1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" 2 "http://www.w3.org/TR/html4/strict.dtd"> 3<html> 4<head> 5<title>Introduction to the Clang AST</title> 6<link type="text/css" rel="stylesheet" href="../menu.css" /> 7<link type="text/css" rel="stylesheet" href="../content.css" /> 8</head> 9<body> 10 11<!--#include virtual="../menu.html.incl"--> 12 13<div id="content"> 14 15<h1>Introduction to the Clang AST</h1> 16<p>This document gives a gentle introduction to the mysteries of the Clang AST. 17It is targeted at developers who either want to contribute to Clang, or use 18tools that work based on Clang's AST, like the AST matchers.</p> 19<!-- FIXME: Add link once we have an AST matcher document --> 20 21<!-- ======================================================================= --> 22<h2 id="intro">Introduction</h2> 23<!-- ======================================================================= --> 24 25<p>Clang's AST is different from ASTs produced by some other compilers in that it closely 26resembles both the written C++ code and the C++ standard. For example, 27parenthesis expressions and compile time constants are available in an unreduced 28form in the AST. This makes Clang's AST a good fit for refactoring tools.</p> 29 30<p>Documentation for all Clang AST nodes is available via the generated 31<a href="http://clang.llvm.org/doxygen">Doxygen</a>. The doxygen online 32documentation is also indexed by your favorite search engine, which will make 33a search for clang and the AST node's class name usually turn up the doxygen 34of the class you're looking for (for example, search for: clang ParenExpr).</p> 35 36<!-- ======================================================================= --> 37<h2 id="examine">Examining the AST</h2> 38<!-- ======================================================================= --> 39 40<p>A good way to familarize yourself with the Clang AST is to actually look 41at it on some simple example code. Clang has a builtin AST-dump modes, which 42can be enabled with the flags -ast-dump and -ast-dump-xml. Note that -ast-dump-xml 43currently only works with debug-builds of clang.</p> 44 45<p>Let's look at a simple example AST:</p> 46<pre> 47# cat test.cc 48int f(int x) { 49 int result = (x / 42); 50 return result; 51} 52 53# Clang by default is a frontend for many tools; -cc1 tells it to directly 54# use the C++ compiler mode. -undef leaves out some internal declarations. 55$ clang -cc1 -undef -ast-dump-xml test.cc 56... cutting out internal declarations of clang ... 57<TranslationUnit ptr="0x4871160"> 58 <Function ptr="0x48a5800" name="f" prototype="true"> 59 <FunctionProtoType ptr="0x4871de0" canonical="0x4871de0"> 60 <BuiltinType ptr="0x4871250" canonical="0x4871250"/> 61 <parameters> 62 <BuiltinType ptr="0x4871250" canonical="0x4871250"/> 63 </parameters> 64 </FunctionProtoType> 65 <ParmVar ptr="0x4871d80" name="x" initstyle="c"> 66 <BuiltinType ptr="0x4871250" canonical="0x4871250"/> 67 </ParmVar> 68 <Stmt> 69(CompoundStmt 0x48a5a38 <t2.cc:1:14, line:4:1> 70 (DeclStmt 0x48a59c0 <line:2:3, col:24> 71 0x48a58c0 "int result = 72 (ParenExpr 0x48a59a0 <col:16, col:23> 'int' 73 (BinaryOperator 0x48a5978 <col:17, col:21> 'int' '/' 74 (ImplicitCastExpr 0x48a5960 <col:17> 'int' <LValueToRValue> 75 (DeclRefExpr 0x48a5918 <col:17> 'int' lvalue ParmVar 0x4871d80 'x' 'int')) 76 (IntegerLiteral 0x48a5940 <col:21> 'int' 42)))") 77 (ReturnStmt 0x48a5a18 <line:3:3, col:10> 78 (ImplicitCastExpr 0x48a5a00 <col:10> 'int' <LValueToRValue> 79 (DeclRefExpr 0x48a59d8 <col:10> 'int' lvalue Var 0x48a58c0 'result' 'int')))) 80 81 </Stmt> 82 </Function> 83</TranslationUnit> 84</pre> 85<p>In general, -ast-dump-xml dumps declarations in an XML-style format and 86statements in an S-expression-style format. 87The toplevel declaration in a translation unit is always the 88<a href="http://clang.llvm.org/doxygen/classclang_1_1TranslationUnitDecl.html">translation unit declaration</a>. 89In this example, our first user written declaration is the 90<a href="http://clang.llvm.org/doxygen/classclang_1_1FunctionDecl.html">function declaration</a> 91of 'f'. The body of 'f' is a <a href="http://clang.llvm.org/doxygen/classclang_1_1CompoundStmt.html">compound statement</a>, 92whose child nodes are a <a href="http://clang.llvm.org/doxygen/classclang_1_1DeclStmt.html">declaration statement</a> 93that declares our result variable, and the 94<a href="http://clang.llvm.org/doxygen/classclang_1_1ReturnStmt.html">return statement</a>.</p> 95 96<!-- ======================================================================= --> 97<h2 id="context">AST Context</h2> 98<!-- ======================================================================= --> 99 100<p>All information about the AST for a translation unit is bundled up in the class 101<a href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html">ASTContext</a>. 102It allows traversal of the whole translation unit starting from 103<a href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html#abd909fb01ef10cfd0244832a67b1dd64">getTranslationUnitDecl</a>, 104or to access Clang's <a href="http://clang.llvm.org/doxygen/classclang_1_1ASTContext.html#a4f95adb9958e22fbe55212ae6482feb4">table of identifiers</a> 105for the parsed translation unit.</p> 106 107<!-- ======================================================================= --> 108<h2 id="nodes">AST Nodes</h2> 109<!-- ======================================================================= --> 110 111<p>Clang's AST nodes are modeled on a class hierarchy that does not have a common 112ancestor. Instead, there are multiple larger hierarchies for basic node types like 113<a href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a> and 114<a href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>. Many 115important AST nodes derive from <a href="http://clang.llvm.org/doxygen/classclang_1_1Type.html">Type</a>, 116<a href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a>, 117<a href="http://clang.llvm.org/doxygen/classclang_1_1DeclContext.html">DeclContext</a> or 118<a href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>, 119with some classes deriving from both Decl and DeclContext.</p> 120<p>There are also a multitude of nodes in the AST that are not part of a 121larger hierarchy, and are only reachable from specific other nodes, 122like <a href="http://clang.llvm.org/doxygen/classclang_1_1CXXBaseSpecifier.html">CXXBaseSpecifier</a>. 123</p> 124 125<p>Thus, to traverse the full AST, one starts from the <a href="http://clang.llvm.org/doxygen/classclang_1_1TranslationUnitDecl.html">TranslationUnitDecl</a> 126and then recursively traverses everything that can be reached from that node 127- this information has to be encoded for each specific node type. This algorithm 128is encoded in the <a href="http://clang.llvm.org/doxygen/classclang_1_1RecursiveASTVisitor.html">RecursiveASTVisitor</a>. 129See the <a href="http://clang.llvm.org/docs/RAVFrontendAction.html">RecursiveASTVisitor tutorial</a>.</p> 130 131<p>The two most basic nodes in the Clang AST are statements (<a href="http://clang.llvm.org/doxygen/classclang_1_1Stmt.html">Stmt</a>) 132and declarations (<a href="http://clang.llvm.org/doxygen/classclang_1_1Decl.html">Decl</a>). 133Note that expressions (<a href="http://clang.llvm.org/doxygen/classclang_1_1Expr.html">Expr</a>) 134are also statements in Clang's AST.</p> 135 136</div> 137</body> 138</html> 139 140