llvm/docs/CompilerDriver.html
2008-12-11 17:34:48 +00:00

421 lines
28 KiB
HTML

<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<meta name="generator" content="Docutils 0.4: http://docutils.sourceforge.net/" />
<title>Customizing LLVMC: Reference Manual</title>
<link rel="stylesheet" href="llvm.css" type="text/css" />
</head>
<body>
<div class="document" id="customizing-llvmc-reference-manual">
<div class="doc_title">Customizing LLVMC: Reference Manual</div>
<div class="doc_warning">
<p>Note: This document is a work-in-progress. Additions and clarifications
are welcome.</p>
</div>
<p>LLVMC is a generic compiler driver, designed to be customizable and
extensible. It plays the same role for LLVM as the <tt class="docutils literal"><span class="pre">gcc</span></tt> program
does for GCC - LLVMC's job is essentially to transform a set of input
files into a set of targets depending on configuration rules and user
options. What makes LLVMC different is that these transformation rules
are completely customizable - in fact, LLVMC knows nothing about the
specifics of transformation (even the command-line options are mostly
not hard-coded) and regards the transformation structure as an
abstract graph. This makes it possible to adapt LLVMC for other
purposes - for example, as a build tool for game resources.</p>
<p>Because LLVMC employs TableGen <a class="footnote-reference" href="#id2" id="id1" name="id1">[1]</a> as its configuration language, you
need to be familiar with it to customize LLVMC.</p>
<div class="contents topic">
<ul class="simple">
<li><a class="reference" href="#compiling-with-llvmc" id="id3" name="id3">Compiling with LLVMC</a></li>
<li><a class="reference" href="#predefined-options" id="id4" name="id4">Predefined options</a></li>
<li><a class="reference" href="#customizing-llvmc-the-compilation-graph" id="id5" name="id5">Customizing LLVMC: the compilation graph</a></li>
<li><a class="reference" href="#writing-a-tool-description" id="id6" name="id6">Writing a tool description</a></li>
<li><a class="reference" href="#option-list-specifying-all-options-in-a-single-place" id="id7" name="id7">Option list - specifying all options in a single place</a></li>
<li><a class="reference" href="#using-hooks-and-environment-variables-in-the-cmd-line-property" id="id8" name="id8">Using hooks and environment variables in the <tt class="docutils literal"><span class="pre">cmd_line</span></tt> property</a></li>
<li><a class="reference" href="#conditional-evaluation-the-case-expression" id="id9" name="id9">Conditional evaluation: the <tt class="docutils literal"><span class="pre">case</span></tt> expression</a></li>
<li><a class="reference" href="#language-map" id="id10" name="id10">Language map</a></li>
<li><a class="reference" href="#references" id="id11" name="id11">References</a></li>
</ul>
</div>
<div class="doc_author">Written by Mikhail Glushenkov</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id3" id="compiling-with-llvmc" name="compiling-with-llvmc">Compiling with LLVMC</a></div>
<p>LLVMC tries hard to be as compatible with <tt class="docutils literal"><span class="pre">gcc</span></tt> as possible,
although there are some small differences. Most of the time, however,
you shouldn't be able to notice them:</p>
<pre class="literal-block">
$ # This works as expected:
$ llvmc2 -O3 -Wall hello.cpp
$ ./a.out
hello
</pre>
<p>One nice feature of LLVMC is that one doesn't have to distinguish
between different compilers for different languages (think <tt class="docutils literal"><span class="pre">g++</span></tt> and
<tt class="docutils literal"><span class="pre">gcc</span></tt>) - the right toolchain is chosen automatically based on input
language names (which are, in turn, determined from file
extensions). If you want to force files ending with &quot;.c&quot; to compile as
C++, use the <tt class="docutils literal"><span class="pre">-x</span></tt> option, just like you would do it with <tt class="docutils literal"><span class="pre">gcc</span></tt>:</p>
<pre class="literal-block">
$ llvmc2 -x c hello.cpp
$ # hello.cpp is really a C file
$ ./a.out
hello
</pre>
<p>On the other hand, when using LLVMC as a linker to combine several C++
object files you should provide the <tt class="docutils literal"><span class="pre">--linker</span></tt> option since it's
impossible for LLVMC to choose the right linker in that case:</p>
<pre class="literal-block">
$ llvmc2 -c hello.cpp
$ llvmc2 hello.o
[A lot of link-time errors skipped]
$ llvmc2 --linker=c++ hello.o
$ ./a.out
hello
</pre>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id4" id="predefined-options" name="predefined-options">Predefined options</a></div>
<p>LLVMC has some built-in options that can't be overridden in the
configuration files:</p>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">-o</span> <span class="pre">FILE</span></tt> - Output file name.</li>
<li><tt class="docutils literal"><span class="pre">-x</span> <span class="pre">LANGUAGE</span></tt> - Specify the language of the following input files
until the next -x option.</li>
<li><tt class="docutils literal"><span class="pre">-v</span></tt> - Enable verbose mode, i.e. print out all executed commands.</li>
<li><tt class="docutils literal"><span class="pre">--view-graph</span></tt> - Show a graphical representation of the compilation
graph. Requires that you have <tt class="docutils literal"><span class="pre">dot</span></tt> and <tt class="docutils literal"><span class="pre">gv</span></tt> commands
installed. Hidden option, useful for debugging.</li>
<li><tt class="docutils literal"><span class="pre">--write-graph</span></tt> - Write a <tt class="docutils literal"><span class="pre">compilation-graph.dot</span></tt> file in the
current directory with the compilation graph description in the
Graphviz format. Hidden option, useful for debugging.</li>
<li><tt class="docutils literal"><span class="pre">--save-temps</span></tt> - Write temporary files to the current directory
and do not delete them on exit. Hidden option, useful for debugging.</li>
<li><tt class="docutils literal"><span class="pre">--help</span></tt>, <tt class="docutils literal"><span class="pre">--help-hidden</span></tt>, <tt class="docutils literal"><span class="pre">--version</span></tt> - These options have
their standard meaning.</li>
</ul>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id5" id="customizing-llvmc-the-compilation-graph" name="customizing-llvmc-the-compilation-graph">Customizing LLVMC: the compilation graph</a></div>
<p>At the time of writing LLVMC does not support on-the-fly reloading of
configuration, so to customize LLVMC you'll have to recompile the
source code (which lives under <tt class="docutils literal"><span class="pre">$LLVM_DIR/tools/llvmc2</span></tt>). The
default configuration files are <tt class="docutils literal"><span class="pre">Common.td</span></tt> (contains common
definitions, don't forget to <tt class="docutils literal"><span class="pre">include</span></tt> it in your configuration
files), <tt class="docutils literal"><span class="pre">Tools.td</span></tt> (tool descriptions) and <tt class="docutils literal"><span class="pre">Graph.td</span></tt> (compilation
graph definition).</p>
<p>To compile LLVMC with your own configuration file (say,``MyGraph.td``),
run <tt class="docutils literal"><span class="pre">make</span></tt> like this:</p>
<pre class="literal-block">
$ cd $LLVM_DIR/tools/llvmc2
$ make GRAPH=MyGraph.td TOOLNAME=my_llvmc
</pre>
<p>This will build an executable named <tt class="docutils literal"><span class="pre">my_llvmc</span></tt>. There are also
several sample configuration files in the <tt class="docutils literal"><span class="pre">llvmc2/examples</span></tt>
subdirectory that should help to get you started.</p>
<p>Internally, LLVMC stores information about possible source
transformations in form of a graph. Nodes in this graph represent
tools, and edges between two nodes represent a transformation path. A
special &quot;root&quot; node is used to mark entry points for the
transformations. LLVMC also assigns a weight to each edge (more on
this later) to choose between several alternative edges.</p>
<p>The definition of the compilation graph (see file <tt class="docutils literal"><span class="pre">Graph.td</span></tt>) is
just a list of edges:</p>
<pre class="literal-block">
def CompilationGraph : CompilationGraph&lt;[
Edge&lt;root, llvm_gcc_c&gt;,
Edge&lt;root, llvm_gcc_assembler&gt;,
...
Edge&lt;llvm_gcc_c, llc&gt;,
Edge&lt;llvm_gcc_cpp, llc&gt;,
...
OptionalEdge&lt;llvm_gcc_c, opt, [(switch_on &quot;opt&quot;)]&gt;,
OptionalEdge&lt;llvm_gcc_cpp, opt, [(switch_on &quot;opt&quot;)]&gt;,
...
OptionalEdge&lt;llvm_gcc_assembler, llvm_gcc_cpp_linker,
(case (input_languages_contain &quot;c++&quot;), (inc_weight),
(or (parameter_equals &quot;linker&quot;, &quot;g++&quot;),
(parameter_equals &quot;linker&quot;, &quot;c++&quot;)), (inc_weight))&gt;,
...
]&gt;;
</pre>
<p>As you can see, the edges can be either default or optional, where
optional edges are differentiated by sporting a <tt class="docutils literal"><span class="pre">case</span></tt> expression
used to calculate the edge's weight.</p>
<p>The default edges are assigned a weight of 1, and optional edges get a
weight of 0 + 2*N where N is the number of tests that evaluated to
true in the <tt class="docutils literal"><span class="pre">case</span></tt> expression. It is also possible to provide an
integer parameter to <tt class="docutils literal"><span class="pre">inc_weight</span></tt> and <tt class="docutils literal"><span class="pre">dec_weight</span></tt> - in this case,
the weight is increased (or decreased) by the provided value instead
of the default 2.</p>
<p>When passing an input file through the graph, LLVMC picks the edge
with the maximum weight. To avoid ambiguity, there should be only one
default edge between two nodes (with the exception of the root node,
which gets a special treatment - there you are allowed to specify one
default edge <em>per language</em>).</p>
<p>To get a visual representation of the compilation graph (useful for
debugging), run <tt class="docutils literal"><span class="pre">llvmc2</span> <span class="pre">--view-graph</span></tt>. You will need <tt class="docutils literal"><span class="pre">dot</span></tt> and
<tt class="docutils literal"><span class="pre">gsview</span></tt> installed for this to work properly.</p>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id6" id="writing-a-tool-description" name="writing-a-tool-description">Writing a tool description</a></div>
<p>As was said earlier, nodes in the compilation graph represent tools,
which are described separately. A tool definition looks like this
(taken from the <tt class="docutils literal"><span class="pre">Tools.td</span></tt> file):</p>
<pre class="literal-block">
def llvm_gcc_cpp : Tool&lt;[
(in_language &quot;c++&quot;),
(out_language &quot;llvm-assembler&quot;),
(output_suffix &quot;bc&quot;),
(cmd_line &quot;llvm-g++ -c $INFILE -o $OUTFILE -emit-llvm&quot;),
(sink)
]&gt;;
</pre>
<p>This defines a new tool called <tt class="docutils literal"><span class="pre">llvm_gcc_cpp</span></tt>, which is an alias for
<tt class="docutils literal"><span class="pre">llvm-g++</span></tt>. As you can see, a tool definition is just a list of
properties; most of them should be self-explanatory. The <tt class="docutils literal"><span class="pre">sink</span></tt>
property means that this tool should be passed all command-line
options that lack explicit descriptions.</p>
<p>The complete list of the currently implemented tool properties follows:</p>
<ul class="simple">
<li>Possible tool properties:<ul>
<li><tt class="docutils literal"><span class="pre">in_language</span></tt> - input language name.</li>
<li><tt class="docutils literal"><span class="pre">out_language</span></tt> - output language name.</li>
<li><tt class="docutils literal"><span class="pre">output_suffix</span></tt> - output file suffix.</li>
<li><tt class="docutils literal"><span class="pre">cmd_line</span></tt> - the actual command used to run the tool. You can
use <tt class="docutils literal"><span class="pre">$INFILE</span></tt> and <tt class="docutils literal"><span class="pre">$OUTFILE</span></tt> variables, output redirection
with <tt class="docutils literal"><span class="pre">&gt;</span></tt>, hook invocations (<tt class="docutils literal"><span class="pre">$CALL</span></tt>), environment variables
(via <tt class="docutils literal"><span class="pre">$ENV</span></tt>) and the <tt class="docutils literal"><span class="pre">case</span></tt> construct (more on this below).</li>
<li><tt class="docutils literal"><span class="pre">join</span></tt> - this tool is a &quot;join node&quot; in the graph, i.e. it gets a
list of input files and joins them together. Used for linkers.</li>
<li><tt class="docutils literal"><span class="pre">sink</span></tt> - all command-line options that are not handled by other
tools are passed to this tool.</li>
</ul>
</li>
</ul>
<p>The next tool definition is slightly more complex:</p>
<pre class="literal-block">
def llvm_gcc_linker : Tool&lt;[
(in_language &quot;object-code&quot;),
(out_language &quot;executable&quot;),
(output_suffix &quot;out&quot;),
(cmd_line &quot;llvm-gcc $INFILE -o $OUTFILE&quot;),
(join),
(prefix_list_option &quot;L&quot;, (forward),
(help &quot;add a directory to link path&quot;)),
(prefix_list_option &quot;l&quot;, (forward),
(help &quot;search a library when linking&quot;)),
(prefix_list_option &quot;Wl&quot;, (unpack_values),
(help &quot;pass options to linker&quot;))
]&gt;;
</pre>
<p>This tool has a &quot;join&quot; property, which means that it behaves like a
linker. This tool also defines several command-line options: <tt class="docutils literal"><span class="pre">-l</span></tt>,
<tt class="docutils literal"><span class="pre">-L</span></tt> and <tt class="docutils literal"><span class="pre">-Wl</span></tt> which have their usual meaning. An option has two
attributes: a name and a (possibly empty) list of properties. All
currently implemented option types and properties are described below:</p>
<ul>
<li><p class="first">Possible option types:</p>
<blockquote>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">switch_option</span></tt> - a simple boolean switch, for example <tt class="docutils literal"><span class="pre">-time</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">parameter_option</span></tt> - option that takes an argument, for example
<tt class="docutils literal"><span class="pre">-std=c99</span></tt>;</li>
<li><tt class="docutils literal"><span class="pre">parameter_list_option</span></tt> - same as the above, but more than one
occurence of the option is allowed.</li>
<li><tt class="docutils literal"><span class="pre">prefix_option</span></tt> - same as the parameter_option, but the option name
and parameter value are not separated.</li>
<li><tt class="docutils literal"><span class="pre">prefix_list_option</span></tt> - same as the above, but more than one
occurence of the option is allowed; example: <tt class="docutils literal"><span class="pre">-lm</span> <span class="pre">-lpthread</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">alias_option</span></tt> - a special option type for creating
aliases. Unlike other option types, aliases are not allowed to
have any properties besides the aliased option name. Usage
example: <tt class="docutils literal"><span class="pre">(alias_option</span> <span class="pre">&quot;preprocess&quot;,</span> <span class="pre">&quot;E&quot;)</span></tt></li>
</ul>
</blockquote>
</li>
<li><p class="first">Possible option properties:</p>
<blockquote>
<ul class="simple">
<li><tt class="docutils literal"><span class="pre">append_cmd</span></tt> - append a string to the tool invocation command.</li>
<li><tt class="docutils literal"><span class="pre">forward</span></tt> - forward this option unchanged.</li>
<li><tt class="docutils literal"><span class="pre">output_suffix</span></tt> - modify the output suffix of this
tool. Example : <tt class="docutils literal"><span class="pre">(switch</span> <span class="pre">&quot;E&quot;,</span> <span class="pre">(output_suffix</span> <span class="pre">&quot;i&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">stop_compilation</span></tt> - stop compilation after this phase.</li>
<li><tt class="docutils literal"><span class="pre">unpack_values</span></tt> - used for for splitting and forwarding
comma-separated lists of options, e.g. <tt class="docutils literal"><span class="pre">-Wa,-foo=bar,-baz</span></tt> is
converted to <tt class="docutils literal"><span class="pre">-foo=bar</span> <span class="pre">-baz</span></tt> and appended to the tool invocation
command.</li>
<li><tt class="docutils literal"><span class="pre">help</span></tt> - help string associated with this option. Used for
<tt class="docutils literal"><span class="pre">--help</span></tt> output.</li>
<li><tt class="docutils literal"><span class="pre">required</span></tt> - this option is obligatory.</li>
</ul>
</blockquote>
</li>
</ul>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id7" id="option-list-specifying-all-options-in-a-single-place" name="option-list-specifying-all-options-in-a-single-place">Option list - specifying all options in a single place</a></div>
<p>It can be handy to have all information about options gathered in a
single place to provide an overview. This can be achieved by using a
so-called <tt class="docutils literal"><span class="pre">OptionList</span></tt>:</p>
<pre class="literal-block">
def Options : OptionList&lt;[
(switch_option &quot;E&quot;, (help &quot;Help string&quot;)),
(alias_option &quot;quiet&quot;, &quot;q&quot;)
...
]&gt;;
</pre>
<p><tt class="docutils literal"><span class="pre">OptionList</span></tt> is also a good place to specify option aliases.</p>
<p>Tool-specific option properties like <tt class="docutils literal"><span class="pre">append_cmd</span></tt> have (obviously)
no meaning in the context of <tt class="docutils literal"><span class="pre">OptionList</span></tt>, so the only properties
allowed there are <tt class="docutils literal"><span class="pre">help</span></tt> and <tt class="docutils literal"><span class="pre">required</span></tt>.</p>
<p>Option lists are used at the file scope. See file
<tt class="docutils literal"><span class="pre">examples/Clang.td</span></tt> for an example of <tt class="docutils literal"><span class="pre">OptionList</span></tt> usage.</p>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id8" id="using-hooks-and-environment-variables-in-the-cmd-line-property" name="using-hooks-and-environment-variables-in-the-cmd-line-property">Using hooks and environment variables in the <tt class="docutils literal"><span class="pre">cmd_line</span></tt> property</a></div>
<p>Normally, LLVMC executes programs from the system <tt class="docutils literal"><span class="pre">PATH</span></tt>. Sometimes,
this is not sufficient: for example, we may want to specify tool names
in the configuration file. This can be achieved via the mechanism of
hooks - to compile LLVMC with your hooks, just drop a .cpp file into
<tt class="docutils literal"><span class="pre">tools/llvmc2</span></tt> directory. Hooks should live in the <tt class="docutils literal"><span class="pre">hooks</span></tt>
namespace and have the signature <tt class="docutils literal"><span class="pre">std::string</span> <span class="pre">hooks::MyHookName</span>
<span class="pre">(void)</span></tt>. They can be used from the <tt class="docutils literal"><span class="pre">cmd_line</span></tt> tool property:</p>
<pre class="literal-block">
(cmd_line &quot;$CALL(MyHook)/path/to/file -o $CALL(AnotherHook)&quot;)
</pre>
<p>It is also possible to use environment variables in the same manner:</p>
<pre class="literal-block">
(cmd_line &quot;$ENV(VAR1)/path/to/file -o $ENV(VAR2)&quot;)
</pre>
<p>To change the command line string based on user-provided options use
the <tt class="docutils literal"><span class="pre">case</span></tt> expression (documented below):</p>
<pre class="literal-block">
(cmd_line
(case
(switch_on &quot;E&quot;),
&quot;llvm-g++ -E -x c $INFILE -o $OUTFILE&quot;,
(default),
&quot;llvm-g++ -c -x c $INFILE -o $OUTFILE -emit-llvm&quot;))
</pre>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id9" id="conditional-evaluation-the-case-expression" name="conditional-evaluation-the-case-expression">Conditional evaluation: the <tt class="docutils literal"><span class="pre">case</span></tt> expression</a></div>
<p>The 'case' construct can be used to calculate weights of the optional
edges and to choose between several alternative command line strings
in the <tt class="docutils literal"><span class="pre">cmd_line</span></tt> tool property. It is designed after the
similarly-named construct in functional languages and takes the form
<tt class="docutils literal"><span class="pre">(case</span> <span class="pre">(test_1),</span> <span class="pre">statement_1,</span> <span class="pre">(test_2),</span> <span class="pre">statement_2,</span> <span class="pre">...</span> <span class="pre">(test_N),</span>
<span class="pre">statement_N)</span></tt>. The statements are evaluated only if the corresponding
tests evaluate to true.</p>
<p>Examples:</p>
<pre class="literal-block">
// Increases edge weight by 5 if &quot;-A&quot; is provided on the
// command-line, and by 5 more if &quot;-B&quot; is also provided.
(case
(switch_on &quot;A&quot;), (inc_weight 5),
(switch_on &quot;B&quot;), (inc_weight 5))
// Evaluates to &quot;cmdline1&quot; if option &quot;-A&quot; is provided on the
// command line, otherwise to &quot;cmdline2&quot;
(case
(switch_on &quot;A&quot;), &quot;cmdline1&quot;,
(switch_on &quot;B&quot;), &quot;cmdline2&quot;,
(default), &quot;cmdline3&quot;)
</pre>
<p>Note the slight difference in 'case' expression handling in contexts
of edge weights and command line specification - in the second example
the value of the <tt class="docutils literal"><span class="pre">&quot;B&quot;</span></tt> switch is never checked when switch <tt class="docutils literal"><span class="pre">&quot;A&quot;</span></tt> is
enabled, and the whole expression always evaluates to <tt class="docutils literal"><span class="pre">&quot;cmdline1&quot;</span></tt> in
that case.</p>
<p>Case expressions can also be nested, i.e. the following is legal:</p>
<pre class="literal-block">
(case (switch_on &quot;E&quot;), (case (switch_on &quot;o&quot;), ..., (default), ...)
(default), ...)
</pre>
<p>You should, however, try to avoid doing that because it hurts
readability. It is usually better to split tool descriptions and/or
use TableGen inheritance instead.</p>
<ul class="simple">
<li>Possible tests are:<ul>
<li><tt class="docutils literal"><span class="pre">switch_on</span></tt> - Returns true if a given command-line option is
provided by the user. Example: <tt class="docutils literal"><span class="pre">(switch_on</span> <span class="pre">&quot;opt&quot;)</span></tt>. Note that
you have to define all possible command-line options separately in
the tool descriptions. See the next doc_text for the discussion of
different kinds of command-line options.</li>
<li><tt class="docutils literal"><span class="pre">parameter_equals</span></tt> - Returns true if a command-line parameter equals
a given value. Example: <tt class="docutils literal"><span class="pre">(parameter_equals</span> <span class="pre">&quot;W&quot;,</span> <span class="pre">&quot;all&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">element_in_list</span></tt> - Returns true if a command-line parameter list
includes a given value. Example: <tt class="docutils literal"><span class="pre">(parameter_in_list</span> <span class="pre">&quot;l&quot;,</span> <span class="pre">&quot;pthread&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">input_languages_contain</span></tt> - Returns true if a given language
belongs to the current input language set. Example:
<tt class="docutils literal"><span class="pre">`(input_languages_contain</span> <span class="pre">&quot;c++&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">in_language</span></tt> - Evaluates to true if the language of the input
file equals to the argument. Valid only when using <tt class="docutils literal"><span class="pre">case</span></tt>
expression in a <tt class="docutils literal"><span class="pre">cmd_line</span></tt> tool property. Example:
<tt class="docutils literal"><span class="pre">`(in_language</span> <span class="pre">&quot;c++&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">not_empty</span></tt> - Returns true if a given option (which should be
either a parameter or a parameter list) is set by the
user. Example: <tt class="docutils literal"><span class="pre">`(not_empty</span> <span class="pre">&quot;o&quot;)</span></tt>.</li>
<li><tt class="docutils literal"><span class="pre">default</span></tt> - Always evaluates to true. Should always be the last
test in the <tt class="docutils literal"><span class="pre">case</span></tt> expression.</li>
<li><tt class="docutils literal"><span class="pre">and</span></tt> - A standard logical combinator that returns true iff all
of its arguments return true. Used like this: <tt class="docutils literal"><span class="pre">(and</span> <span class="pre">(test1),</span>
<span class="pre">(test2),</span> <span class="pre">...</span> <span class="pre">(testN))</span></tt>. Nesting of <tt class="docutils literal"><span class="pre">and</span></tt> and <tt class="docutils literal"><span class="pre">or</span></tt> is allowed,
but not encouraged.</li>
<li><tt class="docutils literal"><span class="pre">or</span></tt> - Another logical combinator that returns true only if any
one of its arguments returns true. Example: <tt class="docutils literal"><span class="pre">(or</span> <span class="pre">(test1),</span>
<span class="pre">(test2),</span> <span class="pre">...</span> <span class="pre">(testN))</span></tt>.</li>
</ul>
</li>
</ul>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id10" id="language-map" name="language-map">Language map</a></div>
<p>One last thing that you will need to modify when adding support for a
new language to LLVMC is the language map, which defines mappings from
file extensions to language names. It is used to choose the proper
toolchain(s) for a given input file set. Language map definition is
located in the file <tt class="docutils literal"><span class="pre">Tools.td</span></tt> and looks like this:</p>
<pre class="literal-block">
def LanguageMap : LanguageMap&lt;
[LangToSuffixes&lt;&quot;c++&quot;, [&quot;cc&quot;, &quot;cp&quot;, &quot;cxx&quot;, &quot;cpp&quot;, &quot;CPP&quot;, &quot;c++&quot;, &quot;C&quot;]&gt;,
LangToSuffixes&lt;&quot;c&quot;, [&quot;c&quot;]&gt;,
...
]&gt;;
</pre>
</div>
<div class="doc_text">
<div class="doc_section"><a class="toc-backref" href="#id11" id="references" name="references">References</a></div>
<table class="docutils footnote" frame="void" id="id2" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id1" name="id2">[1]</a></td><td>TableGen Fundamentals
<a class="reference" href="http://llvm.cs.uiuc.edu/docs/TableGenFundamentals.html">http://llvm.cs.uiuc.edu/docs/TableGenFundamentals.html</a></td></tr>
</tbody>
</table>
</div>
</div>
<hr />
<address>
<a href="http://jigsaw.w3.org/css-validator/check/referer"><img
src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS" /></a>
<a href="http://validator.w3.org/check/referer"><img
src="http://www.w3.org/Icons/valid-xhtml10" alt="Valid XHTML 1.0!" /></a>
<a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br/>
Last modified: $Date$
</address>
</body>
</html>