Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 34c0533a005ad5be0356dbb913ab3a05fb6b3ca5 / . / docs / InternalsManual.html
blob: 42b3b784d92f753b186e7d4b6245e4cca7b659c4 [file] [log] [blame]
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]1<html>
2<head>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]3<title>"Clang" CFE Internals Manual</title>
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]4<link type="text/css" rel="stylesheet" href="../menu.css" />
5<link type="text/css" rel="stylesheet" href="../content.css" />
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]6<style type="text/css">
7td {
8 vertical-align: top;
9}
10</style>
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]11</head>
12<body>
13
14<!--#include virtual="../menu.html.incl"-->
15
16<div id="content">
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]17
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]18<h1>"Clang" CFE Internals Manual</h1>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]19
20<ul>
21<li><a href="#intro">Introduction</a></li>
22<li><a href="#libsystem">LLVM System and Support Libraries</a></li>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]23<li><a href="#libbasic">The Clang 'Basic' Library</a>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]24 <ul>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]25 <li><a href="#Diagnostics">The Diagnostics Subsystem</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]26 <li><a href="#SourceLocation">The SourceLocation and SourceManager
27 classes</a></li>
28 </ul>
29</li>
30<li><a href="#liblex">The Lexer and Preprocessor Library</a>
31 <ul>
32 <li><a href="#Token">The Token class</a></li>
33 <li><a href="#Lexer">The Lexer class</a></li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]34 <li><a href="#AnnotationToken">Annotation Tokens</a></li>
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]35 <li><a href="#TokenLexer">The TokenLexer class</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]36 <li><a href="#MultipleIncludeOpt">The MultipleIncludeOpt class</a></li>
37 </ul>
38</li>
39<li><a href="#libparse">The Parser Library</a>
40 <ul>
41 </ul>
42</li>
43<li><a href="#libast">The AST Library</a>
44 <ul>
45 <li><a href="#Type">The Type class and its subclasses</a></li>
46 <li><a href="#QualType">The QualType class</a></li>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]47 <li><a href="#DeclarationName">Declaration names</a></li>
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]48 <li><a href="#DeclContext">Declaration contexts</a>
49 <ul>
50 <li><a href="#Redeclarations">Redeclarations and Overloads</a></li>
51 <li><a href="#LexicalAndSemanticContexts">Lexical and Semantic
52 Contexts</a></li>
53 <li><a href="#TransparentContexts">Transparent Declaration Contexts</a></li>
54 <li><a href="#MultiDeclContext">Multiply-Defined Declaration Contexts</a></li>
55 </ul>
56 </li>
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]57 <li><a href="#CFG">The CFG class</a></li>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]58 <li><a href="#Constants">Constant Folding in the Clang AST</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]59 </ul>
60</li>
61</ul>
62
63
64<!-- ======================================================================= -->
65<h2 id="intro">Introduction</h2>
66<!-- ======================================================================= -->
67
68<p>This document describes some of the more important APIs and internal design
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]69decisions made in the Clang C front-end. The purpose of this document is to
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]70both capture some of this high level information and also describe some of the
71design decisions behind it. This is meant for people interested in hacking on
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]72Clang, not for end-users. The description below is categorized by
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]73libraries, and does not describe any of the clients of the libraries.</p>
74
75<!-- ======================================================================= -->
76<h2 id="libsystem">LLVM System and Support Libraries</h2>
77<!-- ======================================================================= -->
78
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]79<p>The LLVM libsystem library provides the basic Clang system abstraction layer,
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]80which is used for file system access. The LLVM libsupport library provides many
81underlying libraries and <a
82href="http://llvm.org/docs/ProgrammersManual.html">data-structures</a>,
83 including command line option
84processing and various containers.</p>
85
86<!-- ======================================================================= -->
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]87<h2 id="libbasic">The Clang 'Basic' Library</h2>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]88<!-- ======================================================================= -->
89
90<p>This library certainly needs a better name. The 'basic' library contains a
91number of low-level utilities for tracking and manipulating source buffers,
92locations within the source buffers, diagnostics, tokens, target abstraction,
93and information about the subset of the language being compiled for.</p>
94
95<p>Part of this infrastructure is specific to C (such as the TargetInfo class),
96other parts could be reused for other non-C-based languages (SourceLocation,
97SourceManager, Diagnostics, FileManager). When and if there is future demand
98we can figure out if it makes sense to introduce a new library, move the general
99classes somewhere else, or introduce some other solution.</p>
100
101<p>We describe the roles of these classes in order of their dependencies.</p>
102
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]103
104<!-- ======================================================================= -->
105<h3 id="Diagnostics">The Diagnostics Subsystem</h3>
106<!-- ======================================================================= -->
107
108<p>The Clang Diagnostics subsystem is an important part of how the compiler
109communicates with the human. Diagnostics are the warnings and errors produced
110when the code is incorrect or dubious. In Clang, each diagnostic produced has
111(at the minimum) a unique ID, a <a href="#SourceLocation">SourceLocation</a> to
112"put the caret", an English translation associated with it, and a severity (e.g.
113<tt>WARNING</tt> or <tt>ERROR</tt>). They can also optionally include a number
114of arguments to the dianostic (which fill in "%0"'s in the string) as well as a
115number of source ranges that related to the diagnostic.</p>
116
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]117<p>In this section, we'll be giving examples produced by the Clang command line
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]118driver, but diagnostics can be <a href="#DiagnosticClient">rendered in many
119different ways</a> depending on how the DiagnosticClient interface is
120implemented. A representative example of a diagonstic is:</p>
121
122<pre>
123t.c:38:15: error: invalid operands to binary expression ('int *' and '_Complex float')
124 <font color="darkgreen">P = (P-42) + Gamma*4;</font>
125 <font color="blue">~~~~~~ ^ ~~~~~~~</font>
126</pre>
127
128<p>In this example, you can see the English translation, the severity (error),
129you can see the source location (the caret ("^") and file/line/column info),
130the source ranges "~~~~", arguments to the diagnostic ("int*" and "_Complex
131float"). You'll have to believe me that there is a unique ID backing the
132diagnostic :).</p>
133
134<p>Getting all of this to happen has several steps and involves many moving
135pieces, this section describes them and talks about best practices when adding
136a new diagnostic.</p>
137
138<!-- ============================ -->
139<h4>The DiagnosticKinds.def file</h4>
140<!-- ============================ -->
141
142<p>Diagnostics are created by adding an entry to the <tt><a
143href="http://llvm.org/svn/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticKinds.def"
144>DiagnosticKinds.def</a></tt> file. This file encodes the unique ID of the
145diagnostic (as an enum, the first argument), the severity of the diagnostic
146(second argument) and the English translation + format string.</p>
147
148<p>There is little sanity with the naming of the unique ID's right now. Some
149start with err_, warn_, ext_ to encode the severity into the name. Since the
150enum is referenced in the C++ code that produces the diagnostic, it is somewhat
151useful for it to be reasonably short.</p>
152
153<p>The severity of the diagnostic comes from the set {<tt>NOTE</tt>,
154<tt>WARNING</tt>, <tt>EXTENSION</tt>, <tt>EXTWARN</tt>, <tt>ERROR</tt>}. The
155<tt>ERROR</tt> severity is used for diagnostics indicating the program is never
156acceptable under any circumstances. When an error is emitted, the AST for the
157input code may not be fully built. The <tt>EXTENSION</tt> and <tt>EXTWARN</tt>
158severities are used for extensions to the language that Clang accepts. This
159means that Clang fully understands and can represent them in the AST, but we
160produce diagnostics to tell the user their code is non-portable. The difference
161is that the former are ignored by default, and the later warn by default. The
162<tt>WARNING</tt> severity is used for constructs that are valid in the currently
163selected source language but that are dubious in some way. The <tt>NOTE</tt>
Daniel Dunbar426b8632009-02-17 15:49:03 +0000[diff] [blame]164level is used to staple more information onto previous diagnostics.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]165
166<p>These <em>severities</em> are mapped into a smaller set (the
167Diagnostic::Level enum, {<tt>Ignored</tt>, <tt>Note</tt>, <tt>Warning</tt>,
Chris Lattner0aad2972009-02-05 22:49:08 +0000[diff] [blame]168<tt>Error</tt>, <tt>Fatal</tt> }) of output <em>levels</em> by the diagnostics
Chris Lattnera180fdd2009-02-17 07:07:29 +0000[diff] [blame]169subsystem based on various configuration options. Clang internally supports a
170fully fine grained mapping mechanism that allows you to map almost any
171diagnostic to the output level that you want. The only diagnostics that cannot
172be mapped are <tt>NOTE</tt>s, which always follow the severity of the previously
173emitted diagnostic and <tt>ERROR</tt>s, which can only be mapped to
174<tt>Fatal</tt> (it is not possible to turn an error into a warning,
175for example).</p>
176
177<p>Diagnostic mappings are used in many ways. For example, if the user
178specifies <tt>-pedantic</tt>, <tt>EXTENSION</tt> maps to <tt>Warning</tt>, if
179they specify <tt>-pedantic-errors</tt>, it turns into <tt>Error</tt>. This is
180used to implement options like <tt>-Wunused_macros</tt>, <tt>-Wundef</tt> etc.
181</p>
182
183<p>
184Mapping to <tt>Fatal</tt> should only be used for diagnostics that are
185considered so severe that error recovery won't be able to recover sensibly from
186them (thus spewing a ton of bogus errors). One example of this class of error
187are failure to #include a file.
188</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]189
190<!-- ================= -->
191<h4>The Format String</h4>
192<!-- ================= -->
193
194<p>The format string for the diagnostic is very simple, but it has some power.
195It takes the form of a string in English with markers that indicate where and
196how arguments to the diagnostic are inserted and formatted. For example, here
197are some simple format strings:</p>
198
199<pre>
200 "binary integer literals are an extension"
201 "format string contains '\\0' within the string body"
202 "more '<b>%%</b>' conversions than data arguments"
Chris Lattner545b3682008-11-23 20:27:13 +0000[diff] [blame]203 "invalid operands to binary expression (<b>%0</b> and <b>%1</b>)"
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]204 "overloaded '<b>%0</b>' must be a <b>%select{unary|binary|unary or binary}2</b> operator"
205 " (has <b>%1</b> parameter<b>%s1</b>)"
206</pre>
207
208<p>These examples show some important points of format strings. You can use any
209 plain ASCII character in the diagnostic string except "%" without a problem,
210 but these are C strings, so you have to use and be aware of all the C escape
211 sequences (as in the second example). If you want to produce a "%" in the
212 output, use the "%%" escape sequence, like the third diagnostic. Finally,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]213 Clang uses the "%...[digit]" sequences to specify where and how arguments to
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]214 the diagnostic are formatted.</p>
215
216<p>Arguments to the diagnostic are numbered according to how they are specified
217 by the C++ code that <a href="#producingdiag">produces them</a>, and are
218 referenced by <tt>%0</tt> .. <tt>%9</tt>. If you have more than 10 arguments
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]219 to your diagnostic, you are doing something wrong :). Unlike printf, there
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]220 is no requirement that arguments to the diagnostic end up in the output in
221 the same order as they are specified, you could have a format string with
222 <tt>"%1 %0"</tt> that swaps them, for example. The text in between the
223 percent and digit are formatting instructions. If there are no instructions,
224 the argument is just turned into a string and substituted in.</p>
225
226<p>Here are some "best practices" for writing the English format string:</p>
227
228<ul>
229<li>Keep the string short. It should ideally fit in the 80 column limit of the
230 <tt>DiagnosticKinds.def</tt> file. This avoids the diagnostic wrapping when
231 printed, and forces you to think about the important point you are conveying
232 with the diagnostic.</li>
233<li>Take advantage of location information. The user will be able to see the
234 line and location of the caret, so you don't need to tell them that the
235 problem is with the 4th argument to the function: just point to it.</li>
236<li>Do not capitalize the diagnostic string, and do not end it with a
237 period.</li>
238<li>If you need to quote something in the diagnostic string, use single
239 quotes.</li>
240</ul>
241
242<p>Diagnostics should never take random English strings as arguments: you
243shouldn't use <tt>"you have a problem with %0"</tt> and pass in things like
244<tt>"your argument"</tt> or <tt>"your return value"</tt> as arguments. Doing
245this prevents <a href="translation">translating</a> the Clang diagnostics to
246other languages (because they'll get random English words in their otherwise
247localized diagnostic). The exceptions to this are C/C++ language keywords
248(e.g. auto, const, mutable, etc) and C/C++ operators (<tt>/=</tt>). Note
249that things like "pointer" and "reference" are not keywords. On the other
250hand, you <em>can</em> include anything that comes from the user's source code,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]251including variable names, types, labels, etc. The 'select' format can be
252used to achieve this sort of thing in a localizable way, see below.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]253
254<!-- ==================================== -->
255<h4>Formatting a Diagnostic Argument</a></h4>
256<!-- ==================================== -->
257
258<p>Arguments to diagnostics are fully typed internally, and come from a couple
259different classes: integers, types, names, and random strings. Depending on
260the class of the argument, it can be optionally formatted in different ways.
261This gives the DiagnosticClient information about what the argument means
262without requiring it to use a specific presentation (consider this MVC for
263Clang :).</p>
264
265<p>Here are the different diagnostic argument formats currently supported by
266Clang:</p>
267
268<table>
269<tr><td colspan="2"><b>"s" format</b></td></tr>
270<tr><td>Example:</td><td><tt>"requires %1 parameter%s1"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]271<tr><td>Class:</td><td>Integers</td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]272<tr><td>Description:</td><td>This is a simple formatter for integers that is
273 useful when producing English diagnostics. When the integer is 1, it prints
274 as nothing. When the integer is not 1, it prints as "s". This allows some
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]275 simple grammatical forms to be to be handled correctly, and eliminates the
276 need to use gross things like <tt>"requires %1 parameter(s)"</tt>.</td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]277
278<tr><td colspan="2"><b>"select" format</b></td></tr>
279<tr><td>Example:</td><td><tt>"must be a %select{unary|binary|unary or binary}2
280 operator"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]281<tr><td>Class:</td><td>Integers</td></tr>
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]282<tr><td>Description:</td><td>This format specifier is used to merge multiple
283 related diagnostics together into one common one, without requiring the
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]284 difference to be specified as an English string argument. Instead of
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]285 specifying the string, the diagnostic gets an integer argument and the
286 format string selects the numbered option. In this case, the "%2" value
287 must be an integer in the range [0..2]. If it is 0, it prints 'unary', if
288 it is 1 it prints 'binary' if it is 2, it prints 'unary or binary'. This
289 allows other language translations to substitute reasonable words (or entire
290 phrases) based on the semantics of the diagnostic instead of having to do
291 things textually.</td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]292
293<tr><td colspan="2"><b>"plural" format</b></td></tr>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]294<tr><td>Example:</td><td><tt>"you have %1 %plural{1:mouse|:mice}1 connected to
295 your computer"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]296<tr><td>Class:</td><td>Integers</td></tr>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]297<tr><td>Description:</td><td><p>This is a formatter for complex plural forms.
298 It is designed to handle even the requirements of languages with very
299 complex plural forms, as many Baltic languages have. The argument consists
300 of a series of expression/form pairs, separated by ':', where the first form
301 whose expression evaluates to true is the result of the modifier.</p>
302 <p>An expression can be empty, in which case it is always true. See the
303 example at the top. Otherwise, it is a series of one or more numeric
304 conditions, separated by ','. If any condition matches, the expression
305 matches. Each numeric condition can take one of three forms.</p>
306 <ul>
307 <li>number: A simple decimal number matches if the argument is the same
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]308 as the number. Example: <tt>"%plural{1:mouse|:mice}4"</tt></li>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]309 <li>range: A range in square brackets matches if the argument is within
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]310 the range. Then range is inclusive on both ends. Example:
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]311 <tt>"%plural{0:none|1:one|[2,5]:some|:many}2"</tt></li>
312 <li>modulo: A modulo operator is followed by a number, and
313 equals sign and either a number or a range. The tests are the
314 same as for plain
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]315 numbers and ranges, but the argument is taken modulo the number first.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]316 Example: <tt>"%plural{%100=0:even hundred|%100=[1,50]:lower half|:everything
317 else}1"</tt></li>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]318 </ul>
319 <p>The parser is very unforgiving. A syntax error, even whitespace, will
320 abort, as will a failure to match the argument against any
321 expression.</p></td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]322
Chris Lattner077bf5e2008-11-24 03:33:13 +0000[diff] [blame]323<tr><td colspan="2"><b>"objcclass" format</b></td></tr>
324<tr><td>Example:</td><td><tt>"method %objcclass0 not found"</tt></td></tr>
325<tr><td>Class:</td><td>DeclarationName</td></tr>
326<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
327 DeclarationName corresponds to an Objective-C class method selector. As
328 such, it prints the selector with a leading '+'.</p></td></tr>
329
330<tr><td colspan="2"><b>"objcinstance" format</b></td></tr>
331<tr><td>Example:</td><td><tt>"method %objcinstance0 not found"</tt></td></tr>
332<tr><td>Class:</td><td>DeclarationName</td></tr>
333<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
334 DeclarationName corresponds to an Objective-C instance method selector. As
335 such, it prints the selector with a leading '-'.</p></td></tr>
336
Douglas Gregor47b9a1c2009-02-04 17:27:36 +0000[diff] [blame]337<tr><td colspan="2"><b>"q" format</b></td></tr>
338<tr><td>Example:</td><td><tt>"candidate found by name lookup is %q0"</tt></td></tr>
339<tr><td>Class:</td><td>NamedDecl*</td></tr>
340<tr><td>Description</td><td><p>This formatter indicates that the fully-qualified name of the declaration should be printed, e.g., "std::vector" rather than "vector".</p></td></tr>
341
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]342</table>
343
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]344<p>It is really easy to add format specifiers to the Clang diagnostics system,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]345but they should be discussed before they are added. If you are creating a lot
346of repetitive diagnostics and/or have an idea for a useful formatter, please
347bring it up on the cfe-dev mailing list.</p>
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]348
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]349<!-- ===================================================== -->
350<h4><a name="#producingdiag">Producing the Diagnostic</a></h4>
351<!-- ===================================================== -->
352
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]353<p>Now that you've created the diagnostic in the DiagnosticKinds.def file, you
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]354need to write the code that detects the condition in question and emits the
355new diagnostic. Various components of Clang (e.g. the preprocessor, Sema,
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]356etc) provide a helper function named "Diag". It creates a diagnostic and
357accepts the arguments, ranges, and other information that goes along with
358it.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]359
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]360<p>For example, the binary expression error comes from code like this:</p>
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]361
362<pre>
363 if (various things that are bad)
364 Diag(Loc, diag::err_typecheck_invalid_operands)
365 &lt;&lt; lex-&gt;getType() &lt;&lt; rex-&gt;getType()
366 &lt;&lt; lex-&gt;getSourceRange() &lt;&lt; rex-&gt;getSourceRange();
367</pre>
368
369<p>This shows that use of the Diag method: they take a location (a <a
370href="#SourceLocation">SourceLocation</a> object) and a diagnostic enum value
371(which matches the name from DiagnosticKinds.def). If the diagnostic takes
372arguments, they are specified with the &lt;&lt; operator: the first argument
373becomes %0, the second becomes %1, etc. The diagnostic interface allows you to
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]374specify arguments of many different types, including <tt>int</tt> and
375<tt>unsigned</tt> for integer arguments, <tt>const char*</tt> and
376<tt>std::string</tt> for string arguments, <tt>DeclarationName</tt> and
377<tt>const IdentifierInfo*</tt> for names, <tt>QualType</tt> for types, etc.
378SourceRanges are also specified with the &lt;&lt; operator, but do not have a
379specific ordering requirement.</p>
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]380
381<p>As you can see, adding and producing a diagnostic is pretty straightforward.
382The hard part is deciding exactly what you need to say to help the user, picking
383a suitable wording, and providing the information needed to format it correctly.
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]384The good news is that the call site that issues a diagnostic should be
385completely independent of how the diagnostic is formatted and in what language
386it is rendered.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]387</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]388
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]389<!-- ==================================================== -->
390<h4 id="code-modification-hints">Code Modification Hints</h4>
391<!-- ==================================================== -->
392
393<p>In some cases, the front end emits diagnostics when it is clear
394that some small change to the source code would fix the problem. For
395example, a missing semicolon at the end of a statement or a use of
Chris Lattner34c05332009-02-27 19:31:12 +0000[diff] [blame^]396deprecated syntax that is easily rewritten into a more modern form.
397Clang tries very hard to emit the diagnostic and recover gracefully
398in these and other cases.</p>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]399
Chris Lattner34c05332009-02-27 19:31:12 +0000[diff] [blame^]400<p>However, for these cases where the fix is obvious, the diagnostic
401can be annotated with a code
402modification "hint" that describes how to change the code referenced
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]403by the diagnostic to fix the problem. For example, it might add the
404missing semicolon at the end of the statement or rewrite the use of a
405deprecated construct into something more palatable. Here is one such
406example C++ front end, where we warn about the right-shift operator
407changing meaning from C++98 to C++0x:</p>
408
409<pre>
410test.cpp:3:7: warning: use of right-shift operator ('&gt;&gt;') in template argument will require parentheses in C++0x
411A&lt;100 &gt;&gt; 2&gt; *a;
412 ^
413 ( )
414</pre>
415
416<p>Here, the code modification hint is suggesting that parentheses be
417added, and showing exactly where those parentheses would be inserted
418into the source code. The code modification hints themselves describe
419what changes to make to the source code in an abstract manner, which
420the text diagnostic printer renders as a line of "insertions" below
421the caret line. <a href="#DiagnosticClient">Other diagnostic
422clients</a> might choose to render the code differently (e.g., as
423markup inline) or even give the user the ability to automatically fix
424the problem.</p>
425
426<p>All code modification hints are described by the
427<code>CodeModificationHint</code> class, instances of which should be
428attached to the diagnostic using the &lt;&lt; operator in the same way
429that highlighted source ranges and arguments are passed to the
430diagnostic. Code modification hints can be created with one of three
431constructors:</p>
432
433<dl>
434 <dt><code>CodeModificationHint::CreateInsertion(Loc, Code)</code></dt>
435 <dd>Specifies that the given <code>Code</code> (a string) should be inserted
436 before the source location <code>Loc</code>.</dd>
437
438 <dt><code>CodeModificationHint::CreateRemoval(Range)</code></dt>
439 <dd>Specifies that the code in the given source <code>Range</code>
440 should be removed.</dd>
441
442 <dt><code>CodeModificationHint::CreateReplacement(Range, Code)</code></dt>
443 <dd>Specifies that the code in the given source <code>Range</code>
444 should be removed, and replaced with the given <code>Code</code> string.</dd>
445</dl>
446
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]447<!-- ============================================================= -->
448<h4><a name="DiagnosticClient">The DiagnosticClient Interface</a></h4>
449<!-- ============================================================= -->
450
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]451<p>Once code generates a diagnostic with all of the arguments and the rest of
452the relevant information, Clang needs to know what to do with it. As previously
453mentioned, the diagnostic machinery goes through some filtering to map a
454severity onto a diagnostic level, then (assuming the diagnostic is not mapped to
455"<tt>Ignore</tt>") it invokes an object that implements the DiagnosticClient
456interface with the information.</p>
457
458<p>It is possible to implement this interface in many different ways. For
459example, the normal Clang DiagnosticClient (named 'TextDiagnosticPrinter') turns
460the arguments into strings (according to the various formatting rules), prints
461out the file/line/column information and the string, then prints out the line of
462code, the source ranges, and the caret. However, this behavior isn't required.
463</p>
464
465<p>Another implementation of the DiagnosticClient interface is the
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]466'TextDiagnosticBuffer' class, which is used when Clang is in -verify mode.
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]467Instead of formatting and printing out the diagnostics, this implementation just
468captures and remembers the diagnostics as they fly by. Then -verify compares
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]469the list of produced diagnostics to the list of expected ones. If they disagree,
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]470it prints out its own output.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]471</p>
472
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]473<p>There are many other possible implementations of this interface, and this is
474why we prefer diagnostics to pass down rich structured information in arguments.
475For example, an HTML output might want declaration names be linkified to where
476they come from in the source. Another example is that a GUI might let you click
477on typedefs to expand them. This application would want to pass significantly
478more information about types through to the GUI than a simple flat string. The
479interface allows this to happen.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]480
481<!-- ====================================================== -->
482<h4><a name="translation">Adding Translations to Clang</a></h4>
483<!-- ====================================================== -->
484
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]485<p>Not possible yet! Diagnostic strings should be written in UTF-8, the client
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]486can translate to the relevant code page if needed. Each translation completely
487replaces the format string for the diagnostic.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]488
489
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]490<!-- ======================================================================= -->
491<h3 id="SourceLocation">The SourceLocation and SourceManager classes</h3>
492<!-- ======================================================================= -->
493
494<p>Strangely enough, the SourceLocation class represents a location within the
495source code of the program. Important design points include:</p>
496
497<ol>
498<li>sizeof(SourceLocation) must be extremely small, as these are embedded into
499 many AST nodes and are passed around often. Currently it is 32 bits.</li>
500<li>SourceLocation must be a simple value object that can be efficiently
501 copied.</li>
502<li>We should be able to represent a source location for any byte of any input
503 file. This includes in the middle of tokens, in whitespace, in trigraphs,
504 etc.</li>
505<li>A SourceLocation must encode the current #include stack that was active when
506 the location was processed. For example, if the location corresponds to a
507 token, it should contain the set of #includes active when the token was
508 lexed. This allows us to print the #include stack for a diagnostic.</li>
509<li>SourceLocation must be able to describe macro expansions, capturing both
510 the ultimate instantiation point and the source of the original character
511 data.</li>
512</ol>
513
514<p>In practice, the SourceLocation works together with the SourceManager class
Chris Lattner18376dd2009-01-16 07:00:50 +0000[diff] [blame]515to encode two pieces of information about a location: it's spelling location
Chris Lattner88054de2009-01-16 07:15:35 +0000[diff] [blame]516and it's instantiation location. For most tokens, these will be the same. However,
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]517for a macro expansion (or tokens that came from a _Pragma directive) these will
518describe the location of the characters corresponding to the token and the
519location where the token was used (i.e. the macro instantiation point or the
520location of the _Pragma itself).</p>
521
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]522<p>For efficiency, we only track one level of macro instantiations: if a token was
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]523produced by multiple instantiations, we only track the source and ultimate
524destination. Though we could track the intermediate instantiation points, this
525would require extra bookkeeping and no known client would benefit substantially
526from this.</p>
527
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]528<p>The Clang front-end inherently depends on the location of a token being
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]529tracked correctly. If it is ever incorrect, the front-end may get confused and
530die. The reason for this is that the notion of the 'spelling' of a Token in
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]531Clang depends on being able to find the original input characters for the token.
Chris Lattner18376dd2009-01-16 07:00:50 +0000[diff] [blame]532This concept maps directly to the "spelling location" for the token.</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]533
534<!-- ======================================================================= -->
535<h2 id="liblex">The Lexer and Preprocessor Library</h2>
536<!-- ======================================================================= -->
537
538<p>The Lexer library contains several tightly-connected classes that are involved
539with the nasty process of lexing and preprocessing C source code. The main
540interface to this library for outside clients is the large <a
541href="#Preprocessor">Preprocessor</a> class.
542It contains the various pieces of state that are required to coherently read
543tokens out of a translation unit.</p>
544
545<p>The core interface to the Preprocessor object (once it is set up) is the
546Preprocessor::Lex method, which returns the next <a href="#Token">Token</a> from
547the preprocessor stream. There are two types of token providers that the
548preprocessor is capable of reading from: a buffer lexer (provided by the <a
549href="#Lexer">Lexer</a> class) and a buffered token stream (provided by the <a
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]550href="#TokenLexer">TokenLexer</a> class).
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]551
552
553<!-- ======================================================================= -->
554<h3 id="Token">The Token class</h3>
555<!-- ======================================================================= -->
556
557<p>The Token class is used to represent a single lexed token. Tokens are
558intended to be used by the lexer/preprocess and parser libraries, but are not
559intended to live beyond them (for example, they should not live in the ASTs).<p>
560
561<p>Tokens most often live on the stack (or some other location that is efficient
562to access) as the parser is running, but occasionally do get buffered up. For
563example, macro definitions are stored as a series of tokens, and the C++
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]564front-end periodically needs to buffer tokens up for tentative parsing and
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]565various pieces of look-ahead. As such, the size of a Token matter. On a 32-bit
566system, sizeof(Token) is currently 16 bytes.</p>
567
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]568<p>Tokens occur in two forms: "<a href="#AnnotationToken">Annotation
569Tokens</a>" and normal tokens. Normal tokens are those returned by the lexer,
570annotation tokens represent semantic information and are produced by the parser,
571replacing normal tokens in the token stream. Normal tokens contain the
572following information:</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]573
574<ul>
575<li><b>A SourceLocation</b> - This indicates the location of the start of the
576token.</li>
577
578<li><b>A length</b> - This stores the length of the token as stored in the
579SourceBuffer. For tokens that include them, this length includes trigraphs and
580escaped newlines which are ignored by later phases of the compiler. By pointing
581into the original source buffer, it is always possible to get the original
582spelling of a token completely accurately.</li>
583
584<li><b>IdentifierInfo</b> - If a token takes the form of an identifier, and if
585identifier lookup was enabled when the token was lexed (e.g. the lexer was not
586reading in 'raw' mode) this contains a pointer to the unique hash value for the
587identifier. Because the lookup happens before keyword identification, this
588field is set even for language keywords like 'for'.</li>
589
590<li><b>TokenKind</b> - This indicates the kind of token as classified by the
591lexer. This includes things like <tt>tok::starequal</tt> (for the "*="
592operator), <tt>tok::ampamp</tt> for the "&amp;&amp;" token, and keyword values
593(e.g. <tt>tok::kw_for</tt>) for identifiers that correspond to keywords. Note
594that some tokens can be spelled multiple ways. For example, C++ supports
595"operator keywords", where things like "and" are treated exactly like the
596"&amp;&amp;" operator. In these cases, the kind value is set to
597<tt>tok::ampamp</tt>, which is good for the parser, which doesn't have to
598consider both forms. For something that cares about which form is used (e.g.
599the preprocessor 'stringize' operator) the spelling indicates the original
600form.</li>
601
602<li><b>Flags</b> - There are currently four flags tracked by the
603lexer/preprocessor system on a per-token basis:
604
605 <ol>
606 <li><b>StartOfLine</b> - This was the first token that occurred on its input
607 source line.</li>
608 <li><b>LeadingSpace</b> - There was a space character either immediately
609 before the token or transitively before the token as it was expanded
610 through a macro. The definition of this flag is very closely defined by
611 the stringizing requirements of the preprocessor.</li>
612 <li><b>DisableExpand</b> - This flag is used internally to the preprocessor to
613 represent identifier tokens which have macro expansion disabled. This
614 prevents them from being considered as candidates for macro expansion ever
615 in the future.</li>
616 <li><b>NeedsCleaning</b> - This flag is set if the original spelling for the
617 token includes a trigraph or escaped newline. Since this is uncommon,
618 many pieces of code can fast-path on tokens that did not need cleaning.
619 </p>
620 </ol>
621</li>
622</ul>
623
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]624<p>One interesting (and somewhat unusual) aspect of normal tokens is that they
625don't contain any semantic information about the lexed value. For example, if
626the token was a pp-number token, we do not represent the value of the number
627that was lexed (this is left for later pieces of code to decide). Additionally,
628the lexer library has no notion of typedef names vs variable names: both are
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]629returned as identifiers, and the parser is left to decide whether a specific
630identifier is a typedef or a variable (tracking this requires scope information
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]631among other things). The parser can do this translation by replacing tokens
632returned by the preprocessor with "Annotation Tokens".</p>
633
634<!-- ======================================================================= -->
635<h3 id="AnnotationToken">Annotation Tokens</h3>
636<!-- ======================================================================= -->
637
638<p>Annotation Tokens are tokens that are synthesized by the parser and injected
639into the preprocessor's token stream (replacing existing tokens) to record
640semantic information found by the parser. For example, if "foo" is found to be
641a typedef, the "foo" <tt>tok::identifier</tt> token is replaced with an
642<tt>tok::annot_typename</tt>. This is useful for a couple of reasons: 1) this
643makes it easy to handle qualified type names (e.g. "foo::bar::baz&lt;42&gt;::t")
644in C++ as a single "token" in the parser. 2) if the parser backtracks, the
645reparse does not need to redo semantic analysis to determine whether a token
646sequence is a variable, type, template, etc.</p>
647
648<p>Annotation Tokens are created by the parser and reinjected into the parser's
649token stream (when backtracking is enabled). Because they can only exist in
650tokens that the preprocessor-proper is done with, it doesn't need to keep around
651flags like "start of line" that the preprocessor uses to do its job.
652Additionally, an annotation token may "cover" a sequence of preprocessor tokens
653(e.g. <tt>a::b::c</tt> is five preprocessor tokens). As such, the valid fields
654of an annotation token are different than the fields for a normal token (but
655they are multiplexed into the normal Token fields):</p>
656
657<ul>
658<li><b>SourceLocation "Location"</b> - The SourceLocation for the annotation
659token indicates the first token replaced by the annotation token. In the example
660above, it would be the location of the "a" identifier.</li>
661
662<li><b>SourceLocation "AnnotationEndLoc"</b> - This holds the location of the
663last token replaced with the annotation token. In the example above, it would
664be the location of the "c" identifier.</li>
665
666<li><b>void* "AnnotationValue"</b> - This contains an opaque object that the
667parser gets from Sema through an Actions module, it is passed around and Sema
668intepretes it, based on the type of annotation token.</li>
669
670<li><b>TokenKind "Kind"</b> - This indicates the kind of Annotation token this
671is. See below for the different valid kinds.</li>
672</ul>
673
674<p>Annotation tokens currently come in three kinds:</p>
675
676<ol>
677<li><b>tok::annot_typename</b>: This annotation token represents a
678resolved typename token that is potentially qualified. The AnnotationValue
Steve Naroffb43a50f2009-01-28 19:39:02 +0000[diff] [blame]679field contains a pointer returned by Action::getTypeName(). In the case of the
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]680Sema actions module, this is a <tt>Decl*</tt> for the type.</li>
681
682<li><b>tok::annot_cxxscope</b>: This annotation token represents a C++ scope
683specifier, such as "A::B::". This corresponds to the grammar productions "::"
684and ":: [opt] nested-name-specifier". The AnnotationValue pointer is returned
685by the Action::ActOnCXXGlobalScopeSpecifier and
686Action::ActOnCXXNestedNameSpecifier callbacks. In the case of Sema, this is a
687<tt>DeclContext*</tt>.</li>
688
Douglas Gregor39a8de12009-02-25 19:37:18 +0000[diff] [blame]689<li><b>tok::annot_template_id</b>: This annotation token represents a
690C++ template-id such as "foo&lt;int, 4&gt;", where "foo" is the name
691of a template. The AnnotationValue pointer is a pointer to a malloc'd
692TemplateIdAnnotation object. Depending on the context, a parsed template-id that names a type might become a typename annotation token (if all we care about is the named type, e.g., because it occurs in a type specifier) or might remain a template-id token (if we want to retain more source location information or produce a new type, e.g., in a declaration of a class template specialization). template-id annotation tokens that refer to a type can be "upgraded" to typename annotation tokens by the parser.</li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]693
694</ol>
695
Cedric Venetda76b282009-01-06 16:22:54 +0000[diff] [blame]696<p>As mentioned above, annotation tokens are not returned by the preprocessor,
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]697they are formed on demand by the parser. This means that the parser has to be
698aware of cases where an annotation could occur and form it where appropriate.
699This is somewhat similar to how the parser handles Translation Phase 6 of C99:
700String Concatenation (see C99 5.1.1.2). In the case of string concatenation,
701the preprocessor just returns distinct tok::string_literal and
702tok::wide_string_literal tokens and the parser eats a sequence of them wherever
703the grammar indicates that a string literal can occur.</p>
704
705<p>In order to do this, whenever the parser expects a tok::identifier or
706tok::coloncolon, it should call the TryAnnotateTypeOrScopeToken or
707TryAnnotateCXXScopeToken methods to form the annotation token. These methods
708will maximally form the specified annotation tokens and replace the current
709token with them, if applicable. If the current tokens is not valid for an
710annotation token, it will remain an identifier or :: token.</p>
711
712
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]713
714<!-- ======================================================================= -->
715<h3 id="Lexer">The Lexer class</h3>
716<!-- ======================================================================= -->
717
718<p>The Lexer class provides the mechanics of lexing tokens out of a source
719buffer and deciding what they mean. The Lexer is complicated by the fact that
720it operates on raw buffers that have not had spelling eliminated (this is a
721necessity to get decent performance), but this is countered with careful coding
722as well as standard performance techniques (for example, the comment handling
723code is vectorized on X86 and PowerPC hosts).</p>
724
725<p>The lexer has a couple of interesting modal features:</p>
726
727<ul>
728<li>The lexer can operate in 'raw' mode. This mode has several features that
729 make it possible to quickly lex the file (e.g. it stops identifier lookup,
730 doesn't specially handle preprocessor tokens, handles EOF differently, etc).
731 This mode is used for lexing within an "<tt>#if 0</tt>" block, for
732 example.</li>
733<li>The lexer can capture and return comments as tokens. This is required to
734 support the -C preprocessor mode, which passes comments through, and is
735 used by the diagnostic checker to identifier expect-error annotations.</li>
736<li>The lexer can be in ParsingFilename mode, which happens when preprocessing
Chris Lattner84386242007-09-16 19:25:23 +0000[diff] [blame]737 after reading a #include directive. This mode changes the parsing of '&lt;'
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]738 to return an "angled string" instead of a bunch of tokens for each thing
739 within the filename.</li>
740<li>When parsing a preprocessor directive (after "<tt>#</tt>") the
741 ParsingPreprocessorDirective mode is entered. This changes the parser to
742 return EOM at a newline.</li>
743<li>The Lexer uses a LangOptions object to know whether trigraphs are enabled,
744 whether C++ or ObjC keywords are recognized, etc.</li>
745</ul>
746
747<p>In addition to these modes, the lexer keeps track of a couple of other
748 features that are local to a lexed buffer, which change as the buffer is
749 lexed:</p>
750
751<ul>
752<li>The Lexer uses BufferPtr to keep track of the current character being
753 lexed.</li>
754<li>The Lexer uses IsAtStartOfLine to keep track of whether the next lexed token
755 will start with its "start of line" bit set.</li>
756<li>The Lexer keeps track of the current #if directives that are active (which
757 can be nested).</li>
758<li>The Lexer keeps track of an <a href="#MultipleIncludeOpt">
759 MultipleIncludeOpt</a> object, which is used to
760 detect whether the buffer uses the standard "<tt>#ifndef XX</tt> /
761 <tt>#define XX</tt>" idiom to prevent multiple inclusion. If a buffer does,
762 subsequent includes can be ignored if the XX macro is defined.</li>
763</ul>
764
765<!-- ======================================================================= -->
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]766<h3 id="TokenLexer">The TokenLexer class</h3>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]767<!-- ======================================================================= -->
768
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]769<p>The TokenLexer class is a token provider that returns tokens from a list
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]770of tokens that came from somewhere else. It typically used for two things: 1)
771returning tokens from a macro definition as it is being expanded 2) returning
772tokens from an arbitrary buffer of tokens. The later use is used by _Pragma and
773will most likely be used to handle unbounded look-ahead for the C++ parser.</p>
774
775<!-- ======================================================================= -->
776<h3 id="MultipleIncludeOpt">The MultipleIncludeOpt class</h3>
777<!-- ======================================================================= -->
778
779<p>The MultipleIncludeOpt class implements a really simple little state machine
780that is used to detect the standard "<tt>#ifndef XX</tt> / <tt>#define XX</tt>"
781idiom that people typically use to prevent multiple inclusion of headers. If a
782buffer uses this idiom and is subsequently #include'd, the preprocessor can
783simply check to see whether the guarding condition is defined or not. If so,
784the preprocessor can completely ignore the include of the header.</p>
785
786
787
788<!-- ======================================================================= -->
789<h2 id="libparse">The Parser Library</h2>
790<!-- ======================================================================= -->
791
792<!-- ======================================================================= -->
793<h2 id="libast">The AST Library</h2>
794<!-- ======================================================================= -->
795
796<!-- ======================================================================= -->
797<h3 id="Type">The Type class and its subclasses</h3>
798<!-- ======================================================================= -->
799
800<p>The Type class (and its subclasses) are an important part of the AST. Types
801are accessed through the ASTContext class, which implicitly creates and uniques
802them as they are needed. Types have a couple of non-obvious features: 1) they
803do not capture type qualifiers like const or volatile (See
804<a href="#QualType">QualType</a>), and 2) they implicitly capture typedef
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]805information. Once created, types are immutable (unlike decls).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]806
807<p>Typedefs in C make semantic analysis a bit more complex than it would
808be without them. The issue is that we want to capture typedef information
809and represent it in the AST perfectly, but the semantics of operations need to
810"see through" typedefs. For example, consider this code:</p>
811
812<code>
813void func() {<br>
Bill Wendling30d17752007-10-06 01:56:01 +0000[diff] [blame]814&nbsp;&nbsp;typedef int foo;<br>
815&nbsp;&nbsp;foo X, *Y;<br>
816&nbsp;&nbsp;typedef foo* bar;<br>
817&nbsp;&nbsp;bar Z;<br>
818&nbsp;&nbsp;*X; <i>// error</i><br>
819&nbsp;&nbsp;**Y; <i>// error</i><br>
820&nbsp;&nbsp;**Z; <i>// error</i><br>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]821}<br>
822</code>
823
824<p>The code above is illegal, and thus we expect there to be diagnostics emitted
825on the annotated lines. In this example, we expect to get:</p>
826
827<pre>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]828<b>test.c:6:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]829*X; // error
830<font color="blue">^~</font>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]831<b>test.c:7:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]832**Y; // error
833<font color="blue">^~~</font>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]834<b>test.c:8:1: error: indirection requires pointer operand ('foo' invalid)</b>
835**Z; // error
836<font color="blue">^~~</font>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]837</pre>
838
839<p>While this example is somewhat silly, it illustrates the point: we want to
840retain typedef information where possible, so that we can emit errors about
841"<tt>std::string</tt>" instead of "<tt>std::basic_string&lt;char, std:...</tt>".
842Doing this requires properly keeping typedef information (for example, the type
843of "X" is "foo", not "int"), and requires properly propagating it through the
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]844various operators (for example, the type of *Y is "foo", not "int"). In order
845to retain this information, the type of these expressions is an instance of the
846TypedefType class, which indicates that the type of these expressions is a
847typedef for foo.
848</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]849
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]850<p>Representing types like this is great for diagnostics, because the
851user-specified type is always immediately available. There are two problems
852with this: first, various semantic checks need to make judgements about the
Chris Lattner33fc68a2007-07-31 18:54:50 +0000[diff] [blame]853<em>actual structure</em> of a type, ignoring typdefs. Second, we need an
854efficient way to query whether two types are structurally identical to each
855other, ignoring typedefs. The solution to both of these problems is the idea of
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]856canonical types.</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]857
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]858<!-- =============== -->
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]859<h4>Canonical Types</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]860<!-- =============== -->
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]861
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]862<p>Every instance of the Type class contains a canonical type pointer. For
863simple types with no typedefs involved (e.g. "<tt>int</tt>", "<tt>int*</tt>",
864"<tt>int**</tt>"), the type just points to itself. For types that have a
865typedef somewhere in their structure (e.g. "<tt>foo</tt>", "<tt>foo*</tt>",
866"<tt>foo**</tt>", "<tt>bar</tt>"), the canonical type pointer points to their
867structurally equivalent type without any typedefs (e.g. "<tt>int</tt>",
868"<tt>int*</tt>", "<tt>int**</tt>", and "<tt>int*</tt>" respectively).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]869
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]870<p>This design provides a constant time operation (dereferencing the canonical
871type pointer) that gives us access to the structure of types. For example,
872we can trivially tell that "bar" and "foo*" are the same type by dereferencing
873their canonical type pointers and doing a pointer comparison (they both point
874to the single "<tt>int*</tt>" type).</p>
875
876<p>Canonical types and typedef types bring up some complexities that must be
877carefully managed. Specifically, the "isa/cast/dyncast" operators generally
878shouldn't be used in code that is inspecting the AST. For example, when type
879checking the indirection operator (unary '*' on a pointer), the type checker
880must verify that the operand has a pointer type. It would not be correct to
881check that with "<tt>isa&lt;PointerType&gt;(SubExpr-&gt;getType())</tt>",
882because this predicate would fail if the subexpression had a typedef type.</p>
883
884<p>The solution to this problem are a set of helper methods on Type, used to
885check their properties. In this case, it would be correct to use
886"<tt>SubExpr-&gt;getType()-&gt;isPointerType()</tt>" to do the check. This
887predicate will return true if the <em>canonical type is a pointer</em>, which is
888true any time the type is structurally a pointer type. The only hard part here
889is remembering not to use the <tt>isa/cast/dyncast</tt> operations.</p>
890
891<p>The second problem we face is how to get access to the pointer type once we
892know it exists. To continue the example, the result type of the indirection
893operator is the pointee type of the subexpression. In order to determine the
894type, we need to get the instance of PointerType that best captures the typedef
895information in the program. If the type of the expression is literally a
896PointerType, we can return that, otherwise we have to dig through the
897typedefs to find the pointer type. For example, if the subexpression had type
898"<tt>foo*</tt>", we could return that type as the result. If the subexpression
899had type "<tt>bar</tt>", we want to return "<tt>foo*</tt>" (note that we do
900<em>not</em> want "<tt>int*</tt>"). In order to provide all of this, Type has
Chris Lattner11406c12007-07-31 16:50:51 +0000[diff] [blame]901a getAsPointerType() method that checks whether the type is structurally a
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]902PointerType and, if so, returns the best one. If not, it returns a null
903pointer.</p>
904
905<p>This structure is somewhat mystical, but after meditating on it, it will
906make sense to you :).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]907
908<!-- ======================================================================= -->
909<h3 id="QualType">The QualType class</h3>
910<!-- ======================================================================= -->
911
912<p>The QualType class is designed as a trivial value class that is small,
913passed by-value and is efficient to query. The idea of QualType is that it
914stores the type qualifiers (const, volatile, restrict) separately from the types
915themselves: QualType is conceptually a pair of "Type*" and bits for the type
916qualifiers.</p>
917
918<p>By storing the type qualifiers as bits in the conceptual pair, it is
919extremely efficient to get the set of qualifiers on a QualType (just return the
920field of the pair), add a type qualifier (which is a trivial constant-time
921operation that sets a bit), and remove one or more type qualifiers (just return
922a QualType with the bitfield set to empty).</p>
923
924<p>Further, because the bits are stored outside of the type itself, we do not
925need to create duplicates of types with different sets of qualifiers (i.e. there
926is only a single heap allocated "int" type: "const int" and "volatile const int"
927both point to the same heap allocated "int" type). This reduces the heap size
928used to represent bits and also means we do not have to consider qualifiers when
929uniquing types (<a href="#Type">Type</a> does not even contain qualifiers).</p>
930
931<p>In practice, on hosts where it is safe, the 3 type qualifiers are stored in
932the low bit of the pointer to the Type object. This means that QualType is
933exactly the same size as a pointer, and this works fine on any system where
934malloc'd objects are at least 8 byte aligned.</p>
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]935
936<!-- ======================================================================= -->
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]937<h3 id="DeclarationName">Declaration names</h3>
938<!-- ======================================================================= -->
939
940<p>The <tt>DeclarationName</tt> class represents the name of a
941 declaration in Clang. Declarations in the C family of languages can
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]942 take several different forms. Most declarations are named by
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]943 simple identifiers, e.g., "<code>f</code>" and "<code>x</code>" in
944 the function declaration <code>f(int x)</code>. In C++, declaration
945 names can also name class constructors ("<code>Class</code>"
946 in <code>struct Class { Class(); }</code>), class destructors
947 ("<code>~Class</code>"), overloaded operator names ("operator+"),
948 and conversion functions ("<code>operator void const *</code>"). In
949 Objective-C, declaration names can refer to the names of Objective-C
950 methods, which involve the method name and the parameters,
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]951 collectively called a <i>selector</i>, e.g.,
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]952 "<code>setWidth:height:</code>". Since all of these kinds of
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]953 entities - variables, functions, Objective-C methods, C++
954 constructors, destructors, and operators - are represented as
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]955 subclasses of Clang's common <code>NamedDecl</code>
956 class, <code>DeclarationName</code> is designed to efficiently
957 represent any kind of name.</p>
958
959<p>Given
960 a <code>DeclarationName</code> <code>N</code>, <code>N.getNameKind()</code>
Douglas Gregor2def4832008-11-17 20:34:05 +0000[diff] [blame]961 will produce a value that describes what kind of name <code>N</code>
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]962 stores. There are 8 options (all of the names are inside
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]963 the <code>DeclarationName</code> class)</p>
964<dl>
965 <dt>Identifier</dt>
966 <dd>The name is a simple
967 identifier. Use <code>N.getAsIdentifierInfo()</code> to retrieve the
968 corresponding <code>IdentifierInfo*</code> pointing to the actual
969 identifier. Note that C++ overloaded operators (e.g.,
970 "<code>operator+</code>") are represented as special kinds of
971 identifiers. Use <code>IdentifierInfo</code>'s <code>getOverloadedOperatorID</code>
972 function to determine whether an identifier is an overloaded
973 operator name.</dd>
974
975 <dt>ObjCZeroArgSelector, ObjCOneArgSelector,
976 ObjCMultiArgSelector</dt>
977 <dd>The name is an Objective-C selector, which can be retrieved as a
978 <code>Selector</code> instance
979 via <code>N.getObjCSelector()</code>. The three possible name
980 kinds for Objective-C reflect an optimization within
981 the <code>DeclarationName</code> class: both zero- and
982 one-argument selectors are stored as a
983 masked <code>IdentifierInfo</code> pointer, and therefore require
984 very little space, since zero- and one-argument selectors are far
985 more common than multi-argument selectors (which use a different
986 structure).</dd>
987
988 <dt>CXXConstructorName</dt>
989 <dd>The name is a C++ constructor
990 name. Use <code>N.getCXXNameType()</code> to retrieve
991 the <a href="#QualType">type</a> that this constructor is meant to
992 construct. The type is always the canonical type, since all
993 constructors for a given type have the same name.</dd>
994
995 <dt>CXXDestructorName</dt>
996 <dd>The name is a C++ destructor
997 name. Use <code>N.getCXXNameType()</code> to retrieve
998 the <a href="#QualType">type</a> whose destructor is being
999 named. This type is always a canonical type.</dd>
1000
1001 <dt>CXXConversionFunctionName</dt>
1002 <dd>The name is a C++ conversion function. Conversion functions are
1003 named according to the type they convert to, e.g., "<code>operator void
1004 const *</code>". Use <code>N.getCXXNameType()</code> to retrieve
1005 the type that this conversion function converts to. This type is
1006 always a canonical type.</dd>
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1007
1008 <dt>CXXOperatorName</dt>
1009 <dd>The name is a C++ overloaded operator name. Overloaded operators
1010 are named according to their spelling, e.g.,
1011 "<code>operator+</code>" or "<code>operator new
1012 []</code>". Use <code>N.getCXXOverloadedOperator()</code> to
1013 retrieve the overloaded operator (a value of
1014 type <code>OverloadedOperatorKind</code>).</dd>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1015</dl>
1016
1017<p><code>DeclarationName</code>s are cheap to create, copy, and
1018 compare. They require only a single pointer's worth of storage in
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1019 the common cases (identifiers, zero-
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1020 and one-argument Objective-C selectors) and use dense, uniqued
1021 storage for the other kinds of
1022 names. Two <code>DeclarationName</code>s can be compared for
1023 equality (<code>==</code>, <code>!=</code>) using a simple bitwise
1024 comparison, can be ordered
1025 with <code>&lt;</code>, <code>&gt;</code>, <code>&lt;=</code>,
1026 and <code>&gt;=</code> (which provide a lexicographical ordering for
1027 normal identifiers but an unspecified ordering for other kinds of
1028 names), and can be placed into LLVM <code>DenseMap</code>s
1029 and <code>DenseSet</code>s.</p>
1030
1031<p><code>DeclarationName</code> instances can be created in different
1032 ways depending on what kind of name the instance will store. Normal
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1033 identifiers (<code>IdentifierInfo</code> pointers) and Objective-C selectors
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1034 (<code>Selector</code>) can be implicitly converted
1035 to <code>DeclarationName</code>s. Names for C++ constructors,
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1036 destructors, conversion functions, and overloaded operators can be retrieved from
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1037 the <code>DeclarationNameTable</code>, an instance of which is
1038 available as <code>ASTContext::DeclarationNames</code>. The member
1039 functions <code>getCXXConstructorName</code>, <code>getCXXDestructorName</code>,
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1040 <code>getCXXConversionFunctionName</code>, and <code>getCXXOperatorName</code>, respectively,
1041 return <code>DeclarationName</code> instances for the four kinds of
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1042 C++ special function names.</p>
1043
1044<!-- ======================================================================= -->
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1045<h3 id="DeclContext">Declaration contexts</h3>
1046<!-- ======================================================================= -->
1047<p>Every declaration in a program exists within some <i>declaration
1048 context</i>, such as a translation unit, namespace, class, or
1049 function. Declaration contexts in Clang are represented by
1050 the <code>DeclContext</code> class, from which the various
1051 declaration-context AST nodes
1052 (<code>TranslationUnitDecl</code>, <code>NamespaceDecl</code>, <code>RecordDecl</code>, <code>FunctionDecl</code>,
1053 etc.) will derive. The <code>DeclContext</code> class provides
1054 several facilities common to each declaration context:</p>
1055<dl>
1056 <dt>Source-centric vs. Semantics-centric View of Declarations</dt>
1057 <dd><code>DeclContext</code> provides two views of the declarations
1058 stored within a declaration context. The source-centric view
1059 accurately represents the program source code as written, including
1060 multiple declarations of entities where present (see the
1061 section <a href="#Redeclarations">Redeclarations and
1062 Overloads</a>), while the semantics-centric view represents the
1063 program semantics. The two views are kept synchronized by semantic
1064 analysis while the ASTs are being constructed.</dd>
1065
1066 <dt>Storage of declarations within that context</dt>
1067 <dd>Every declaration context can contain some number of
1068 declarations. For example, a C++ class (represented
1069 by <code>RecordDecl</code>) contains various member functions,
1070 fields, nested types, and so on. All of these declarations will be
1071 stored within the <code>DeclContext</code>, and one can iterate
1072 over the declarations via
1073 [<code>DeclContext::decls_begin()</code>,
1074 <code>DeclContext::decls_end()</code>). This mechanism provides
1075 the source-centric view of declarations in the context.</dd>
1076
1077 <dt>Lookup of declarations within that context</dt>
1078 <dd>The <code>DeclContext</code> structure provides efficient name
1079 lookup for names within that declaration context. For example,
1080 if <code>N</code> is a namespace we can look for the
1081 name <code>N::f</code>
1082 using <code>DeclContext::lookup</code>. The lookup itself is
1083 based on a lazily-constructed array (for declaration contexts
1084 with a small number of declarations) or hash table (for
1085 declaration contexts with more declarations). The lookup
1086 operation provides the semantics-centric view of the declarations
1087 in the context.</dd>
1088
1089 <dt>Ownership of declarations</dt>
1090 <dd>The <code>DeclContext</code> owns all of the declarations that
1091 were declared within its declaration context, and is responsible
1092 for the management of their memory as well as their
1093 (de-)serialization.</dd>
1094</dl>
1095
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1096<p>All declarations are stored within a declaration context, and one
1097 can query
1098 information about the context in which each declaration lives. One
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1099 can retrieve the <code>DeclContext</code> that contains a
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1100 particular <code>Decl</code>
1101 using <code>Decl::getDeclContext</code>. However, see the
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1102 section <a href="#LexicalAndSemanticContexts">Lexical and Semantic
1103 Contexts</a> for more information about how to interpret this
1104 context information.</p>
1105
1106<h4 id="Redeclarations">Redeclarations and Overloads</h4>
1107<p>Within a translation unit, it is common for an entity to be
1108declared several times. For example, we might declare a function "f"
1109 and then later re-declare it as part of an inlined definition:</p>
1110
1111<pre>
1112void f(int x, int y, int z = 1);
1113
1114inline void f(int x, int y, int z) { /* ... */ }
1115</pre>
1116
1117<p>The representation of "f" differs in the source-centric and
1118 semantics-centric views of a declaration context. In the
1119 source-centric view, all redeclarations will be present, in the
1120 order they occurred in the source code, making
1121 this view suitable for clients that wish to see the structure of
1122 the source code. In the semantics-centric view, only the most recent "f"
1123 will be found by the lookup, since it effectively replaces the first
1124 declaration of "f".</p>
1125
1126<p>In the semantics-centric view, overloading of functions is
1127 represented explicitly. For example, given two declarations of a
1128 function "g" that are overloaded, e.g.,</p>
1129<pre>
1130void g();
1131void g(int);
1132</pre>
1133<p>the <code>DeclContext::lookup</code> operation will return
1134 an <code>OverloadedFunctionDecl</code> that contains both
1135 declarations of "g". Clients that perform semantic analysis on a
1136 program that is not concerned with the actual source code will
1137 primarily use this semantics-centric view.</p>
1138
1139<h4 id="LexicalAndSemanticContexts">Lexical and Semantic Contexts</h4>
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1140<p>Each declaration has two potentially different
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1141 declaration contexts: a <i>lexical</i> context, which corresponds to
1142 the source-centric view of the declaration context, and
1143 a <i>semantic</i> context, which corresponds to the
1144 semantics-centric view. The lexical context is accessible
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1145 via <code>Decl::getLexicalDeclContext</code> while the
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1146 semantic context is accessible
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1147 via <code>Decl::getDeclContext</code>, both of which return
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1148 <code>DeclContext</code> pointers. For most declarations, the two
1149 contexts are identical. For example:</p>
1150
1151<pre>
1152class X {
1153public:
1154 void f(int x);
1155};
1156</pre>
1157
1158<p>Here, the semantic and lexical contexts of <code>X::f</code> are
1159 the <code>DeclContext</code> associated with the
1160 class <code>X</code> (itself stored as a <code>RecordDecl</code> AST
1161 node). However, we can now define <code>X::f</code> out-of-line:</p>
1162
1163<pre>
1164void X::f(int x = 17) { /* ... */ }
1165</pre>
1166
1167<p>This definition of has different lexical and semantic
1168 contexts. The lexical context corresponds to the declaration
1169 context in which the actual declaration occurred in the source
1170 code, e.g., the translation unit containing <code>X</code>. Thus,
1171 this declaration of <code>X::f</code> can be found by traversing
1172 the declarations provided by
1173 [<code>decls_begin()</code>, <code>decls_end()</code>) in the
1174 translation unit.</p>
1175
1176<p>The semantic context of <code>X::f</code> corresponds to the
1177 class <code>X</code>, since this member function is (semantically) a
1178 member of <code>X</code>. Lookup of the name <code>f</code> into
1179 the <code>DeclContext</code> associated with <code>X</code> will
1180 then return the definition of <code>X::f</code> (including
1181 information about the default argument).</p>
1182
1183<h4 id="TransparentContexts">Transparent Declaration Contexts</h4>
1184<p>In C and C++, there are several contexts in which names that are
1185 logically declared inside another declaration will actually "leak"
1186 out into the enclosing scope from the perspective of name
1187 lookup. The most obvious instance of this behavior is in
1188 enumeration types, e.g.,</p>
1189<pre>
1190enum Color {
1191 Red,
1192 Green,
1193 Blue
1194};
1195</pre>
1196
1197<p>Here, <code>Color</code> is an enumeration, which is a declaration
1198 context that contains the
1199 enumerators <code>Red</code>, <code>Green</code>,
1200 and <code>Blue</code>. Thus, traversing the list of declarations
1201 contained in the enumeration <code>Color</code> will
1202 yield <code>Red</code>, <code>Green</code>,
1203 and <code>Blue</code>. However, outside of the scope
1204 of <code>Color</code> one can name the enumerator <code>Red</code>
1205 without qualifying the name, e.g.,</p>
1206
1207<pre>
1208Color c = Red;
1209</pre>
1210
1211<p>There are other entities in C++ that provide similar behavior. For
1212 example, linkage specifications that use curly braces:</p>
1213
1214<pre>
1215extern "C" {
1216 void f(int);
1217 void g(int);
1218}
1219// f and g are visible here
1220</pre>
1221
1222<p>For source-level accuracy, we treat the linkage specification and
1223 enumeration type as a
1224 declaration context in which its enclosed declarations ("Red",
1225 "Green", and "Blue"; "f" and "g")
1226 are declared. However, these declarations are visible outside of the
1227 scope of the declaration context.</p>
1228
1229<p>These language features (and several others, described below) have
1230 roughly the same set of
1231 requirements: declarations are declared within a particular lexical
1232 context, but the declarations are also found via name lookup in
1233 scopes enclosing the declaration itself. This feature is implemented
1234 via <i>transparent</i> declaration contexts
1235 (see <code>DeclContext::isTransparentContext()</code>), whose
1236 declarations are visible in the nearest enclosing non-transparent
1237 declaration context. This means that the lexical context of the
1238 declaration (e.g., an enumerator) will be the
1239 transparent <code>DeclContext</code> itself, as will the semantic
1240 context, but the declaration will be visible in every outer context
1241 up to and including the first non-transparent declaration context (since
1242 transparent declaration contexts can be nested).</p>
1243
1244<p>The transparent <code>DeclContexts</code> are:</p>
1245<ul>
1246 <li>Enumerations (but not C++0x "scoped enumerations"):
1247 <pre>
1248enum Color {
1249 Red,
1250 Green,
1251 Blue
1252};
1253// Red, Green, and Blue are in scope
1254 </pre></li>
1255 <li>C++ linkage specifications:
1256 <pre>
1257extern "C" {
1258 void f(int);
1259 void g(int);
1260}
1261// f and g are in scope
1262 </pre></li>
1263 <li>Anonymous unions and structs:
1264 <pre>
1265struct LookupTable {
1266 bool IsVector;
1267 union {
1268 std::vector&lt;Item&gt; *Vector;
1269 std::set&lt;Item&gt; *Set;
1270 };
1271};
1272
1273LookupTable LT;
1274LT.Vector = 0; // Okay: finds Vector inside the unnamed union
1275 </pre>
1276 </li>
1277 <li>C++0x inline namespaces:
1278<pre>
1279namespace mylib {
1280 inline namespace debug {
1281 class X;
1282 }
1283}
1284mylib::X *xp; // okay: mylib::X refers to mylib::debug::X
1285</pre>
1286</li>
1287</ul>
1288
1289
1290<h4 id="MultiDeclContext">Multiply-Defined Declaration Contexts</h4>
1291<p>C++ namespaces have the interesting--and, so far, unique--property that
1292the namespace can be defined multiple times, and the declarations
1293provided by each namespace definition are effectively merged (from
1294the semantic point of view). For example, the following two code
1295snippets are semantically indistinguishable:</p>
1296<pre>
1297// Snippet #1:
1298namespace N {
1299 void f();
1300}
1301namespace N {
1302 void f(int);
1303}
1304
1305// Snippet #2:
1306namespace N {
1307 void f();
1308 void f(int);
1309}
1310</pre>
1311
1312<p>In Clang's representation, the source-centric view of declaration
1313 contexts will actually have two separate <code>NamespaceDecl</code>
1314 nodes in Snippet #1, each of which is a declaration context that
1315 contains a single declaration of "f". However, the semantics-centric
1316 view provided by name lookup into the namespace <code>N</code> for
1317 "f" will return an <code>OverloadedFunctionDecl</code> that contains
1318 both declarations of "f".</p>
1319
1320<p><code>DeclContext</code> manages multiply-defined declaration
1321 contexts internally. The
1322 function <code>DeclContext::getPrimaryContext</code> retrieves the
1323 "primary" context for a given <code>DeclContext</code> instance,
1324 which is the <code>DeclContext</code> responsible for maintaining
1325 the lookup table used for the semantics-centric view. Given the
1326 primary context, one can follow the chain
1327 of <code>DeclContext</code> nodes that define additional
1328 declarations via <code>DeclContext::getNextContext</code>. Note that
1329 these functions are used internally within the lookup and insertion
1330 methods of the <code>DeclContext</code>, so the vast majority of
1331 clients can ignore them.</p>
1332
1333<!-- ======================================================================= -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1334<h3 id="CFG">The <tt>CFG</tt> class</h3>
1335<!-- ======================================================================= -->
1336
1337<p>The <tt>CFG</tt> class is designed to represent a source-level
1338control-flow graph for a single statement (<tt>Stmt*</tt>). Typically
1339instances of <tt>CFG</tt> are constructed for function bodies (usually
1340an instance of <tt>CompoundStmt</tt>), but can also be instantiated to
1341represent the control-flow of any class that subclasses <tt>Stmt</tt>,
1342which includes simple expressions. Control-flow graphs are especially
1343useful for performing
1344<a href="http://en.wikipedia.org/wiki/Data_flow_analysis#Sensitivities">flow-
1345or path-sensitive</a> program analyses on a given function.</p>
1346
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1347<!-- ============ -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1348<h4>Basic Blocks</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1349<!-- ============ -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1350
1351<p>Concretely, an instance of <tt>CFG</tt> is a collection of basic
1352blocks. Each basic block is an instance of <tt>CFGBlock</tt>, which
1353simply contains an ordered sequence of <tt>Stmt*</tt> (each referring
1354to statements in the AST). The ordering of statements within a block
1355indicates unconditional flow of control from one statement to the
1356next. <a href="#ConditionalControlFlow">Conditional control-flow</a>
1357is represented using edges between basic blocks. The statements
1358within a given <tt>CFGBlock</tt> can be traversed using
1359the <tt>CFGBlock::*iterator</tt> interface.</p>
1360
1361<p>
Ted Kremenek18e17e72007-10-18 22:50:52 +0000[diff] [blame]1362A <tt>CFG</tt> object owns the instances of <tt>CFGBlock</tt> within
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1363the control-flow graph it represents. Each <tt>CFGBlock</tt> within a
1364CFG is also uniquely numbered (accessible
1365via <tt>CFGBlock::getBlockID()</tt>). Currently the number is
1366based on the ordering the blocks were created, but no assumptions
1367should be made on how <tt>CFGBlock</tt>s are numbered other than their
1368numbers are unique and that they are numbered from 0..N-1 (where N is
1369the number of basic blocks in the CFG).</p>
1370
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1371<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1372<h4>Entry and Exit Blocks</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1373<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1374
1375Each instance of <tt>CFG</tt> contains two special blocks:
1376an <i>entry</i> block (accessible via <tt>CFG::getEntry()</tt>), which
1377has no incoming edges, and an <i>exit</i> block (accessible
1378via <tt>CFG::getExit()</tt>), which has no outgoing edges. Neither
1379block contains any statements, and they serve the role of providing a
1380clear entrance and exit for a body of code such as a function body.
1381The presence of these empty blocks greatly simplifies the
1382implementation of many analyses built on top of CFGs.
1383
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1384<!-- ===================================================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1385<h4 id ="ConditionalControlFlow">Conditional Control-Flow</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1386<!-- ===================================================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1387
1388<p>Conditional control-flow (such as those induced by if-statements
1389and loops) is represented as edges between <tt>CFGBlock</tt>s.
1390Because different C language constructs can induce control-flow,
1391each <tt>CFGBlock</tt> also records an extra <tt>Stmt*</tt> that
1392represents the <i>terminator</i> of the block. A terminator is simply
1393the statement that caused the control-flow, and is used to identify
1394the nature of the conditional control-flow between blocks. For
1395example, in the case of an if-statement, the terminator refers to
1396the <tt>IfStmt</tt> object in the AST that represented the given
1397branch.</p>
1398
1399<p>To illustrate, consider the following code example:</p>
1400
1401<code>
1402int foo(int x) {<br>
1403&nbsp;&nbsp;x = x + 1;<br>
1404<br>
1405&nbsp;&nbsp;if (x > 2) x++;<br>
1406&nbsp;&nbsp;else {<br>
1407&nbsp;&nbsp;&nbsp;&nbsp;x += 2;<br>
1408&nbsp;&nbsp;&nbsp;&nbsp;x *= 2;<br>
1409&nbsp;&nbsp;}<br>
1410<br>
1411&nbsp;&nbsp;return x;<br>
1412}
1413</code>
1414
1415<p>After invoking the parser+semantic analyzer on this code fragment,
1416the AST of the body of <tt>foo</tt> is referenced by a
1417single <tt>Stmt*</tt>. We can then construct an instance
1418of <tt>CFG</tt> representing the control-flow graph of this function
1419body by single call to a static class method:</p>
1420
1421<code>
1422&nbsp;&nbsp;Stmt* FooBody = ...<br>
1423&nbsp;&nbsp;CFG* FooCFG = <b>CFG::buildCFG</b>(FooBody);
1424</code>
1425
1426<p>It is the responsibility of the caller of <tt>CFG::buildCFG</tt>
1427to <tt>delete</tt> the returned <tt>CFG*</tt> when the CFG is no
1428longer needed.</p>
1429
1430<p>Along with providing an interface to iterate over
1431its <tt>CFGBlock</tt>s, the <tt>CFG</tt> class also provides methods
1432that are useful for debugging and visualizing CFGs. For example, the
1433method
1434<tt>CFG::dump()</tt> dumps a pretty-printed version of the CFG to
1435standard error. This is especially useful when one is using a
1436debugger such as gdb. For example, here is the output
1437of <tt>FooCFG->dump()</tt>:</p>
1438
1439<code>
1440&nbsp;[ B5 (ENTRY) ]<br>
1441&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (0):<br>
1442&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B4<br>
1443<br>
1444&nbsp;[ B4 ]<br>
1445&nbsp;&nbsp;&nbsp;&nbsp;1: x = x + 1<br>
1446&nbsp;&nbsp;&nbsp;&nbsp;2: (x > 2)<br>
1447&nbsp;&nbsp;&nbsp;&nbsp;<b>T: if [B4.2]</b><br>
1448&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B5<br>
1449&nbsp;&nbsp;&nbsp;&nbsp;Successors (2): B3 B2<br>
1450<br>
1451&nbsp;[ B3 ]<br>
1452&nbsp;&nbsp;&nbsp;&nbsp;1: x++<br>
1453&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B4<br>
1454&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B1<br>
1455<br>
1456&nbsp;[ B2 ]<br>
1457&nbsp;&nbsp;&nbsp;&nbsp;1: x += 2<br>
1458&nbsp;&nbsp;&nbsp;&nbsp;2: x *= 2<br>
1459&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B4<br>
1460&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B1<br>
1461<br>
1462&nbsp;[ B1 ]<br>
1463&nbsp;&nbsp;&nbsp;&nbsp;1: return x;<br>
1464&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (2): B2 B3<br>
1465&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B0<br>
1466<br>
1467&nbsp;[ B0 (EXIT) ]<br>
1468&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B1<br>
1469&nbsp;&nbsp;&nbsp;&nbsp;Successors (0):
1470</code>
1471
1472<p>For each block, the pretty-printed output displays for each block
1473the number of <i>predecessor</i> blocks (blocks that have outgoing
1474control-flow to the given block) and <i>successor</i> blocks (blocks
1475that have control-flow that have incoming control-flow from the given
1476block). We can also clearly see the special entry and exit blocks at
1477the beginning and end of the pretty-printed output. For the entry
1478block (block B5), the number of predecessor blocks is 0, while for the
1479exit block (block B0) the number of successor blocks is 0.</p>
1480
1481<p>The most interesting block here is B4, whose outgoing control-flow
1482represents the branching caused by the sole if-statement
1483in <tt>foo</tt>. Of particular interest is the second statement in
1484the block, <b><tt>(x > 2)</tt></b>, and the terminator, printed
1485as <b><tt>if [B4.2]</tt></b>. The second statement represents the
1486evaluation of the condition of the if-statement, which occurs before
1487the actual branching of control-flow. Within the <tt>CFGBlock</tt>
1488for B4, the <tt>Stmt*</tt> for the second statement refers to the
1489actual expression in the AST for <b><tt>(x > 2)</tt></b>. Thus
1490pointers to subclasses of <tt>Expr</tt> can appear in the list of
1491statements in a block, and not just subclasses of <tt>Stmt</tt> that
1492refer to proper C statements.</p>
1493
1494<p>The terminator of block B4 is a pointer to the <tt>IfStmt</tt>
1495object in the AST. The pretty-printer outputs <b><tt>if
1496[B4.2]</tt></b> because the condition expression of the if-statement
1497has an actual place in the basic block, and thus the terminator is
1498essentially
1499<i>referring</i> to the expression that is the second statement of
1500block B4 (i.e., B4.2). In this manner, conditions for control-flow
1501(which also includes conditions for loops and switch statements) are
1502hoisted into the actual basic block.</p>
1503
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1504<!-- ===================== -->
1505<!-- <h4>Implicit Control-Flow</h4> -->
1506<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1507
1508<!--
1509<p>A key design principle of the <tt>CFG</tt> class was to not require
1510any transformations to the AST in order to represent control-flow.
1511Thus the <tt>CFG</tt> does not perform any "lowering" of the
1512statements in an AST: loops are not transformed into guarded gotos,
1513short-circuit operations are not converted to a set of if-statements,
1514and so on.</p>
1515-->
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]1516
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1517
1518<!-- ======================================================================= -->
1519<h3 id="Constants">Constant Folding in the Clang AST</h3>
1520<!-- ======================================================================= -->
1521
1522<p>There are several places where constants and constant folding matter a lot to
1523the Clang front-end. First, in general, we prefer the AST to retain the source
1524code as close to how the user wrote it as possible. This means that if they
1525wrote "5+4", we want to keep the addition and two constants in the AST, we don't
1526want to fold to "9". This means that constant folding in various ways turns
1527into a tree walk that needs to handle the various cases.</p>
1528
1529<p>However, there are places in both C and C++ that require constants to be
1530folded. For example, the C standard defines what an "integer constant
1531expression" (i-c-e) is with very precise and specific requirements. The
1532language then requires i-c-e's in a lot of places (for example, the size of a
1533bitfield, the value for a case statement, etc). For these, we have to be able
1534to constant fold the constants, to do semantic checks (e.g. verify bitfield size
1535is non-negative and that case statements aren't duplicated). We aim for Clang
1536to be very pedantic about this, diagnosing cases when the code does not use an
1537i-c-e where one is required, but accepting the code unless running with
1538<tt>-pedantic-errors</tt>.</p>
1539
1540<p>Things get a little bit more tricky when it comes to compatibility with
1541real-world source code. Specifically, GCC has historically accepted a huge
1542superset of expressions as i-c-e's, and a lot of real world code depends on this
1543unfortuate accident of history (including, e.g., the glibc system headers). GCC
1544accepts anything its "fold" optimizer is capable of reducing to an integer
1545constant, which means that the definition of what it accepts changes as its
1546optimizer does. One example is that GCC accepts things like "case X-X:" even
1547when X is a variable, because it can fold this to 0.</p>
1548
1549<p>Another issue are how constants interact with the extensions we support, such
1550as __builtin_constant_p, __builtin_inf, __extension__ and many others. C99
1551obviously does not specify the semantics of any of these extensions, and the
1552definition of i-c-e does not include them. However, these extensions are often
1553used in real code, and we have to have a way to reason about them.</p>
1554
1555<p>Finally, this is not just a problem for semantic analysis. The code
1556generator and other clients have to be able to fold constants (e.g. to
1557initialize global variables) and has to handle a superset of what C99 allows.
1558Further, these clients can benefit from extended information. For example, we
1559know that "foo()||1" always evaluates to true, but we can't replace the
1560expression with true because it has side effects.</p>
1561
1562<!-- ======================= -->
1563<h4>Implementation Approach</h4>
1564<!-- ======================= -->
1565
1566<p>After trying several different approaches, we've finally converged on a
1567design (Note, at the time of this writing, not all of this has been implemented,
1568consider this a design goal!). Our basic approach is to define a single
1569recursive method evaluation method (<tt>Expr::Evaluate</tt>), which is
1570implemented in <tt>AST/ExprConstant.cpp</tt>. Given an expression with 'scalar'
1571type (integer, fp, complex, or pointer) this method returns the following
1572information:</p>
1573
1574<ul>
1575<li>Whether the expression is an integer constant expression, a general
1576 constant that was folded but has no side effects, a general constant that
1577 was folded but that does have side effects, or an uncomputable/unfoldable
1578 value.
1579</li>
1580<li>If the expression was computable in any way, this method returns the APValue
1581 for the result of the expression.</li>
1582<li>If the expression is not evaluatable at all, this method returns
1583 information on one of the problems with the expression. This includes a
1584 SourceLocation for where the problem is, and a diagnostic ID that explains
1585 the problem. The diagnostic should be have ERROR type.</li>
1586<li>If the expression is not an integer constant expression, this method returns
1587 information on one of the problems with the expression. This includes a
1588 SourceLocation for where the problem is, and a diagnostic ID that explains
1589 the problem. The diagnostic should be have EXTENSION type.</li>
1590</ul>
1591
1592<p>This information gives various clients the flexibility that they want, and we
1593will eventually have some helper methods for various extensions. For example,
1594Sema should have a <tt>Sema::VerifyIntegerConstantExpression</tt> method, which
1595calls Evaluate on the expression. If the expression is not foldable, the error
1596is emitted, and it would return true. If the expression is not an i-c-e, the
1597EXTENSION diagnostic is emitted. Finally it would return false to indicate that
1598the AST is ok.</p>
1599
1600<p>Other clients can use the information in other ways, for example, codegen can
1601just use expressions that are foldable in any way.</p>
1602
1603<!-- ========== -->
1604<h4>Extensions</h4>
1605<!-- ========== -->
1606
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]1607<p>This section describes how some of the various extensions Clang supports
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1608interacts with constant evaluation:</p>
1609
1610<ul>
1611<li><b><tt>__extension__</tt></b>: The expression form of this extension causes
1612 any evaluatable subexpression to be accepted as an integer constant
1613 expression.</li>
1614<li><b><tt>__builtin_constant_p</tt></b>: This returns true (as a integer
Chris Lattner28daa532008-12-12 06:55:44 +0000[diff] [blame]1615 constant expression) if the operand is any evaluatable constant. As a
1616 special case, if <tt>__builtin_constant_p</tt> is the (potentially
1617 parenthesized) condition of a conditional operator expression ("?:"), only
Chris Lattner42b83dd2008-12-12 18:00:51 +0000[diff] [blame]1618 the true side of the conditional operator is considered, and it is evaluated
1619 with full constant folding.</li>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1620<li><b><tt>__builtin_choose_expr</tt></b>: The condition is required to be an
1621 integer constant expression, but we accept any constant as an "extension of
1622 an extension". This only evaluates one operand depending on which way the
1623 condition evaluates.</li>
1624<li><b><tt>__builtin_classify_type</tt></b>: This always returns an integer
1625 constant expression.</li>
1626<li><b><tt>__builtin_inf,nan,..</tt></b>: These are treated just like a
1627 floating-point literal.</li>
1628<li><b><tt>__builtin_abs,copysign,..</tt></b>: These are constant folded as
1629 general constant expressions.</li>
1630</ul>
1631
1632
1633
1634
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]1635</div>
1636</body>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1637</html>