path: root/doc/user/separate.xml

<!--

  Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011 The SCons Foundation

  Permission is hereby granted, free of charge, to any person obtaining
  a copy of this software and associated documentation files (the
  "Software"), to deal in the Software without restriction, including
  without limitation the rights to use, copy, modify, merge, publish,
  distribute, sublicense, and/or sell copies of the Software, and to
  permit persons to whom the Software is furnished to do so, subject to
  the following conditions:

  The above copyright notice and this permission notice shall be included
  in all copies or substantial portions of the Software.

  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY
  KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
  WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
  LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
  OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
  WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

-->

<!--

=head1 Separating source and build trees

It's often desirable to keep any derived files from the build completely
separate from the source files. This makes it much easier to keep track of
just what is a source file, and also makes it simpler to handle B<variant>
builds, especially if you want the variant builds to co-exist.

=head2 Separating build and source directories using the Link command

Cons provides a simple mechanism that handles all of these requirements. The
C<Link> command is invoked as in this example:

  Link 'build' => 'src';

The specified directories are ``linked'' to the specified source
directory. Let's suppose that you setup a source directory, F<src>, with the
sub-directories F<world> and F<hello> below it, as in the previous
example. You could then substitute for the original build lines the
following:

  Build qw(
	build/world/Conscript
	build/hello/Conscript
  );

Notice that you treat the F<Conscript> file as if it existed in the build
directory. Now if you type the same command as before, you will get the
following results:

  % cons export
  Install build/world/world.h as export/include/world.h
  cc -Iexport/include -c build/hello/hello.c -o build/hello/hello.o
  cc -Iexport/include -c build/world/world.c -o build/world/world.o
  ar r build/world/libworld.a build/world/world.o
  ar: creating build/world/libworld.a
  ranlib build/world/libworld.a
  Install build/world/libworld.a as export/lib/libworld.a
  cc -o build/hello/hello build/hello/hello.o -Lexport/lib -lworld
  Install build/hello/hello as export/bin/hello

Again, Cons has taken care of the details for you. In particular, you will
notice that all the builds are done using source files and object files from
the build directory. For example, F<build/world/world.o> is compiled from
F<build/world/world.c>, and F<export/include/world.h> is installed from
F<build/world/world.h>. This is accomplished on most systems by the simple
expedient of ``hard'' linking the required files from each source directory
into the appropriate build directory.

The links are maintained correctly by Cons, no matter what you do to the
source directory. If you modify a source file, your editor may do this ``in
place'' or it may rename it first and create a new file. In the latter case,
any hard link will be lost. Cons will detect this condition the next time
the source file is needed, and will relink it appropriately.

You'll also notice, by the way, that B<no> changes were required to the
underlying F<Conscript> files. And we can go further, as we shall see in the
next section.

=head2 Explicit references to the source directory

When using the C<Link> command on some operating systems or with some
tool chains, it's sometimes useful to have a command actually use
the path name to the source directory, not the build directory.  For
example, on systems that must copy, not "hard link," the F<src/> and
F<build/> copies of C<Linked> files, using the F<src/> path of a file
name might make an editor aware that a syntax error must be fixed in the
source directory, not the build directory.

You can tell Cons that you want to use the "source path" for a file by
preceding the file name with a ``!'' (exclamation point).  For example,
if we add a ``!'' to the beginning of a source file:

  Program $env "foo", "!foo.c";	# Notice initial ! on foo.c

Cons will compile the target as follows:

  cc -c src/foo.c -o build/foo.o
  cc -o build/foo build/foo.o

Notice that Cons has compiled the program from the the F<src/foo.c>
source file.  Without the initial ``!'', Cons would have compiled the
program using the F<build/foo.c> path name.

-->

  <para>

  It's often useful to keep any built files completely
  separate from the source files.
  In &SCons;, this is usually done by creating one or more separate
  <emphasis>variant directory trees</emphasis>
  that are used to hold the built objects files, libraries,
  and executable programs, etc.
  for a specific flavor, or variant, of build.
  &SCons; provides two ways to do this,
  one through the &SConscript; function that we've already seen,
  and the second through a more flexible &VariantDir; function.

  </para>

  <para>

  One historical note:  the &VariantDir; function
  used to be called &BuildDir;.
  That name is still supported
  but has been deprecated
  because the &SCons; functionality
  differs from the model of a "build directory"
  implemented by other build systems like the GNU Autotools.

  </para>

  <section>
  <title>Specifying a Variant Directory Tree as Part of an &SConscript; Call</title>

    <para>

    The most straightforward way to establish a variant directory tree
    uses the fact that the usual way to
    set up a build hierarchy is to have an
    &SConscript; file in the source subdirectory.
    If you then pass a &variant_dir; argument to the
    &SConscript; function call:

    </para>

    <programlisting>
      SConscript('src/SConscript', variant_dir='build')
    </programlisting>

    <para>

    &SCons; will then build all of the files in
    the &build; subdirectory:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      SConscript  hello.c
      % <userinput>scons -Q</userinput>
      cc -o build/hello.o -c build/hello.c
      cc -o build/hello build/hello.o
      % <userinput>ls build</userinput>
      SConscript  hello  hello.c  hello.o
    </screen>

    <para>

    But wait a minute--what's going on here?
    &SCons; created the object file
    <filename>build/hello.o</filename>
    in the &build; subdirectory,
    as expected.
    But even though our &hello_c; file lives in the &src; subdirectory,
    &SCons; has actually compiled a
    <filename>build/hello.c</filename> file
    to create the object file.

    </para>

    <para>

    What's happened is that &SCons; has <emphasis>duplicated</emphasis>
    the &hello_c; file from the &src; subdirectory
    to the &build; subdirectory,
    and built the program from there.
    The next section explains why &SCons; does this.

    </para>

  </section>

  <section>
  <title>Why &SCons; Duplicates Source Files in a Variant Directory Tree</title>

    <para>

    &SCons; duplicates source files in variant directory trees
    because it's the most straightforward way to guarantee a correct build
    <emphasis>regardless of include-file directory paths,
    relative references between files,
    or tool support for putting files in different locations</emphasis>,
    and the &SCons; philosophy is to, by default,
    guarantee a correct build in all cases.

    </para>

    <para>

    The most direct reason to duplicate source files
    in variant directories
    is simply that some tools (mostly older versions)
    are written to only build their output files
    in the same directory as the source files.
    In this case, the choices are either
    to build the output file in the source directory
    and move it to the variant directory,
    or to duplicate the source files in the variant directory.

    </para>

    <para>

    Additionally,
    relative references between files
    can cause problems if we don't
    just duplicate the hierarchy of source files
    in the variant directory.
    You can see this at work in
    use of the C preprocessor <literal>#include</literal>
    mechanism with double quotes, not angle brackets:

    </para>

    <programlisting>
      #include "file.h"
    </programlisting>

    <para>

    The <emphasis>de facto</emphasis> standard behavior
    for most C compilers in this case
    is to first look in the same directory
    as the source file that contains the <literal>#include</literal> line,
    then to look in the directories in the preprocessor search path.
    Add to this that the &SCons; implementation of
    support for code repositories
    (described below)
    means not all of the files
    will be found in the same directory hierarchy,
    and the simplest way to make sure
    that the right include file is found
    is to duplicate the source files into the variant directory,
    which provides a correct build
    regardless of the original location(s) of the source files.

    </para>

    <para>

    Although source-file duplication guarantees a correct build
    even in these end-cases,
    it <emphasis>can</emphasis> usually be safely disabled.
    The next section describes
    how you can disable the duplication of source files
    in the variant directory.

    </para>

  </section>

  <section>
  <title>Telling &SCons; to Not Duplicate Source Files in the Variant Directory Tree</title>

    <para>

    In most cases and with most tool sets,
    &SCons; can place its target files in a build subdirectory
    <emphasis>without</emphasis>
    duplicating the source files
    and everything will work just fine.
    You can disable the default &SCons; behavior
    by specifying <literal>duplicate=0</literal>
    when you call the &SConscript; function:

    </para>

    <programlisting>
      SConscript('src/SConscript', variant_dir='build', duplicate=0)
    </programlisting>

    <para>

    When this flag is specified,
    &SCons; uses the variant directory
    like most people expect--that is,
    the output files are placed in the variant directory
    while the source files stay in the source directory:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      SConscript
      hello.c
      % <userinput>scons -Q</userinput>
      cc -c src/hello.c -o build/hello.o
      cc -o build/hello build/hello.o
      % <userinput>ls build</userinput>
      hello
      hello.o
    </screen>

  </section>

  <section>
  <title>The &VariantDir; Function</title>

    <para>

    Use the &VariantDir; function to establish that target
    files should be built in a separate directory
    from the source files:

    </para>

    <programlisting>
      VariantDir('build', 'src')
      env = Environment()
      env.Program('build/hello.c')
    </programlisting>

    <para>

    Note that when you're not using
    an &SConscript; file in the &src; subdirectory,
    you must actually specify that
    the program must be built from
    the <filename>build/hello.c</filename>
    file that &SCons; will duplicate in the
    &build; subdirectory.

    </para>

    <para>

    When using the &VariantDir; function directly,
    &SCons; still duplicates the source files
    in the variant directory by default:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      hello.c
      % <userinput>scons -Q</userinput>
      cc -o build/hello.o -c build/hello.c
      cc -o build/hello build/hello.o
      % <userinput>ls build</userinput>
      hello  hello.c  hello.o
    </screen>

    <para>

    You can specify the same <literal>duplicate=0</literal> argument
    that you can specify for an &SConscript; call:

    </para>

    <programlisting>
      VariantDir('build', 'src', duplicate=0)
      env = Environment()
      env.Program('build/hello.c')
    </programlisting>

    <para>

    In which case &SCons;
    will disable duplication of the source files:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      hello.c
      % <userinput>scons -Q</userinput>
      cc -o build/hello.o -c src/hello.c
      cc -o build/hello build/hello.o
      % <userinput>ls build</userinput>
      hello  hello.o
    </screen>

  </section>

  <section>
  <title>Using &VariantDir; With an &SConscript; File</title>

    <para>

    Even when using the &VariantDir; function,
    it's much more natural to use it with
    a subsidiary &SConscript; file.
    For example, if the
    <filename>src/SConscript</filename>
    looks like this:

    </para>

    <programlisting>
      env = Environment()
      env.Program('hello.c')
    </programlisting>

    <para>

    Then our &SConstruct; file could look like:

    </para>

    
    <programlisting>
      VariantDir('build', 'src')
      SConscript('build/SConscript')
      </programlisting>

    <para>

    Yielding the following output:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      SConscript  hello.c
      % <userinput>scons -Q</userinput>
      cc -o build/hello.o -c build/hello.c
      cc -o build/hello build/hello.o
      % <userinput>ls build</userinput>
      SConscript  hello  hello.c  hello.o
    </screen>

    <para>

    Notice that this is completely equivalent
    to the use of &SConscript; that we
    learned about in the previous section.

    </para>

  </section>

  <section>
  <title>Using &Glob; with &VariantDir;</title>

    <para>

    The &Glob; file name pattern matching function
    works just as usual when using &VariantDir;.
    For example, if the
    <filename>src/SConscript</filename>
    looks like this:

    </para>

    <programlisting>
      env = Environment()
      env.Program('hello', Glob('*.c'))
    </programlisting>

    <para>

    Then with the same &SConstruct; file as in the previous section,
    and source files <filename>f1.c</filename>
    and <filename>f2.c</filename> in src,
    we would see the following output:

    </para>

    <screen>
      % <userinput>ls src</userinput>
      SConscript  f1.c  f2.c  f2.h
      % <userinput>scons -Q</userinput>
      cc -o build/f1.o -c build/f1.c
      cc -o build/f2.o -c build/f2.c
      cc -o build/hello build/f1.o build/f2.o
      % <userinput>ls build</userinput>
      SConscript  f1.c  f1.o  f2.c  f2.h  f2.o  hello
    </screen>

    <para>

    The &Glob; function returns Nodes in the
    <filename>build/</filename> tree, as you'd expect.

    </para>

  </section>

  <!--

  <section>
  <title>Why You'd Want to Call &VariantDir; Instead of &SConscript;</title>

    <para>

    XXX why call VariantDir() instead of SConscript(variant_dir=)

    </para>

  </section>

  -->