2
0
mirror of https://github.com/boostorg/build.git synced 2026-02-11 11:42:14 +00:00
Files
build/doc/src/reference.xml
Vladimir Prus 8091df49bb link-runtime -> runtime-link.
Patch from Craig Rodrigues.


[SVN r29765]
2005-06-24 09:42:16 +00:00

1071 lines
41 KiB
XML

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE chapter PUBLIC "-//Boost//DTD BoostBook XML V1.0//EN"
"http://www.boost.org/tools/boostbook/dtd/boostbook.dtd">
<chapter id="bbv2.reference">
<title>Detailed reference</title>
<section id="bbv2.reference.general">
<title>General information</title>
<section id="bbv2.reference.init">
<title>Initialization</title>
<para>bjam's first job upon startup is to load the Jam code that
implements the build system. To do this, it searches for a file
called <filename>boost-build.jam</filename>, first in the invocation directory, then
in its parent and so forth up to the filesystem root, and finally
in the directories specified by the environment variable
BOOST_BUILD_PATH. When found, the file is interpreted, and should
specify the build system location by calling the boost-build
rule:</para>
<programlisting>
rule boost-build ( location ? )
</programlisting>
<para>
If location is a relative path, it is treated as relative to
the directory of <filename>boost-build.jam</filename>. The directory specified by
that location and the directories in BOOST_BUILD_PATH are then searched for
a file called <filename>bootstrap.jam</filename>, which is expected to
bootstrap the build system. This arrangement allows the build
system to work without any command-line or environment variable
settings. For example, if the build system files were located in a
directory "build-system/" at your project root, you might place a
<filename>boost-build.jam</filename> at the project root containing:
<programlisting>
boost-build build-system ;
</programlisting>
In this case, running bjam anywhere in the project tree will
automatically find the build system.</para>
<para>The default <filename>bootstrap.jam</filename>, after loading some standard
definitions, loads two files, which can be provided/customised by
user: <filename>site-config.jam</filename> and <filename>user-config.jam</filename>.</para>
<para>Locations where those files a search are summarized below:</para>
<table id="bbv2.reference.init.config">
<title>Search paths for configuration files</title>
<tgroup cols="3">
<thead>
<row>
<entry></entry>
<entry>site-config.jam</entry>
<entry>user-config.jam</entry>
</row>
</thead>
<tbody>
<row>
<entry>Linux</entry>
<entry>
<simpara><code>/etc</code></simpara>
<simpara><code>$HOME</code></simpara>
<simpara><code>$BOOST_BUILD_PATH</code></simpara>
</entry>
<entry>
<simpara><code>$HOME</code></simpara>
<simpara><code>$BOOST_BUILD_PATH</code></simpara>
</entry>
</row>
<row>
<entry>Windows</entry>
<entry>
<simpara><code>%SystemRoot%</code></simpara>
<simpara><code>%HOMEDRIVE%%HOMEPATH%</code></simpara>
<simpara><code>%HOME%</code></simpara>
<simpara><code>%BOOST_BUILD_PATH%</code></simpara>
</entry>
<entry>
<simpara><code>%HOMEDRIVE%%HOMEPATH%</code></simpara>
<simpara><code>%HOME%</code></simpara>
<simpara><code>%BOOST_BUILD_PATH%</code></simpara>
</entry>
</row>
</tbody>
</tgroup>
</table>
<para>
Boost.Build comes with default versions of those files,
<!-- Where are those files installed? The user can't use them as templates unless she can find them -->
which can serve as templates for customized versions.
</para>
</section>
<section id="bbv2.reference.commandline">
<title>Command line</title>
<para>The command line may contain:</para>
<itemizedlist>
<listitem><simpara>Jam options,</simpara></listitem>
<listitem><simpara>Boost.Build <link linkend=
"bbv2.reference.init.options">options</link>,</simpara></listitem>
<listitem><simpara>Command line arguments</simpara></listitem>
</itemizedlist>
<section id="bbv2.reference.init.args">
<title>Command line arguments</title>
<para>
Command line arguments specify targets and build
request using the following rules.
</para>
<itemizedlist>
<listitem>
<simpara>
An argument that does not contain slashes or the <code>=</code>
symbol is either a value of an implicit feature or of a target to
be built. It is taken to be value of a feature if an appropriate
feature exists. Otherwise, it is considered a <link linkend=
"bbv2.reference.ids">target id</link>. Building the
special target name “clean” has the same effect as
using the <code>--clean</code> option.
</simpara>
</listitem>
<listitem>
<para>
An argument containing either slashes or
the <code>=</code> symbol specifies a number of build
request elements (see <xref
linkend="bbv2.reference.buildreq"/>). In its simplest
form, it's just a set of properties, separated by
slashes, which become a single build request element,
for example:
<programlisting>
borland/&lt;runtime-link&gt;static
</programlisting>
A more complex form can be used to save typing. For example,
instead of
<programlisting>
borland/runtime-link=static borland/runtime-link=dynamic
</programlisting>
one can use
<programlisting>
borland/runtime-link=static,dynamic
</programlisting>
Exactly, the conversion from argument to build request
elements is performed by (1) splitting the argument at each slash,
(2) converting each split part into a set of properties and (3)
taking all possible combinations
<!-- Be specific. Do you mean the cross-product? -->
of the property sets. Each split
part should have the either the form
<programlisting>
<emphasis>feature-name</emphasis>=<emphasis>feature-value1</emphasis>[","<emphasis>feature-valueN</emphasis>]*
</programlisting>
or, in case of implicit features
<programlisting>
<emphasis>feature-value1</emphasis>[","<emphasis>feature-valueN</emphasis>;]*
</programlisting>
will be converted into the property set
<programlisting>
&lt;feature-name&gt;feature-value1 .... &lt;feature-name&gt;feature-valueN
</programlisting>
<!-- There's absolutely no explanation of how arguments are combined. Fix that. -->
</para>
</listitem>
</itemizedlist>
<para>
For example, the command line
<programlisting>
target1 debug gcc/runtime-link=dynamic,static
</programlisting>
would cause target called <literal>target1</literal> to be rebuilt in
debug mode, except that for gcc, both dynamically and statically
linked binaries would be created.
</para>
</section>
<section id="bbv2.reference.init.options">
<title>Command line options</title>
<para>All of the Boost.Build options start with the "--" prefix.
They are described in the following table.</para>
<table>
<title>Command line options</title>
<tgroup cols="2">
<thead>
<row>
<entry>Option</entry>
<entry>Description</entry>
</row>
</thead>
<tbody>
<row>
<entry><literal>--version</literal></entry>
<entry>Prints information on Boost.Build and Boost.Jam
versions.</entry>
</row>
<row id="bbv2.reference.init.options.help">
<entry><literal>--help</literal></entry>
<entry>Access to the online help system. This prints general
information on how to use the help system with additional
--help* options.</entry>
</row>
<row>
<entry><literal>--clean</literal></entry>
<entry>Removes everything instead of building. Unlike
<literal>clean</literal> target in make, it is possible to clean only
some targets.</entry>
</row>
<row>
<entry><literal>--debug</literal></entry>
<entry>Enables internal checks.</entry>
</row>
<row>
<entry><literal>--dump-projects</literal></entry>
<entry>Cause the project structure to be output.</entry>
</row>
<row>
<entry><literal>--no-error-backtrace</literal></entry>
<entry>Don't print backtrace on errors. Primary useful for
testing.</entry>
</row>
<row>
<entry><literal>--ignore-config</literal></entry>
<entry>Do not load <literal>site-config.jam</literal> and
<literal>user-config.jam</literal></entry>
</row>
</tbody>
</tgroup>
</table>
</section>
</section>
</section>
<section id="bbv2.reference.jamfiles">
<title>Writing Jamfiles</title>
<para>This section describes specific information about writing Jamfiles.</para>
<section id="bbv2.reference.headers">
<title>Generated headers</title>
<para>Usually, Boost.Build handles implicit dependendies completely
automatically. For example, for C++ files, all <literal>#include</literal>
statements are found and handled. The only aspect where user help
might be needed is implicit dependency on generated files.</para>
<para>By default, Boost.Build handles such dependencies within one
main target. For example, assume that main target "app" has two
sources, "app.cpp" and "parser.y". The latter source is converted
into "parser.c" and "parser.h". Then, if "app.cpp" includes
"parser.h", Boost.Build will detect this dependency. Moreover,
since "parser.h" will be generated into a build directory, the
path to that directory will automatically added to include
path.</para>
<para>Making this mechanism work across main target boundaries is
possible, but imposes certain overhead. For that reason, if
there's implicit dependency on files from other main targets, the
<literal>&lt;implicit-dependency&gt;</literal> [ link ] feature must
be used, for example:</para>
<programlisting>
lib parser : parser.y ;
exe app : app.cpp : &lt;implicit-dependency&gt;parser ;
</programlisting>
<para>
The above example tells the build system that when scanning
all sources of "app" for implicit-dependencies, it should consider
targets from "parser" as potential dependencies.
</para>
</section>
</section>
<section id="bbv2.reference.buildprocess">
<title>Build process</title>
<para>The general overview of the build process was given in the
<link linkend="bbv2.advanced.build_process">user documentation</link>.
This section provides additional details, and some specific rules.
</para>
<para>To recap, building a target with specific properties includes the
following steps:
<orderedlist>
<listitem><para>applying default build,</para></listitem>
<listitem><para>selecting the main target alternative to use,
</para></listitem>
<listitem><para>determining "common" properties</para></listitem>
<listitem><para>building targets referred by the sources list and
dependency properties</para></listitem>
<listitem><para>adding the usage requirements produces when building
dependencies to the "common" properties</para></listitem>
<listitem><para>building the target using generators</para></listitem>
<listitem><para>computing the usage requirements to be returned</para></listitem>
</orderedlist>
</para>
<section id="bbv2.reference.buildprocess.alternatives">
<title>Alternative selection</title>
<para>When there are several alternatives, one of them must be
selected. The process is as follows:</para>
<orderedlist>
<listitem>
<simpara>
For each alternative <emphasis>condition</emphasis> is defined
as the set of base properies in requirements. [Note: it might be
better to specify the condition explicitly, as in
conditional requirements].
</simpara>
</listitem>
<listitem>
<simpara>
An alternative is viable only if all properties in condition
are present in build request.
</simpara>
</listitem>
<listitem>
<simpara>
If there's one viable alternative, it's choosen. Otherwise,
an attempt is made to find one best alternative. An alternative
a is better than another alternative b, iff set of properties
in b's condition is strict subset of the set of properities of
'a's condition. If there's one viable alternative, which is
better than all other, it's selected. Otherwise, an error is
reported.
</simpara>
</listitem>
</orderedlist>
</section>
<section id="bbv2.reference.buildprocess.common">
<title>Determining common properties</title>
<para>The "common" properties is a somewhat artificial term. Those are
the intermediate property set from which both the build request for
dependencies and properties for building the target are derived.
</para>
<para>Since default build and alternatives are already handled, we have
only two inputs: build requests and requirements. Here are the rules
about common properties.
</para>
<orderedlist>
<listitem><para>Non-free feature can have only one
value</para></listitem>
<listitem><para>A non-conditional property in requirement in always
present in common properties.</para></listitem>
<listitem><para>A property in build request is present in
common properties, unless (2) tells otherwise.</para></listitem>
<listitem><para>If either build request, or requirements (non-conditional
or conditional) include an expandable property (either composite,
or property with specified subfeature value), the behaviour is
equivalent to explicitly adding all expanded properties to build
request or requirements.</para></listitem>
<listitem><para>If requirements include a conditional property, and
condiiton of this property is true in context of common
properties, then the conditional property should be in common
properties as well.</para></listitem>
<listitem><para>If no value for a feature is given by other rules
here, it has default value in common properties.</para></listitem>
</orderedlist>
<para>Those rules are declarative, they don't specify how to compute the
common properties. However, they provide enough information for the
user. The important point is the handling of conditional
requirements. The condition can be satisfied either by property in
build request, by non-conditional requirements, or even by another
conditional property. For example, the following example works as
expected:
<programlisting>
exe a : a.cpp
: &lt;toolset&gt;gcc:&lt;variant&gt;release
&lt;variant&gt;release:&lt;define&gt;FOO ;
</programlisting>
</para>
</section>
</section>
<section id="bbv2.reference.definitions">
<title>Definitions</title>
<section id="bbv2.reference.features">
<title>Features and properties</title>
<para>A <emphasis>feature</emphasis> is a normalized (toolset-independent)
aspect of a build configuration, such as whether inlining is
enabled. Feature names may not contain the '<literal>&gt;</literal>'
character.</para>
<!--
And what about dash?
-->
<para>Each feature in a build configuration has one or more
associated <emphasis>value</emphasis>s. Feature values for non-free features
may not contain the '<literal>&lt;</literal>', '<literal>:</literal>', or
'<literal>=</literal>' characters. Feature values for free features may not
contain the '<literal>&lt;</literal>' character.</para>
<para>A <emphasis>property</emphasis> is a (feature,value) pair, expressed as
&lt;feature&gt;value.</para>
<para>A <emphasis>subfeature</emphasis> is a feature that only exists in the
presence of its parent feature, and whose identity can be derived
(in the context of its parent) from its value. A subfeature's
parent can never be another subfeature. Thus, features and their
subfeatures form a two-level hierarchy.</para>
<para>A <emphasis>value-string</emphasis> for a feature <emphasis role="bold">F</emphasis> is a string of
the form
<literal>value-subvalue1-subvalue2</literal>...<literal>-subvalueN</literal>, where
<literal>value</literal> is a legal value for <emphasis role="bold">F</emphasis> and
<literal>subvalue1</literal>...<literal>subvalueN</literal> are legal values of some
of <emphasis role="bold">F</emphasis>'s subfeatures. For example, the properties
<literal>&lt;toolset&gt;gcc &lt;toolset-version&gt;3.0.1</literal> can be
expressed more conscisely using a value-string, as
<literal>&lt;toolset&gt;gcc-3.0.1</literal>.</para>
<para>A <emphasis>property set</emphasis> is a set of properties (i.e. a
collection without duplicates), for instance:
<literal>&lt;toolset&gt;gcc &lt;runtime-link&gt;static</literal>.</para>
<para>A <emphasis>property path</emphasis> is a property set whose elements have
been joined into a single string separated by slashes. A property
path representation of the previous example would be
<literal>&lt;toolset&gt;gcc/&lt;runtime-link&gt;static</literal>.</para>
<para>A <emphasis>build specification</emphasis> is a property set that fully
describes the set of features used to build a target.</para>
<section id="bbv2.reference.features.validity">
<title>Property Validity</title>
<para>
For <link linkend=
"bbv2.reference.features.attributes.free">free</link>
features, all values are valid. For all other features,
the valid values are explicitly specified, and the build
system will report an error for the use of an invalid
feature-value. Subproperty validity may be restricted so
that certain values are valid only in the presence of
certain other subproperties. For example, it is possible
to specify that the <code>&lt;gcc-target&gt;mingw</code>
property is only valid in the presence of
<code>&lt;gcc-version&gt;2.95.2</code>.
</para>
</section>
<section id="bbv2.reference.features.attributes">
<title>Feature Attributes</title>
<para>Each feature has a collection of zero or more of the following
attributes. Feature attributes are low-level descriptions of how the
build system should interpret a feature's values when they appear in
a build request. We also refer to the attributes of properties, so
that an <emphasis>incidental</emphasis> property, for example, is
one whose feature has the <emphasis>incidental</emphasis>
attribute.</para>
<itemizedlist>
<listitem>
<para><emphasis>incidental</emphasis></para>
<para>Incidental features are assumed not to affect build
products at all. As a consequence, the build system may use
the same file for targets whose build specification differs
only in incidental features. A feature that controls a
compiler's warning level is one example of a likely
incidental feature.</para>
<para>Non-incidental features are assumed to affect build
products, so the files for targets whose build specification
differs in non-incidental features are placed in different
directories as described in "target paths" below. [ where? ]
</para>
</listitem>
<listitem>
<para>
<anchor id="bbv2.reference.features.attributes.propagated"/>
<emphasis>propagated</emphasis>
</para>
<para>Features of this kind are
propagated to dependencies. That is, if a <link linkend=
"bbv2.advanced.targets.main">main target</link> is built using a
propagated
property, the build systems attempts to use the same property
when building any of its dependencies as part of that main
target. For instance, when an optimized exectuable is
requested, one usually wants it to be linked with optimized
libraries. Thus, the <literal>&lt;optimization&gt;</literal> feature is
propagated.</para>
</listitem>
<listitem>
<para>
<anchor id="bbv2.reference.features.attributes.free"/>
<emphasis>free</emphasis>
</para>
<para>Most features have a finite set of allowed values, and can
only take on a single value from that set in a given build
specification. Free features, on the other hand, can have
several values at a time and each value can be an arbitrary
string. For example, it is possible to have several
preprocessor symbols defined simultaneously:</para>
<programlisting>
&lt;define&gt;NDEBUG=1 &lt;define&gt;HAS_CONFIG_H=1
</programlisting>
</listitem>
<listitem>
<para><emphasis>optional</emphasis></para>
<para>An optional feature is a feature that is not required to
appear in a build specification. Every non-optional non-free
feature has a default value that is used when a value for
the feature is not otherwise specified, either in a target's
requirements or in the user's build request. [A feature's
default value is given by the first value listed in the
feature's declaration. -- move this elsewhere - dwa]</para>
</listitem>
<listitem>
<para><emphasis>symmetric</emphasis></para>
<para>A symmetric feature's default value is not automatically
included in <link linkend=
"bbv2.reference.variants">build variants</link>. Normally
a feature only generates a subvariant directory when its
value differs from the value specified by the build variant,
leading to an assymmetric subvariant directory structure for
certain values of the feature. A symmetric feature, when
relevant to the toolset, always generates a corresponding
subvariant directory.</para>
</listitem>
<listitem>
<para><emphasis>path</emphasis></para>
<para>The value of a path feature specifies a path. The path is
treated as relative to the directory of Jamfile where path
feature is used and is translated appropriately by the build
system when the build is invoked from a different
directory</para>
</listitem>
<listitem>
<para><emphasis>implicit</emphasis></para>
<para>Values of implicit features alone identify the feature.
For example, a user is not required to write
"&lt;toolset&gt;gcc", but can simply write "gcc". Implicit
feature names also don't appear in variant paths, although
the values do. Thus: bin/gcc/... as opposed to
bin/toolset-gcc/.... There should typically be only a few
such features, to avoid possible name clashes.</para>
</listitem>
<listitem>
<para><emphasis>composite</emphasis></para>
<para>Composite features actually correspond to groups of
properties. For example, a build variant is a composite
feature. When generating targets from a set of build
properties, composite features are recursively expanded and
<emphasis>added</emphasis> to the build property set, so rules can find
them if necessary. Non-composite non-free features override
components of composite features in a build property set.</para>
</listitem>
<listitem>
<para><emphasis>dependency</emphasis></para>
<para>The value of dependency feature if a target reference.
When used for building of a main target, the value of
dependency feature is treated as additional dependency.</para>
<para>For example, dependency features allow to state that
library A depends on library B. As the result, whenever an
application will link to A, it will also link to B.
Specifying B as dependency of A is different from adding B to
the sources of A. <!-- Need to clarify this. --></para>
</listitem>
</itemizedlist>
<para>Features that are neither free nor incidental are called
<emphasis>base</emphasis> features.</para>
</section>
<section id="bbv2.reference.features.declaration">
<title>Feature Declaration</title>
<para>The low-level feature declaration interface is the
<literal>feature</literal> rule from the
<literal>feature</literal> module:
<programlisting>
rule feature ( name : allowed-values * : attributes * )
</programlisting>
A feature's allowed-values may be extended with the
<code>feature.extend</code> rule.
</para>
</section>
</section>
<section id="bbv2.reference.variants">
<title>Build Variants</title>
<para>
A build variant, or (simply variant) is a special kind of composite
feature that automatically incorporates the default values of
features that . Typically you'll want at least two separate
variants: one for debugging, and one for your release code. [
Volodya says: "Yea, we'd need to mention that it's a composite
feature and describe how they are declared, in pacticular that
default values of non-optional features are incorporated into
build variant automagically. Also, do we wan't some variant
inheritance/extension/templates. I don't remember how it works in
V1, so can't document this for V2.". Will clean up soon -DWA ]
</para>
</section>
<section id="bbv2.reference.variants.proprefine">
<title>Property refinement</title>
<para>When a target with certain properties is requested, and that
target requires some set of properties, it is needed to find the
set of properties to use for building. This process is called
<emphasis>property refinement</emphasis> and is performed by these rules</para>
<orderedlist>
<listitem>
<simpara>
Each property in the required set is added to the original
property set
</simpara>
</listitem>
<listitem>
<simpara>
If the original property set includes property with a different
value of non free feature, that property is removed.
</simpara>
</listitem>
</orderedlist>
</section>
<section id="bbv2.reference.variants.propcond">
<title>Conditional properties</title>
<para>Sometime it's desirable to apply certain requirements only for
a specific combination of other properties. For example, one of
compilers that you use issues a pointless warning that you want to
suppress by passing a command line option to it. You would not
want to pass that option to other compilers. Conditional
properties allow you to do just that. Their syntax is:</para>
<programlisting>
property ( "," property ) * ":" property
</programlisting>
<para>
For example, the problem above would be solved by:
<programlisting>
exe hello : hello.cpp : &lt;toolset&gt;yfc:&lt;cxxflags&gt;-disable-pointless-warning ;
</programlisting>
</para>
<para>The syntax also allows several properties in the condition, for
example:
<programlisting>
exe hello : hello.cpp : &lt;os&gt;NT,&lt;toolset&gt;gcc:&lt;link&gt;static ;
</programlisting>
</para>
</section>
<section id="bbv2.reference.ids">
<title>Target identifiers and references</title>
<para><emphasis>Target identifier</emphasis> is used to denote a
target. The syntax is:</para>
<programlisting>
target-id -&gt; (project-id | target-name | file-name )
| (project-id | directory-name) "//" target-name
project-id -&gt; path
target-name -&gt; path
file-name -&gt; path
directory-name -&gt; path
</programlisting>
<para>
This grammar allows some elements to be recognized as either
<itemizedlist>
<listitem>
<simpara>
project id (at this point, all project ids start with slash).
</simpara>
</listitem>
<listitem>
<simpara>
name of target declared in current Jamfile (note that target
names may include slash).
</simpara>
</listitem>
<listitem>
<simpara>
a regular file, denoted by absolute name or name relative to
project's sources location.
</simpara>
</listitem>
</itemizedlist>
To determine the real meaning a check is made if project-id
by the specified name exists, and then if main target of that
name exists. For example, valid target ids might be:
<screen>
a -- target in current project
lib/b.cpp -- regular file
/boost/thread -- project "/boost/thread"
/home/ghost/build/lr_library//parser -- target in specific project
</screen>
</para>
<para><emphasis role="bold">Rationale:</emphasis>Target is separated from project by special
separator (not just slash), because:</para>
<itemizedlist>
<listitem>
<simpara>
It emphasises that projects and targets are different things.
</simpara>
</listitem>
<listitem>
<simpara>
It allows to have main target names with slashes.
<!-- The motivation for which is:
So, to summarize:
1. The project that extract tarfile may extract all possible kinds
of targets, and it's reasonable to use them directly from other
project.
2. The rule for unpacking tar is inplemented in terms of
"patch-file", for maintainability, and therefore, must use main
target name that contains slashes?
3. Using sub-Jamfile in "foo" to declare extracted file "foo/b" is
not an option, because you should not change existing tree
That makes good rationale for why main target must contain names.
-->
</simpara>
</listitem>
</itemizedlist>
<para id="bbv2.reference.targets.references">
<emphasis>Target reference</emphasis> is used to
specify a source target, and may additionally specify desired
properties for that target. It has this syntax:</para>
<programlisting>
target-reference -&gt; target-id [ "/" requested-properties ]
requested-properties -&gt; property-path
</programlisting>
<para>
For example,
<programlisting>
exe compiler : compiler.cpp libs/cmdline/&lt;optimization&gt;space ;
</programlisting>
would cause the version of <literal>cmdline</literal> library,
optimized for space, to be linked in even if the
<literal>compiler</literal> executable is build with optimization for
speed.
</para>
</section>
</section>
<section id="bbv2.reference.generators">
<title>Generators</title>
<warning><para>The information is this section is likely to be outdated
and misleading.
</para></warning>
<para>To construct a main target with given properties from sources,
it is required to create a dependency graph for that main target,
which will also include actions to be run. The algorithm for
creating the dependency graph is described here.</para>
<para>The fundamental concept is <emphasis>generator</emphasis>. If encapsulates
the notion of build tool and is capable to converting a set of
input targets into a set of output targets, with some properties.
Generator matches a build tool as closely as possible: it works
only when the tool can work with requested properties (for
example, msvc compiler can't work when requested toolset is gcc),
and should produce exactly the same targets as the tool (for
example, if Borland's linker produces additional files with debug
information, generator should also).</para>
<para>Given a set of generators, the fundamental operation is to
construct a target of a given type, with given properties, from a
set of targets. That operation is performed by rule
<literal>generators.construct</literal> and the used algorithm is described
below.</para>
<section>
<title>Selecting and ranking viable generators</title>
<para>Each generator, in addition to target types that it can
produce, have attribute that affects its applicability in
particular sitiation. Those attributes are:</para>
<orderedlist>
<listitem>
<simpara>
Required properties, which are properties absolutely
necessary for the generator to work. For example, generator
encapsulating the gcc compiler would have &lt;toolset&gt;gcc as
required property.
</simpara>
</listitem>
<listitem>
<simpara>
Optional properties, which increase the generators
suitability for a particual build.
</simpara>
</listitem>
</orderedlist>
<para>
Generator's required and optional properties may not include
either free or incidental properties. (Allowing this would
greatly complicate caching targets).
</para>
<para>When trying to construct a target, the first step is to select
all possible generators for the requested target type, which
required properties are a subset of requested properties.
Generators that were already selected up the call stack are
excluded. In addition, if any composing generators were selected
up the call stack, all other composing generators are ignored
(TODO: define composing generators). The found generators
are assigned a rank, which is the number of optional properties
present in requested properties. Finally, generators with highest
rank are selected for futher processing.</para>
</section>
<section>
<title>Running generators</title>
<para>When generators are selected, each is run to produce a list of
created targets. This list might include targets that are not of
requested types, because generators create the same targets as
some tool, and tool's behaviour is fixed. (Note: should specify
that in some cases we actually want extra targets). If generator
fails, it returns an empty list. Generator is free to call
'construct' again, to convert sources to the types it can handle.
It also can pass modified properties to 'construct'. However, a
generator is not allowed to modify any propagated properties,
otherwise when actually consuming properties we might discover
that the set of propagated properties is different from what was
used for building sources.</para>
<para>For all targets that are not of requested types, we try to
convert them to requested type, using a second call to
<literal>construct</literal>. This is done in order to support
transformation sequences where single source file expands to
several later. See <ulink url=
"http://groups.yahoo.com/group/jamboost/message/1667">this
message</ulink> for details.</para>
</section>
<section>
<title>Selecting dependency graph</title>
<para>
After all generators are run,
it is necessary to decide which of successfull invocation will be
taken as final result. At the moment, this is not done. Instead,
it is checked whether all successfull generator invocation
returned the same target list. Error is issued otherwise.
</para>
</section>
<section>
<title>Property adjustment</title>
<para>Because target location is determined by the build system, it
is sometimes necessary to adjust properties, in order to not
break actions. For example, if there's an action that generates
a header, say "a_parser.h", and a source file "a.cpp" which
includes that file, we must make everything work as if a_parser.h
is generated in the same directory where it would be generated
without any subvariants.</para>
<para>Correct property adjustment can be done only after all targets
are created, so the approach taken is:</para>
<orderedlist>
<listitem>
<para>
When dependency graph is constructed, each action can be
assigned a rule for property adjustment.
</para>
</listitem>
<listitem>
<para>
When virtual target is actualized, that rule is run and
return the final set of properties. At this stage it can use
information of all created virtual targets.
</para>
</listitem>
</orderedlist>
<para>In case of quoted includes, no adjustment can give 100% correct
results. If target dirs are not changed by build system, quoted
includes are searched in "." and then in include path, while angle
includes are searched only in include path. When target dirs are
changed, we'd want to make quoted includes to be search in "." then in
additional dirs and then in the include path and make angle includes
be searched in include path, probably with additional paths added at
some position. Unless, include path already has "." as the first
element, this is not possible. So, either generated headers should not
be included with quotes, or first element of include path should be
".", which essentially erases the difference between quoted and angle
includes. <emphasis role="bold">Note:</emphasis> the only way to get
"." as include path into compiler command line is via verbatim
compiler option. In all other case, Boost.Build will convert "." into
directory where it occurs.</para>
</section>
<section>
<title>Transformations cache</title>
<para>
Under certain conditions, an
attempt is made to cache results of transformation search. First,
the sources are replaced with targets with special name and the
found target list is stored. Later, when properties, requested
type, and source type are the same, the store target list is
retrieved and cloned, with appropriate change in names.
</para>
</section>
</section>
</chapter>
<!--
Local Variables:
mode: xml
sgml-indent-data: t
sgml-parent-document: ("userman.xml" "chapter")
sgml-set-face: t
End:
-->