filename> generated by compiling a
                <filename>hs-boot</filename> file is in the same
                machine-generated binary format as any other GHC-generated
                interface file (e.g. <filename>B.hi</filename>). You can
                display its contents with <command>ghc
                  --show-iface</command>. If you specify a directory for
                interface files, the <option>-ohidir</option> flag, then that
                affects <filename>hi-boot</filename> files
                too.</para>
            </listitem>
          </itemizedlist>
        </listitem>

   <listitem><para> If hs-boot files are considered distinct from their parent source
   files, and if a <literal>{-# SOURCE #-}</literal> import is considered to refer to the
   hs-boot file, then the module import graph must have no cycles.  The command
   <command>ghc -M</command> will report an error if a cycle is found.
   </para></listitem>

   <listitem><para> A module <literal>M</literal> that is 
   <literal>{-# SOURCE #-}</literal>-imported in a program will usually also be
   ordinarily imported elsewhere.  If not, <command>ghc --make</command>
   automatically adds <literal>M</literal> to the set of modules it tries to
   compile and link, to ensure that <literal>M</literal>'s implementation is included in
   the final program.
   </para></listitem>
</itemizedlist>
</para>
<para>
A hs-boot file need only contain the bare
      minimum of information needed to get the bootstrapping process
      started.  For example, it doesn't need to contain declarations
      for <emphasis>everything</emphasis> that module
      <literal>A</literal> exports, only the things required by the
      module(s) that import <literal>A</literal> recursively.</para>
<para>A hs-boot file is written in a subset of Haskell:
<itemizedlist>
<listitem><para> The module header (including the export list), and import statements, are exactly as in
Haskell, and so are the scoping rules.  
   Hence, to mention a non-Prelude type or class, you must import it.</para></listitem>
   
<listitem><para> There must be no value declarations, but there can be type signatures for
values.  For example:
<programlisting>
  double :: Int -&#62; Int
</programlisting>
</para></listitem>
<listitem><para> Fixity declarations are exactly as in Haskell.</para></listitem>
<listitem><para> Type synonym declarations are exactly as in Haskell.</para></listitem>
<listitem><para> A data type declaration can either be given in full, exactly as in Haskell, or it 
can be given abstractly, by omitting the '=' sign and everything that follows.  For example:
<programlisting>
  data T a b
</programlisting>
            In a <emphasis>source</emphasis> program
          this would declare TA to have no constructors (a GHC extension: see <xref linkend="nullary-types"/>),
          but in an hi-boot file it means "I don't know or care what the constructors are".
            This is the most common form of data type declaration, because it's easy to get right.
          You <emphasis>can</emphasis> also write out the constructors but, if you do so, you must write
          it out precisely as in its real definition.</para>
          <para>
            If you do not write out the constructors, you may need to give a kind 
            annotation (<xref linkend="kinding"/>), to tell
            GHC the kind of the type variable, if it is not "*".  (In source files, this is worked out
            from the way the type variable is used in the constructors.)  For example:
<programlisting>
  data R (x :: * -&#62; *) y
</programlisting>
You cannot use <literal>deriving</literal> on a data type declaration; write an
<literal>instance</literal> declaration instead.
</para></listitem>
<listitem><para> Class declarations is exactly as in Haskell, except that you may not put
default method declarations.  You can also omit all the superclasses and class
methods entirely; but you must either omit them all or put them all in.
</para></listitem>
<listitem><para> You can include instance declarations just as in Haskell; but omit the "where" part.
 </para></listitem>
</itemizedlist>
</para>
    </sect2>


    <sect2 id="using-make">
      <title>Using <command>make</command></title>

      <indexterm><primary><literal>make</literal></primary></indexterm>

      <para>It is reasonably straightforward to set up a
      <filename>Makefile</filename> to use with GHC, assuming you name
      your source files the same as your modules.  Thus:</para>

<programlisting>
HC      = ghc
HC_OPTS = -cpp $(EXTRA_HC_OPTS)

SRCS = Main.lhs Foo.lhs Bar.lhs
OBJS = Main.o   Foo.o   Bar.o

.SUFFIXES : .o .hs .hi .lhs .hc .s

cool_pgm : $(OBJS)
        rm -f $@
        $(HC) -o $@ $(HC_OPTS) $(OBJS)

# Standard suffix rules
.o.hi:
        @:

.lhs.o:
        $(HC) -c $&#60; $(HC_OPTS)

.hs.o:
        $(HC) -c $&#60; $(HC_OPTS)

.o-boot.hi-boot:
        @:

.lhs-boot.o-boot:
        $(HC) -c $&#60; $(HC_OPTS)

.hs-boot.o-boot:
        $(HC) -c $&#60; $(HC_OPTS)

# Inter-module dependencies
Foo.o Foo.hc Foo.s    : Baz.hi          # Foo imports Baz
Main.o Main.hc Main.s : Foo.hi Baz.hi   # Main imports Foo and Baz
</programlisting>

      <para>(Sophisticated <command>make</command> variants may
      achieve some of the above more elegantly.  Notably,
      <command>gmake</command>'s pattern rules let you write the more
      comprehensible:</para>

<programlisting>
%.o : %.lhs
        $(HC) -c $&#60; $(HC_OPTS)
</programlisting>

      <para>What we've shown should work with any
      <command>make</command>.)</para>

      <para>Note the cheesy <literal>.o.hi</literal> rule: It records
      the dependency of the interface (<filename>.hi</filename>) file
      on the source.  The rule says a <filename>.hi</filename> file
      can be made from a <filename>.o</filename> file by
      doing&hellip;nothing.  Which is true.</para>
      <para> Note that the suffix rules are all repeated twice, once
      for normal Haskell source files, and once for <filename>hs-boot</filename>
      files (see <xref linkend="mutual-recursion"/>).</para>

      <para>Note also the inter-module dependencies at the end of the
      Makefile, which take the form

<programlisting>
Foo.o Foo.hc Foo.s    : Baz.hi          # Foo imports Baz
</programlisting>

      They tell <command>make</command> that if any of
      <literal>Foo.o</literal>, <literal>Foo.hc</literal> or
      <literal>Foo.s</literal> have an earlier modification date than
      <literal>Baz.hi</literal>, then the out-of-date file must be
      brought up to date.  To bring it up to date,
      <literal>make</literal> looks for a rule to do so; one of the
      preceding suffix rules does the job nicely.  These dependencies
      can be generated automatically by <command>ghc</command>; see 
      <xref linkend="makefile-dependencies"/></para>

 </sect2>

      <sect2 id="makefile-dependencies">
        <title>Dependency generation</title>
        <indexterm><primary>dependencies in Makefiles</primary></indexterm>
        <indexterm><primary>Makefile dependencies</primary></indexterm>

        <para>Putting inter-dependencies of the form <literal>Foo.o :
        Bar.hi</literal> into your <filename>Makefile</filename> by
        hand is rather error-prone.  Don't worry, GHC has support for
        automatically generating the required dependencies.  Add the
        following to your <filename>Makefile</filename>:</para>

<programlisting>
depend :
        ghc -M $(HC_OPTS) $(SRCS)
</programlisting>

        <para>Now, before you start compiling, and any time you change
        the <literal>imports</literal> in your program, do
        <command>make depend</command> before you do <command>make
        cool&lowbar;pgm</command>.  The command <command>ghc -M</command> will
        append the needed dependencies to your
        <filename>Makefile</filename>.</para>

        <para>In general, <command>ghc -M Foo</command> does the following.
        For each module <literal>M</literal> in the set 
        <literal>Foo</literal> plus all its imports (transitively),
        it adds to the Makefile:
        <itemizedlist>
        <listitem><para>A line recording the dependence of the object file on the source file.
<programlisting>
M.o : M.hs
</programlisting>
(or <literal>M.lhs</literal> if that is the filename you used).
       </para></listitem>
       <listitem><para> For each import declaration <literal>import X</literal> in <literal>M</literal>,
       a line recording the dependence of <literal>M</literal> on <literal>X</literal>:
<programlisting>
M.o : X.hi
</programlisting></para></listitem>
       <listitem><para> For each import declaration <literal>import {-# SOURCE #-} X</literal> in <literal>M</literal>,
       a line recording the dependence of <literal>M</literal> on <literal>X</literal>:
<programlisting>
M.o : X.hi-boot
</programlisting>
       (See <xref linkend="mutual-recursion"/> for details of
       <literal>hi-boot</literal> style interface files.)
      </para></listitem>
        </itemizedlist> 
        If <literal>M</literal> imports multiple modules, then there will
       be multiple lines with <filename>M.o</filename> as the
       target.</para>
       <para>There is no need to list all of the source files as arguments to the <command>ghc -M</command> command;
       <command>ghc</command> traces the dependencies, just like <command>ghc --make</command>
       (a new feature in GHC 6.4).</para>

    <para>Note that <literal>ghc -M</literal> needs to find a <emphasis>source
        file</emphasis> for each module in the dependency graph, so that it can
      parse the import declarations and follow dependencies.  Any pre-compiled
      modules without source files must therefore belong to a
      package<footnote><para>This is a change in behaviour relative to 6.2 and
        earlier.</para>
      </footnote>.</para>

        <para>By default, <command>ghc -M</command> generates all the
        dependencies, and then concatenates them onto the end of
        <filename>makefile</filename> (or
        <filename>Makefile</filename> if <filename>makefile</filename>
        doesn't exist) bracketed by the lines "<literal>&num; DO NOT
        DELETE: Beginning of Haskell dependencies</literal>" and
        "<literal>&num; DO NOT DELETE: End of Haskell
        dependencies</literal>".  If these lines already exist in the
        <filename>makefile</filename>, then the old dependencies are
        deleted first.</para>

        <para>Don't forget to use the same <option>-package</option>
        options on the <literal>ghc -M</literal> command line as you
        would when compiling; this enables the dependency generator to
        locate any imported modules that come from packages.  The
        package modules won't be included in the dependencies
        generated, though (but see the
        <option>&ndash;&ndash;include-pkg-deps</option> option below).</para>

        <para>The dependency generation phase of GHC can take some
        additional options, which you may find useful.

        The options which affect dependency generation are:</para>

        <variablelist>
          <varlistentry>
            <term><option>-ddump-mod-cycles</option></term>
            <listitem>
              <para>Display a list of the cycles in the module graph. This is
          useful when trying to eliminate such cycles.</para>
            </listitem>
          </varlistentry>

          <varlistentry>
            <term><option>-v2</option></term>
            <listitem>
              <para>Print a full list of the module dependencies to stdout.
                    (This is the standard verbosity flag, so the list will
              also be displayed with <option>-v3</option> and
              <option>-v4</option>;
              <xref linkend ="options-help"/>.)</para>
            </listitem>
          </varlistentry>

          <varlistentry>
            <term><option>-dep-makefile</option> <replaceable>file</replaceable></term>
            <listitem>
              <para>Use <replaceable>file</replaceable> as the makefile,
              rather than <filename>makefile</filename> or
              <filename>Makefile</filename>.  If
              <replaceable>file</replaceable> doesn't exist,
              <command>mkdependHS</command> creates it.  We often use
              <option>-dep-makefile .depend</option> to put the dependencies in
              <filename>.depend</filename> and then
              <command>include</command> the file
              <filename>.depend</filename> into
              <filename>Makefile</filename>.</para>
            </listitem>
          </varlistentry>

          <varlistentry>
            <term><option>-dep-suffix &lt;suf&gt;</option></term>
            <listitem>
              <para>Make extra dependencies that declare that files
              with suffix
              <filename>.&lt;suf&gt;&lowbar;&lt;osuf&gt;</filename>
              depend on interface files with suffix
              <filename>.&lt;suf&gt;&lowbar;hi</filename>, or (for
              <literal>&lcub;-&num; SOURCE &num;-&rcub;</literal>
              imports) on <filename>.hi-boot</filename>.  Multiple
              <option>-dep-suffix</option> flags are permitted.  For example,
              <option>-dep-suffix a -dep-suffix b</option>
              will make dependencies
              for <filename>.hs</filename> on
              <filename>.hi</filename>,
              <filename>.a&lowbar;hs</filename> on
              <filename>.a&lowbar;hi</filename>, and
              <filename>.b&lowbar;hs</filename> on
              <filename>.b&lowbar;hi</filename>.  (Useful in
              conjunction with NoFib "ways".)</para>
            </listitem>
          </varlistentry>

          <varlistentry>
            <term><option>&ndash;&ndash;exclude-module=&lt;file&gt;</option></term>
            <listitem>
              <para>Regard <filename>&lt;file&gt;</filename> as
              "stable"; i.e., exclude it from having dependencies on
              it.</para>
            </listitem>
          </varlistentry>

          <varlistentry>
            <term><option>&ndash;&ndash;include-pkg-deps</option></term>
            <listitem>
              <para>Regard modules imported from packages as unstable,
              i.e., generate dependencies on any imported package modules
              (including <literal>Prelude</literal>, and all other
              standard Haskell libraries).  Dependencies are not traced
              recursively into packages; dependencies are only generated for
              home-package modules on external-package modules directly imported
              by the home package module.
              This option is normally
              only used by the various system libraries.</para>
            </listitem>
          </varlistentry>
        </variablelist>

    </sect2>

    <sect2 id="orphan-modules">
      <title>Orphan modules and instance declarations</title>

<para> Haskell specifies that when compiling module M, any instance
declaration in any module "below" M is visible.  (Module A is "below"
M if A is imported directly by M, or if A is below a module that M imports directly.)
In principle, GHC must therefore read the interface files of every module below M,
just in case they contain an instance declaration that matters to M.  This would
be a disaster in practice, so GHC tries to be clever. </para>

<para>In particular, if an instance declaration is in the same module as the definition
of any type or class mentioned in the <emphasis>head</emphasis> of the instance declaration 
(the part after the &ldquo;<literal>=&gt;</literal>&rdquo;; see <xref linkend="instance-rules"/>), then
GHC has to visit that interface file anyway.  Example:</para>
<programlisting>
  module A where
    instance C a =&gt; D (T a) where ...
    data T a = ...
</programlisting>
<para> The instance declaration is only relevant if the type T is in use, and if
so, GHC will have visited A's interface file to find T's definition. </para>

<para> The only problem comes when a module contains an instance declaration
and GHC has no other reason for visiting the module.  Example:
<programlisting>
  module Orphan where
    instance C a =&gt; D (T a) where ...
    class C a where ...
</programlisting>
Here, neither D nor T is declared in module Orphan.
We call such modules &ldquo;orphan modules&rdquo;.
GHC identifies orphan modules, and visits the interface file of
every orphan module below the module being compiled.  This is usually
wasted work, but there is no avoiding it.  You should therefore do
your best to have as few orphan modules as possible.
</para>
<para>
Functional dependencies complicate matters.  Suppose we have:
<programlisting>
  module B where
    instance E T Int where ...
    data T = ...
</programlisting>
Is this an orphan module?  Apparently not, because <literal>T</literal>
is declared in the same module.  But suppose class <literal>E</literal> had a
functional dependency:
<programlisting>
  module Lib where
    class E x y | y -> x where ...
</programlisting>
Then in some importing module M, the constraint <literal>(E a Int)</literal> should be "improved" by setting
<literal>a = T</literal>, <emphasis>even though there is no explicit mention
of <literal>T</literal> in M</emphasis>.</para>

These considerations lead to the following definition of an orphan module:
<itemizedlist>
  <listitem> <para> An <emphasis>orphan module</emphasis>
  <indexterm><primary>orphan module</primary></indexterm>
  contains at least one <emphasis>orphan instance</emphasis> or at
  least one <emphasis>orphan rule</emphasis>.</para> </listitem>

  <listitem><para> An instance declaration in a module M is an <emphasis>orphan instance</emphasis> if
  <indexterm><primary>orphan instance</primary></indexterm> 
<itemizedlist> 
<listitem><para>
  The class of the instance declaration is not declared in M, and
</para></listitem>
<listitem>
<para> <emphasis>Either</emphasis> the class has no functional dependencies, and none of the type constructors
        in the instance head is declared in M; <emphasis>or</emphasis> there
        is a functional dependency for which none of the type constructors mentioned
        in the <emphasis>non-determined</emphasis> part of the instance head is defined in M.
        </para></listitem>
  </itemizedlist>
  </para>
  <para> Only the instance head 
  counts.  In the example above, it is not good enough for C's declaration
  to be in module A; it must be the declaration of D or T.</para>
  </listitem>

  <listitem><para> A rewrite rule in a module M is an <emphasis>orphan rule</emphasis>
  <indexterm><primary>orphan rule</primary></indexterm>
  if none of the variables, type constructors,
  or classes that are free in the left hand side of the rule are declared in M.
  </para> </listitem>
 </itemizedlist>



<para>If you use the flag <option>-fwarn-orphans</option>, GHC will warn you 
if you are creating an orphan module.
Like any warning, you can switch the warning off with <option>-fno-warn-orphans</option>, 
and <option>-Werror</option>
will make the compilation fail if the warning is issued.
</para>
<para>
You can identify an orphan module by looking in its interface
file, <filename>M.hi</filename>, using the
<link linkend="modes"><option>--show-iface</option> mode</link>.  If there is a <literal>[orphan module]</literal> on the
first line, GHC considers it an orphan module.
</para>
</sect2>

  </sect1>

<!-- Emacs stuff:
     ;;; Local Variables: ***
     ;;; mode: xml ***
     ;;; sgml-parent-document: ("users_guide.xml" "book" "chapter" "sect1") ***
     ;;; End: ***
 -->