Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 4a8dfb511e8f84b2e38b7a86d8ddf05ac1e1a41b / . / docs / InternalsManual.html
blob: da83d89acc61d6eaf506a51435c65e2170260011 [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>
Douglas Gregor715c92a2010-10-27 16:02:28 +0000[diff] [blame]28 <li><a href="#SourceRange">SourceRange and CharSourceRange</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]29 </ul>
30</li>
Daniel Dunbar27d9e9f2009-03-30 06:50:01 +0000[diff] [blame]31<li><a href="#libdriver">The Driver Library</a>
32 <ul>
33 </ul>
34</li>
Douglas Gregor32110df2009-05-20 00:16:32 +0000[diff] [blame]35<li><a href="#pch">Precompiled Headers</a>
Daniel Dunbar27d9e9f2009-03-30 06:50:01 +0000[diff] [blame]36<li><a href="#libfrontend">The Frontend Library</a>
37 <ul>
38 </ul>
39</li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]40<li><a href="#liblex">The Lexer and Preprocessor Library</a>
41 <ul>
42 <li><a href="#Token">The Token class</a></li>
43 <li><a href="#Lexer">The Lexer class</a></li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]44 <li><a href="#AnnotationToken">Annotation Tokens</a></li>
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]45 <li><a href="#TokenLexer">The TokenLexer class</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]46 <li><a href="#MultipleIncludeOpt">The MultipleIncludeOpt class</a></li>
47 </ul>
48</li>
49<li><a href="#libparse">The Parser Library</a>
50 <ul>
51 </ul>
52</li>
53<li><a href="#libast">The AST Library</a>
54 <ul>
55 <li><a href="#Type">The Type class and its subclasses</a></li>
56 <li><a href="#QualType">The QualType class</a></li>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]57 <li><a href="#DeclarationName">Declaration names</a></li>
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]58 <li><a href="#DeclContext">Declaration contexts</a>
59 <ul>
60 <li><a href="#Redeclarations">Redeclarations and Overloads</a></li>
61 <li><a href="#LexicalAndSemanticContexts">Lexical and Semantic
62 Contexts</a></li>
63 <li><a href="#TransparentContexts">Transparent Declaration Contexts</a></li>
64 <li><a href="#MultiDeclContext">Multiply-Defined Declaration Contexts</a></li>
65 </ul>
66 </li>
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]67 <li><a href="#CFG">The CFG class</a></li>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]68 <li><a href="#Constants">Constant Folding in the Clang AST</a></li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]69 </ul>
70</li>
Jeffrey Yasskin28dadd62011-01-28 23:41:54 +0000[diff] [blame]71<li><a href="#Howtos">Howto guides</a>
72 <ul>
73 <li><a href="#AddingAttributes">How to add an attribute</a></li>
Douglas Gregor1f634c62011-09-30 21:32:37 +0000[diff] [blame]74 <li><a href="#AddingExprStmt">How to add a new expression or statement</a></li>
Jeffrey Yasskin28dadd62011-01-28 23:41:54 +0000[diff] [blame]75 </ul>
76</li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]77</ul>
78
79
80<!-- ======================================================================= -->
81<h2 id="intro">Introduction</h2>
82<!-- ======================================================================= -->
83
84<p>This document describes some of the more important APIs and internal design
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]85decisions made in the Clang C front-end. The purpose of this document is to
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]86both capture some of this high level information and also describe some of the
87design decisions behind it. This is meant for people interested in hacking on
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]88Clang, not for end-users. The description below is categorized by
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]89libraries, and does not describe any of the clients of the libraries.</p>
90
91<!-- ======================================================================= -->
92<h2 id="libsystem">LLVM System and Support Libraries</h2>
93<!-- ======================================================================= -->
94
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]95<p>The LLVM libsystem library provides the basic Clang system abstraction layer,
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]96which is used for file system access. The LLVM libsupport library provides many
97underlying libraries and <a
98href="http://llvm.org/docs/ProgrammersManual.html">data-structures</a>,
99 including command line option
100processing and various containers.</p>
101
102<!-- ======================================================================= -->
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]103<h2 id="libbasic">The Clang 'Basic' Library</h2>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]104<!-- ======================================================================= -->
105
106<p>This library certainly needs a better name. The 'basic' library contains a
107number of low-level utilities for tracking and manipulating source buffers,
108locations within the source buffers, diagnostics, tokens, target abstraction,
109and information about the subset of the language being compiled for.</p>
110
111<p>Part of this infrastructure is specific to C (such as the TargetInfo class),
112other parts could be reused for other non-C-based languages (SourceLocation,
113SourceManager, Diagnostics, FileManager). When and if there is future demand
114we can figure out if it makes sense to introduce a new library, move the general
115classes somewhere else, or introduce some other solution.</p>
116
117<p>We describe the roles of these classes in order of their dependencies.</p>
118
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]119
120<!-- ======================================================================= -->
121<h3 id="Diagnostics">The Diagnostics Subsystem</h3>
122<!-- ======================================================================= -->
123
124<p>The Clang Diagnostics subsystem is an important part of how the compiler
125communicates with the human. Diagnostics are the warnings and errors produced
126when the code is incorrect or dubious. In Clang, each diagnostic produced has
Sebastian Redl9bc2a992010-07-07 23:42:27 +0000[diff] [blame]127(at the minimum) a unique ID, an English translation associated with it, a <a
128href="#SourceLocation">SourceLocation</a> to "put the caret", and a severity (e.g.
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]129<tt>WARNING</tt> or <tt>ERROR</tt>). They can also optionally include a number
130of arguments to the dianostic (which fill in "%0"'s in the string) as well as a
131number of source ranges that related to the diagnostic.</p>
132
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]133<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]134driver, but diagnostics can be <a href="#DiagnosticClient">rendered in many
135different ways</a> depending on how the DiagnosticClient interface is
Sebastian Redl9bc2a992010-07-07 23:42:27 +0000[diff] [blame]136implemented. A representative example of a diagnostic is:</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]137
138<pre>
139t.c:38:15: error: invalid operands to binary expression ('int *' and '_Complex float')
140 <font color="darkgreen">P = (P-42) + Gamma*4;</font>
141 <font color="blue">~~~~~~ ^ ~~~~~~~</font>
142</pre>
143
144<p>In this example, you can see the English translation, the severity (error),
145you can see the source location (the caret ("^") and file/line/column info),
146the source ranges "~~~~", arguments to the diagnostic ("int*" and "_Complex
147float"). You'll have to believe me that there is a unique ID backing the
148diagnostic :).</p>
149
150<p>Getting all of this to happen has several steps and involves many moving
151pieces, this section describes them and talks about best practices when adding
152a new diagnostic.</p>
153
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]154<!-- ============================= -->
155<h4>The Diagnostic*Kinds.td files</h4>
156<!-- ============================= -->
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]157
Chris Lattner4c50b692010-05-01 17:35:19 +0000[diff] [blame]158<p>Diagnostics are created by adding an entry to one of the <tt>
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]159clang/Basic/Diagnostic*Kinds.td</tt> files, depending on what library will
160be using it. From this file, tblgen generates the unique ID of the diagnostic,
161the severity of the diagnostic and the English translation + format string.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]162
163<p>There is little sanity with the naming of the unique ID's right now. Some
164start with err_, warn_, ext_ to encode the severity into the name. Since the
165enum is referenced in the C++ code that produces the diagnostic, it is somewhat
166useful for it to be reasonably short.</p>
167
168<p>The severity of the diagnostic comes from the set {<tt>NOTE</tt>,
169<tt>WARNING</tt>, <tt>EXTENSION</tt>, <tt>EXTWARN</tt>, <tt>ERROR</tt>}. The
170<tt>ERROR</tt> severity is used for diagnostics indicating the program is never
171acceptable under any circumstances. When an error is emitted, the AST for the
172input code may not be fully built. The <tt>EXTENSION</tt> and <tt>EXTWARN</tt>
173severities are used for extensions to the language that Clang accepts. This
174means that Clang fully understands and can represent them in the AST, but we
175produce diagnostics to tell the user their code is non-portable. The difference
176is that the former are ignored by default, and the later warn by default. The
177<tt>WARNING</tt> severity is used for constructs that are valid in the currently
178selected source language but that are dubious in some way. The <tt>NOTE</tt>
Daniel Dunbar426b8632009-02-17 15:49:03 +0000[diff] [blame]179level is used to staple more information onto previous diagnostics.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]180
181<p>These <em>severities</em> are mapped into a smaller set (the
182Diagnostic::Level enum, {<tt>Ignored</tt>, <tt>Note</tt>, <tt>Warning</tt>,
Chris Lattner0aad2972009-02-05 22:49:08 +0000[diff] [blame]183<tt>Error</tt>, <tt>Fatal</tt> }) of output <em>levels</em> by the diagnostics
Chris Lattnera180fdd2009-02-17 07:07:29 +0000[diff] [blame]184subsystem based on various configuration options. Clang internally supports a
185fully fine grained mapping mechanism that allows you to map almost any
186diagnostic to the output level that you want. The only diagnostics that cannot
187be mapped are <tt>NOTE</tt>s, which always follow the severity of the previously
188emitted diagnostic and <tt>ERROR</tt>s, which can only be mapped to
189<tt>Fatal</tt> (it is not possible to turn an error into a warning,
190for example).</p>
191
192<p>Diagnostic mappings are used in many ways. For example, if the user
193specifies <tt>-pedantic</tt>, <tt>EXTENSION</tt> maps to <tt>Warning</tt>, if
194they specify <tt>-pedantic-errors</tt>, it turns into <tt>Error</tt>. This is
195used to implement options like <tt>-Wunused_macros</tt>, <tt>-Wundef</tt> etc.
196</p>
197
198<p>
199Mapping to <tt>Fatal</tt> should only be used for diagnostics that are
200considered so severe that error recovery won't be able to recover sensibly from
201them (thus spewing a ton of bogus errors). One example of this class of error
202are failure to #include a file.
203</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]204
205<!-- ================= -->
206<h4>The Format String</h4>
207<!-- ================= -->
208
209<p>The format string for the diagnostic is very simple, but it has some power.
210It takes the form of a string in English with markers that indicate where and
211how arguments to the diagnostic are inserted and formatted. For example, here
212are some simple format strings:</p>
213
214<pre>
215 "binary integer literals are an extension"
216 "format string contains '\\0' within the string body"
217 "more '<b>%%</b>' conversions than data arguments"
Chris Lattner545b3682008-11-23 20:27:13 +0000[diff] [blame]218 "invalid operands to binary expression (<b>%0</b> and <b>%1</b>)"
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]219 "overloaded '<b>%0</b>' must be a <b>%select{unary|binary|unary or binary}2</b> operator"
220 " (has <b>%1</b> parameter<b>%s1</b>)"
221</pre>
222
223<p>These examples show some important points of format strings. You can use any
224 plain ASCII character in the diagnostic string except "%" without a problem,
225 but these are C strings, so you have to use and be aware of all the C escape
226 sequences (as in the second example). If you want to produce a "%" in the
227 output, use the "%%" escape sequence, like the third diagnostic. Finally,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]228 Clang uses the "%...[digit]" sequences to specify where and how arguments to
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]229 the diagnostic are formatted.</p>
230
231<p>Arguments to the diagnostic are numbered according to how they are specified
232 by the C++ code that <a href="#producingdiag">produces them</a>, and are
233 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]234 to your diagnostic, you are doing something wrong :). Unlike printf, there
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]235 is no requirement that arguments to the diagnostic end up in the output in
236 the same order as they are specified, you could have a format string with
237 <tt>"%1 %0"</tt> that swaps them, for example. The text in between the
238 percent and digit are formatting instructions. If there are no instructions,
239 the argument is just turned into a string and substituted in.</p>
240
241<p>Here are some "best practices" for writing the English format string:</p>
242
243<ul>
244<li>Keep the string short. It should ideally fit in the 80 column limit of the
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]245 <tt>DiagnosticKinds.td</tt> file. This avoids the diagnostic wrapping when
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]246 printed, and forces you to think about the important point you are conveying
247 with the diagnostic.</li>
248<li>Take advantage of location information. The user will be able to see the
249 line and location of the caret, so you don't need to tell them that the
250 problem is with the 4th argument to the function: just point to it.</li>
251<li>Do not capitalize the diagnostic string, and do not end it with a
252 period.</li>
253<li>If you need to quote something in the diagnostic string, use single
254 quotes.</li>
255</ul>
256
257<p>Diagnostics should never take random English strings as arguments: you
258shouldn't use <tt>"you have a problem with %0"</tt> and pass in things like
259<tt>"your argument"</tt> or <tt>"your return value"</tt> as arguments. Doing
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]260this prevents <a href="#translation">translating</a> the Clang diagnostics to
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]261other languages (because they'll get random English words in their otherwise
262localized diagnostic). The exceptions to this are C/C++ language keywords
263(e.g. auto, const, mutable, etc) and C/C++ operators (<tt>/=</tt>). Note
264that things like "pointer" and "reference" are not keywords. On the other
265hand, you <em>can</em> include anything that comes from the user's source code,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]266including variable names, types, labels, etc. The 'select' format can be
267used to achieve this sort of thing in a localizable way, see below.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]268
269<!-- ==================================== -->
270<h4>Formatting a Diagnostic Argument</a></h4>
271<!-- ==================================== -->
272
273<p>Arguments to diagnostics are fully typed internally, and come from a couple
274different classes: integers, types, names, and random strings. Depending on
275the class of the argument, it can be optionally formatted in different ways.
276This gives the DiagnosticClient information about what the argument means
277without requiring it to use a specific presentation (consider this MVC for
278Clang :).</p>
279
280<p>Here are the different diagnostic argument formats currently supported by
281Clang:</p>
282
283<table>
284<tr><td colspan="2"><b>"s" format</b></td></tr>
285<tr><td>Example:</td><td><tt>"requires %1 parameter%s1"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]286<tr><td>Class:</td><td>Integers</td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]287<tr><td>Description:</td><td>This is a simple formatter for integers that is
288 useful when producing English diagnostics. When the integer is 1, it prints
289 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]290 simple grammatical forms to be to be handled correctly, and eliminates the
291 need to use gross things like <tt>"requires %1 parameter(s)"</tt>.</td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]292
293<tr><td colspan="2"><b>"select" format</b></td></tr>
294<tr><td>Example:</td><td><tt>"must be a %select{unary|binary|unary or binary}2
295 operator"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]296<tr><td>Class:</td><td>Integers</td></tr>
John McCall3a47e232010-01-14 19:12:17 +0000[diff] [blame]297<tr><td>Description:</td><td><p>This format specifier is used to merge multiple
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]298 related diagnostics together into one common one, without requiring the
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]299 difference to be specified as an English string argument. Instead of
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]300 specifying the string, the diagnostic gets an integer argument and the
301 format string selects the numbered option. In this case, the "%2" value
302 must be an integer in the range [0..2]. If it is 0, it prints 'unary', if
303 it is 1 it prints 'binary' if it is 2, it prints 'unary or binary'. This
304 allows other language translations to substitute reasonable words (or entire
305 phrases) based on the semantics of the diagnostic instead of having to do
John McCall3a47e232010-01-14 19:12:17 +0000[diff] [blame]306 things textually.</p>
307 <p>The selected string does undergo formatting.</p></td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]308
309<tr><td colspan="2"><b>"plural" format</b></td></tr>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]310<tr><td>Example:</td><td><tt>"you have %1 %plural{1:mouse|:mice}1 connected to
311 your computer"</tt></td></tr>
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]312<tr><td>Class:</td><td>Integers</td></tr>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]313<tr><td>Description:</td><td><p>This is a formatter for complex plural forms.
314 It is designed to handle even the requirements of languages with very
315 complex plural forms, as many Baltic languages have. The argument consists
316 of a series of expression/form pairs, separated by ':', where the first form
317 whose expression evaluates to true is the result of the modifier.</p>
318 <p>An expression can be empty, in which case it is always true. See the
319 example at the top. Otherwise, it is a series of one or more numeric
320 conditions, separated by ','. If any condition matches, the expression
321 matches. Each numeric condition can take one of three forms.</p>
322 <ul>
323 <li>number: A simple decimal number matches if the argument is the same
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]324 as the number. Example: <tt>"%plural{1:mouse|:mice}4"</tt></li>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]325 <li>range: A range in square brackets matches if the argument is within
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]326 the range. Then range is inclusive on both ends. Example:
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]327 <tt>"%plural{0:none|1:one|[2,5]:some|:many}2"</tt></li>
328 <li>modulo: A modulo operator is followed by a number, and
329 equals sign and either a number or a range. The tests are the
330 same as for plain
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]331 numbers and ranges, but the argument is taken modulo the number first.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]332 Example: <tt>"%plural{%100=0:even hundred|%100=[1,50]:lower half|:everything
333 else}1"</tt></li>
Sebastian Redl68168562008-11-22 22:16:45 +0000[diff] [blame]334 </ul>
335 <p>The parser is very unforgiving. A syntax error, even whitespace, will
336 abort, as will a failure to match the argument against any
337 expression.</p></td></tr>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]338
John McCall3a47e232010-01-14 19:12:17 +0000[diff] [blame]339<tr><td colspan="2"><b>"ordinal" format</b></td></tr>
340<tr><td>Example:</td><td><tt>"ambiguity in %ordinal0 argument"</tt></td></tr>
341<tr><td>Class:</td><td>Integers</td></tr>
342<tr><td>Description:</td><td><p>This is a formatter which represents the
343 argument number as an ordinal: the value <tt>1</tt> becomes <tt>1st</tt>,
344 <tt>3</tt> becomes <tt>3rd</tt>, and so on. Values less than <tt>1</tt>
345 are not supported.</p>
346 <p>This formatter is currently hard-coded to use English ordinals.</p></td></tr>
347
Chris Lattner077bf5e2008-11-24 03:33:13 +0000[diff] [blame]348<tr><td colspan="2"><b>"objcclass" format</b></td></tr>
349<tr><td>Example:</td><td><tt>"method %objcclass0 not found"</tt></td></tr>
350<tr><td>Class:</td><td>DeclarationName</td></tr>
351<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
352 DeclarationName corresponds to an Objective-C class method selector. As
353 such, it prints the selector with a leading '+'.</p></td></tr>
354
355<tr><td colspan="2"><b>"objcinstance" format</b></td></tr>
356<tr><td>Example:</td><td><tt>"method %objcinstance0 not found"</tt></td></tr>
357<tr><td>Class:</td><td>DeclarationName</td></tr>
358<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
359 DeclarationName corresponds to an Objective-C instance method selector. As
360 such, it prints the selector with a leading '-'.</p></td></tr>
361
Douglas Gregor47b9a1c2009-02-04 17:27:36 +0000[diff] [blame]362<tr><td colspan="2"><b>"q" format</b></td></tr>
363<tr><td>Example:</td><td><tt>"candidate found by name lookup is %q0"</tt></td></tr>
364<tr><td>Class:</td><td>NamedDecl*</td></tr>
365<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>
366
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]367</table>
368
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]369<p>It is really easy to add format specifiers to the Clang diagnostics system,
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]370but they should be discussed before they are added. If you are creating a lot
371of repetitive diagnostics and/or have an idea for a useful formatter, please
372bring it up on the cfe-dev mailing list.</p>
Chris Lattnercc543342008-11-22 23:50:47 +0000[diff] [blame]373
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]374<!-- ===================================================== -->
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]375<h4 id="producingdiag">Producing the Diagnostic</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]376<!-- ===================================================== -->
377
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]378<p>Now that you've created the diagnostic in the DiagnosticKinds.td file, you
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]379need to write the code that detects the condition in question and emits the
380new diagnostic. Various components of Clang (e.g. the preprocessor, Sema,
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]381etc) provide a helper function named "Diag". It creates a diagnostic and
382accepts the arguments, ranges, and other information that goes along with
383it.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]384
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]385<p>For example, the binary expression error comes from code like this:</p>
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]386
387<pre>
388 if (various things that are bad)
389 Diag(Loc, diag::err_typecheck_invalid_operands)
390 &lt;&lt; lex-&gt;getType() &lt;&lt; rex-&gt;getType()
391 &lt;&lt; lex-&gt;getSourceRange() &lt;&lt; rex-&gt;getSourceRange();
392</pre>
393
394<p>This shows that use of the Diag method: they take a location (a <a
395href="#SourceLocation">SourceLocation</a> object) and a diagnostic enum value
Chris Lattnercc2ac1e2011-02-14 06:42:50 +0000[diff] [blame]396(which matches the name from DiagnosticKinds.td). If the diagnostic takes
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]397arguments, they are specified with the &lt;&lt; operator: the first argument
398becomes %0, the second becomes %1, etc. The diagnostic interface allows you to
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]399specify arguments of many different types, including <tt>int</tt> and
400<tt>unsigned</tt> for integer arguments, <tt>const char*</tt> and
401<tt>std::string</tt> for string arguments, <tt>DeclarationName</tt> and
402<tt>const IdentifierInfo*</tt> for names, <tt>QualType</tt> for types, etc.
403SourceRanges are also specified with the &lt;&lt; operator, but do not have a
404specific ordering requirement.</p>
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]405
406<p>As you can see, adding and producing a diagnostic is pretty straightforward.
407The hard part is deciding exactly what you need to say to help the user, picking
408a suitable wording, and providing the information needed to format it correctly.
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]409The good news is that the call site that issues a diagnostic should be
410completely independent of how the diagnostic is formatted and in what language
411it is rendered.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]412</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]413
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]414<!-- ==================================================== -->
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]415<h4 id="fix-it-hints">Fix-It Hints</h4>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]416<!-- ==================================================== -->
417
418<p>In some cases, the front end emits diagnostics when it is clear
419that some small change to the source code would fix the problem. For
420example, a missing semicolon at the end of a statement or a use of
Chris Lattner34c05332009-02-27 19:31:12 +0000[diff] [blame]421deprecated syntax that is easily rewritten into a more modern form.
422Clang tries very hard to emit the diagnostic and recover gracefully
423in these and other cases.</p>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]424
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]425<p>However, for these cases where the fix is obvious, the diagnostic
426can be annotated with a hint (referred to as a "fix-it hint") that
427describes how to change the code referenced by the diagnostic to fix
428the problem. For example, it might add the missing semicolon at the
429end of the statement or rewrite the use of a deprecated construct
430into something more palatable. Here is one such example from the C++
431front end, where we warn about the right-shift operator changing
432meaning from C++98 to C++0x:</p>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]433
434<pre>
435test.cpp:3:7: warning: use of right-shift operator ('&gt;&gt;') in template argument will require parentheses in C++0x
436A&lt;100 &gt;&gt; 2&gt; *a;
437 ^
438 ( )
439</pre>
440
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]441<p>Here, the fix-it hint is suggesting that parentheses be added,
442and showing exactly where those parentheses would be inserted into the
443source code. The fix-it hints themselves describe what changes to make
444to the source code in an abstract manner, which the text diagnostic
445printer renders as a line of "insertions" below the caret line. <a
446href="#DiagnosticClient">Other diagnostic clients</a> might choose
447to render the code differently (e.g., as markup inline) or even give
448the user the ability to automatically fix the problem.</p>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]449
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]450<p>All fix-it hints are described by the <code>FixItHint</code> class,
451instances of which should be attached to the diagnostic using the
452&lt;&lt; operator in the same way that highlighted source ranges and
453arguments are passed to the diagnostic. Fix-it hints can be created
454with one of three constructors:</p>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]455
456<dl>
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]457 <dt><code>FixItHint::CreateInsertion(Loc, Code)</code></dt>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]458 <dd>Specifies that the given <code>Code</code> (a string) should be inserted
459 before the source location <code>Loc</code>.</dd>
460
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]461 <dt><code>FixItHint::CreateRemoval(Range)</code></dt>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]462 <dd>Specifies that the code in the given source <code>Range</code>
463 should be removed.</dd>
464
Peter Collingbourne38448d32011-03-21 01:45:18 +0000[diff] [blame]465 <dt><code>FixItHint::CreateReplacement(Range, Code)</code></dt>
Douglas Gregorb2fb6de2009-02-27 17:53:17 +0000[diff] [blame]466 <dd>Specifies that the code in the given source <code>Range</code>
467 should be removed, and replaced with the given <code>Code</code> string.</dd>
468</dl>
469
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]470<!-- ============================================================= -->
471<h4><a name="DiagnosticClient">The DiagnosticClient Interface</a></h4>
472<!-- ============================================================= -->
473
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]474<p>Once code generates a diagnostic with all of the arguments and the rest of
475the relevant information, Clang needs to know what to do with it. As previously
476mentioned, the diagnostic machinery goes through some filtering to map a
477severity onto a diagnostic level, then (assuming the diagnostic is not mapped to
478"<tt>Ignore</tt>") it invokes an object that implements the DiagnosticClient
479interface with the information.</p>
480
481<p>It is possible to implement this interface in many different ways. For
482example, the normal Clang DiagnosticClient (named 'TextDiagnosticPrinter') turns
483the arguments into strings (according to the various formatting rules), prints
484out the file/line/column information and the string, then prints out the line of
485code, the source ranges, and the caret. However, this behavior isn't required.
486</p>
487
488<p>Another implementation of the DiagnosticClient interface is the
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]489'TextDiagnosticBuffer' class, which is used when Clang is in -verify mode.
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]490Instead of formatting and printing out the diagnostics, this implementation just
491captures and remembers the diagnostics as they fly by. Then -verify compares
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]492the list of produced diagnostics to the list of expected ones. If they disagree,
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]493it prints out its own output.
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]494</p>
495
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]496<p>There are many other possible implementations of this interface, and this is
497why we prefer diagnostics to pass down rich structured information in arguments.
498For example, an HTML output might want declaration names be linkified to where
499they come from in the source. Another example is that a GUI might let you click
500on typedefs to expand them. This application would want to pass significantly
501more information about types through to the GUI than a simple flat string. The
502interface allows this to happen.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]503
504<!-- ====================================================== -->
505<h4><a name="translation">Adding Translations to Clang</a></h4>
506<!-- ====================================================== -->
507
Chris Lattner627b7052008-11-23 00:28:33 +0000[diff] [blame]508<p>Not possible yet! Diagnostic strings should be written in UTF-8, the client
Chris Lattnerfd408ea2008-11-23 00:42:53 +0000[diff] [blame]509can translate to the relevant code page if needed. Each translation completely
510replaces the format string for the diagnostic.</p>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]511
512
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]513<!-- ======================================================================= -->
514<h3 id="SourceLocation">The SourceLocation and SourceManager classes</h3>
515<!-- ======================================================================= -->
516
517<p>Strangely enough, the SourceLocation class represents a location within the
518source code of the program. Important design points include:</p>
519
520<ol>
521<li>sizeof(SourceLocation) must be extremely small, as these are embedded into
522 many AST nodes and are passed around often. Currently it is 32 bits.</li>
523<li>SourceLocation must be a simple value object that can be efficiently
524 copied.</li>
525<li>We should be able to represent a source location for any byte of any input
526 file. This includes in the middle of tokens, in whitespace, in trigraphs,
527 etc.</li>
528<li>A SourceLocation must encode the current #include stack that was active when
529 the location was processed. For example, if the location corresponds to a
530 token, it should contain the set of #includes active when the token was
531 lexed. This allows us to print the #include stack for a diagnostic.</li>
532<li>SourceLocation must be able to describe macro expansions, capturing both
533 the ultimate instantiation point and the source of the original character
534 data.</li>
535</ol>
536
537<p>In practice, the SourceLocation works together with the SourceManager class
Nick Lewycky77561e52010-05-26 21:48:10 +0000[diff] [blame]538to encode two pieces of information about a location: its spelling location
539and its instantiation location. For most tokens, these will be the same.
540However, for a macro expansion (or tokens that came from a _Pragma directive)
541these will describe the location of the characters corresponding to the token
542and the location where the token was used (i.e. the macro instantiation point
543or the location of the _Pragma itself).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]544
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]545<p>The Clang front-end inherently depends on the location of a token being
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]546tracked correctly. If it is ever incorrect, the front-end may get confused and
547die. 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]548Clang depends on being able to find the original input characters for the token.
Chris Lattner18376dd2009-01-16 07:00:50 +0000[diff] [blame]549This concept maps directly to the "spelling location" for the token.</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]550
Douglas Gregor715c92a2010-10-27 16:02:28 +0000[diff] [blame]551
552<!-- ======================================================================= -->
553<h3 id="SourceRange">SourceRange and CharSourceRange</h3>
554<!-- ======================================================================= -->
555<!-- mostly taken from
556 http://lists.cs.uiuc.edu/pipermail/cfe-dev/2010-August/010595.html -->
557
558<p>Clang represents most source ranges by [first, last], where first and last
559each point to the beginning of their respective tokens. For example
560consider the SourceRange of the following statement:</p>
561<pre>
562x = foo + bar;
563^first ^last
564</pre>
565
566<p>To map from this representation to a character-based
567representation, the 'last' location needs to be adjusted to point to
568(or past) the end of that token with either
569<code>Lexer::MeasureTokenLength()</code> or
Chris Lattner7ef5c272010-11-17 07:05:50 +0000[diff] [blame]570<code>Lexer::getLocForEndOfToken()</code>. For the rare cases
Douglas Gregor715c92a2010-10-27 16:02:28 +0000[diff] [blame]571where character-level source ranges information is needed we use
572the <code>CharSourceRange</code> class.</p>
573
574
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]575<!-- ======================================================================= -->
Daniel Dunbar27d9e9f2009-03-30 06:50:01 +0000[diff] [blame]576<h2 id="libdriver">The Driver Library</h2>
577<!-- ======================================================================= -->
578
Ted Kremenekcfa8d572009-04-09 18:08:18 +0000[diff] [blame]579<p>The clang Driver and library are documented <a
580href="DriverInternals.html">here<a>.<p>
581
582<!-- ======================================================================= -->
Douglas Gregor32110df2009-05-20 00:16:32 +0000[diff] [blame]583<h2 id="pch">Precompiled Headers</h2>
Ted Kremenekcfa8d572009-04-09 18:08:18 +0000[diff] [blame]584<!-- ======================================================================= -->
585
Douglas Gregor32110df2009-05-20 00:16:32 +0000[diff] [blame]586<p>Clang supports two implementations of precompiled headers. The
587 default implementation, precompiled headers (<a
588 href="PCHInternals.html">PCH</a>) uses a serialized representation
589 of Clang's internal data structures, encoded with the <a
590 href="http://llvm.org/docs/BitCodeFormat.html">LLVM bitstream
591 format</a>. Pretokenized headers (<a
592 href="PTHInternals.html">PTH</a>), on the other hand, contain a
593 serialized representation of the tokens encountered when
594 preprocessing a header (and anything that header includes).</p>
595
Daniel Dunbar27d9e9f2009-03-30 06:50:01 +0000[diff] [blame]596
597<!-- ======================================================================= -->
598<h2 id="libfrontend">The Frontend Library</h2>
599<!-- ======================================================================= -->
600
601<p>The Frontend library contains functionality useful for building
602tools on top of the clang libraries, for example several methods for
603outputting diagnostics.</p>
604
605<!-- ======================================================================= -->
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]606<h2 id="liblex">The Lexer and Preprocessor Library</h2>
607<!-- ======================================================================= -->
608
609<p>The Lexer library contains several tightly-connected classes that are involved
610with the nasty process of lexing and preprocessing C source code. The main
611interface to this library for outside clients is the large <a
612href="#Preprocessor">Preprocessor</a> class.
613It contains the various pieces of state that are required to coherently read
614tokens out of a translation unit.</p>
615
616<p>The core interface to the Preprocessor object (once it is set up) is the
617Preprocessor::Lex method, which returns the next <a href="#Token">Token</a> from
618the preprocessor stream. There are two types of token providers that the
619preprocessor is capable of reading from: a buffer lexer (provided by the <a
620href="#Lexer">Lexer</a> class) and a buffered token stream (provided by the <a
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]621href="#TokenLexer">TokenLexer</a> class).
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]622
623
624<!-- ======================================================================= -->
625<h3 id="Token">The Token class</h3>
626<!-- ======================================================================= -->
627
628<p>The Token class is used to represent a single lexed token. Tokens are
629intended to be used by the lexer/preprocess and parser libraries, but are not
630intended to live beyond them (for example, they should not live in the ASTs).<p>
631
632<p>Tokens most often live on the stack (or some other location that is efficient
633to access) as the parser is running, but occasionally do get buffered up. For
634example, macro definitions are stored as a series of tokens, and the C++
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]635front-end periodically needs to buffer tokens up for tentative parsing and
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]636various pieces of look-ahead. As such, the size of a Token matter. On a 32-bit
637system, sizeof(Token) is currently 16 bytes.</p>
638
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]639<p>Tokens occur in two forms: "<a href="#AnnotationToken">Annotation
640Tokens</a>" and normal tokens. Normal tokens are those returned by the lexer,
641annotation tokens represent semantic information and are produced by the parser,
642replacing normal tokens in the token stream. Normal tokens contain the
643following information:</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]644
645<ul>
646<li><b>A SourceLocation</b> - This indicates the location of the start of the
647token.</li>
648
649<li><b>A length</b> - This stores the length of the token as stored in the
650SourceBuffer. For tokens that include them, this length includes trigraphs and
651escaped newlines which are ignored by later phases of the compiler. By pointing
652into the original source buffer, it is always possible to get the original
653spelling of a token completely accurately.</li>
654
655<li><b>IdentifierInfo</b> - If a token takes the form of an identifier, and if
656identifier lookup was enabled when the token was lexed (e.g. the lexer was not
657reading in 'raw' mode) this contains a pointer to the unique hash value for the
658identifier. Because the lookup happens before keyword identification, this
659field is set even for language keywords like 'for'.</li>
660
661<li><b>TokenKind</b> - This indicates the kind of token as classified by the
662lexer. This includes things like <tt>tok::starequal</tt> (for the "*="
663operator), <tt>tok::ampamp</tt> for the "&amp;&amp;" token, and keyword values
664(e.g. <tt>tok::kw_for</tt>) for identifiers that correspond to keywords. Note
665that some tokens can be spelled multiple ways. For example, C++ supports
666"operator keywords", where things like "and" are treated exactly like the
667"&amp;&amp;" operator. In these cases, the kind value is set to
668<tt>tok::ampamp</tt>, which is good for the parser, which doesn't have to
669consider both forms. For something that cares about which form is used (e.g.
670the preprocessor 'stringize' operator) the spelling indicates the original
671form.</li>
672
673<li><b>Flags</b> - There are currently four flags tracked by the
674lexer/preprocessor system on a per-token basis:
675
676 <ol>
677 <li><b>StartOfLine</b> - This was the first token that occurred on its input
678 source line.</li>
679 <li><b>LeadingSpace</b> - There was a space character either immediately
680 before the token or transitively before the token as it was expanded
681 through a macro. The definition of this flag is very closely defined by
682 the stringizing requirements of the preprocessor.</li>
683 <li><b>DisableExpand</b> - This flag is used internally to the preprocessor to
684 represent identifier tokens which have macro expansion disabled. This
685 prevents them from being considered as candidates for macro expansion ever
686 in the future.</li>
687 <li><b>NeedsCleaning</b> - This flag is set if the original spelling for the
688 token includes a trigraph or escaped newline. Since this is uncommon,
689 many pieces of code can fast-path on tokens that did not need cleaning.
690 </p>
691 </ol>
692</li>
693</ul>
694
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]695<p>One interesting (and somewhat unusual) aspect of normal tokens is that they
696don't contain any semantic information about the lexed value. For example, if
697the token was a pp-number token, we do not represent the value of the number
698that was lexed (this is left for later pieces of code to decide). Additionally,
699the lexer library has no notion of typedef names vs variable names: both are
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]700returned as identifiers, and the parser is left to decide whether a specific
701identifier is a typedef or a variable (tracking this requires scope information
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]702among other things). The parser can do this translation by replacing tokens
703returned by the preprocessor with "Annotation Tokens".</p>
704
705<!-- ======================================================================= -->
706<h3 id="AnnotationToken">Annotation Tokens</h3>
707<!-- ======================================================================= -->
708
709<p>Annotation Tokens are tokens that are synthesized by the parser and injected
710into the preprocessor's token stream (replacing existing tokens) to record
711semantic information found by the parser. For example, if "foo" is found to be
712a typedef, the "foo" <tt>tok::identifier</tt> token is replaced with an
713<tt>tok::annot_typename</tt>. This is useful for a couple of reasons: 1) this
714makes it easy to handle qualified type names (e.g. "foo::bar::baz&lt;42&gt;::t")
715in C++ as a single "token" in the parser. 2) if the parser backtracks, the
716reparse does not need to redo semantic analysis to determine whether a token
717sequence is a variable, type, template, etc.</p>
718
719<p>Annotation Tokens are created by the parser and reinjected into the parser's
720token stream (when backtracking is enabled). Because they can only exist in
721tokens that the preprocessor-proper is done with, it doesn't need to keep around
722flags like "start of line" that the preprocessor uses to do its job.
723Additionally, an annotation token may "cover" a sequence of preprocessor tokens
724(e.g. <tt>a::b::c</tt> is five preprocessor tokens). As such, the valid fields
725of an annotation token are different than the fields for a normal token (but
726they are multiplexed into the normal Token fields):</p>
727
728<ul>
729<li><b>SourceLocation "Location"</b> - The SourceLocation for the annotation
730token indicates the first token replaced by the annotation token. In the example
731above, it would be the location of the "a" identifier.</li>
732
733<li><b>SourceLocation "AnnotationEndLoc"</b> - This holds the location of the
734last token replaced with the annotation token. In the example above, it would
735be the location of the "c" identifier.</li>
736
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]737<li><b>void* "AnnotationValue"</b> - This contains an opaque object
738that the parser gets from Sema. The parser merely preserves the
739information for Sema to later interpret based on the annotation token
740kind.</li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]741
742<li><b>TokenKind "Kind"</b> - This indicates the kind of Annotation token this
743is. See below for the different valid kinds.</li>
744</ul>
745
746<p>Annotation tokens currently come in three kinds:</p>
747
748<ol>
749<li><b>tok::annot_typename</b>: This annotation token represents a
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]750resolved typename token that is potentially qualified. The
751AnnotationValue field contains the <tt>QualType</tt> returned by
752Sema::getTypeName(), possibly with source location information
753attached.</li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]754
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]755<li><b>tok::annot_cxxscope</b>: This annotation token represents a C++
756scope specifier, such as "A::B::". This corresponds to the grammar
757productions "::" and ":: [opt] nested-name-specifier". The
758AnnotationValue pointer is a <tt>NestedNameSpecifier*</tt> returned by
759the Sema::ActOnCXXGlobalScopeSpecifier and
760Sema::ActOnCXXNestedNameSpecifier callbacks.</li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]761
Douglas Gregor39a8de12009-02-25 19:37:18 +0000[diff] [blame]762<li><b>tok::annot_template_id</b>: This annotation token represents a
763C++ template-id such as "foo&lt;int, 4&gt;", where "foo" is the name
764of a template. The AnnotationValue pointer is a pointer to a malloc'd
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]765TemplateIdAnnotation object. Depending on the context, a parsed
766template-id that names a type might become a typename annotation token
767(if all we care about is the named type, e.g., because it occurs in a
768type specifier) or might remain a template-id token (if we want to
769retain more source location information or produce a new type, e.g.,
770in a declaration of a class template specialization). template-id
771annotation tokens that refer to a type can be "upgraded" to typename
772annotation tokens by the parser.</li>
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]773
774</ol>
775
Cedric Venetda76b282009-01-06 16:22:54 +0000[diff] [blame]776<p>As mentioned above, annotation tokens are not returned by the preprocessor,
Chris Lattner3932fe02009-01-06 06:02:08 +0000[diff] [blame]777they are formed on demand by the parser. This means that the parser has to be
778aware of cases where an annotation could occur and form it where appropriate.
779This is somewhat similar to how the parser handles Translation Phase 6 of C99:
780String Concatenation (see C99 5.1.1.2). In the case of string concatenation,
781the preprocessor just returns distinct tok::string_literal and
782tok::wide_string_literal tokens and the parser eats a sequence of them wherever
783the grammar indicates that a string literal can occur.</p>
784
785<p>In order to do this, whenever the parser expects a tok::identifier or
786tok::coloncolon, it should call the TryAnnotateTypeOrScopeToken or
787TryAnnotateCXXScopeToken methods to form the annotation token. These methods
788will maximally form the specified annotation tokens and replace the current
789token with them, if applicable. If the current tokens is not valid for an
790annotation token, it will remain an identifier or :: token.</p>
791
792
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]793
794<!-- ======================================================================= -->
795<h3 id="Lexer">The Lexer class</h3>
796<!-- ======================================================================= -->
797
798<p>The Lexer class provides the mechanics of lexing tokens out of a source
799buffer and deciding what they mean. The Lexer is complicated by the fact that
800it operates on raw buffers that have not had spelling eliminated (this is a
801necessity to get decent performance), but this is countered with careful coding
802as well as standard performance techniques (for example, the comment handling
803code is vectorized on X86 and PowerPC hosts).</p>
804
805<p>The lexer has a couple of interesting modal features:</p>
806
807<ul>
808<li>The lexer can operate in 'raw' mode. This mode has several features that
809 make it possible to quickly lex the file (e.g. it stops identifier lookup,
810 doesn't specially handle preprocessor tokens, handles EOF differently, etc).
811 This mode is used for lexing within an "<tt>#if 0</tt>" block, for
812 example.</li>
813<li>The lexer can capture and return comments as tokens. This is required to
814 support the -C preprocessor mode, which passes comments through, and is
815 used by the diagnostic checker to identifier expect-error annotations.</li>
816<li>The lexer can be in ParsingFilename mode, which happens when preprocessing
Chris Lattner84386242007-09-16 19:25:23 +0000[diff] [blame]817 after reading a #include directive. This mode changes the parsing of '&lt;'
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]818 to return an "angled string" instead of a bunch of tokens for each thing
819 within the filename.</li>
820<li>When parsing a preprocessor directive (after "<tt>#</tt>") the
821 ParsingPreprocessorDirective mode is entered. This changes the parser to
Peter Collingbourne84021552011-02-28 02:37:51 +0000[diff] [blame]822 return EOD at a newline.</li>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]823<li>The Lexer uses a LangOptions object to know whether trigraphs are enabled,
824 whether C++ or ObjC keywords are recognized, etc.</li>
825</ul>
826
827<p>In addition to these modes, the lexer keeps track of a couple of other
828 features that are local to a lexed buffer, which change as the buffer is
829 lexed:</p>
830
831<ul>
832<li>The Lexer uses BufferPtr to keep track of the current character being
833 lexed.</li>
834<li>The Lexer uses IsAtStartOfLine to keep track of whether the next lexed token
835 will start with its "start of line" bit set.</li>
836<li>The Lexer keeps track of the current #if directives that are active (which
837 can be nested).</li>
838<li>The Lexer keeps track of an <a href="#MultipleIncludeOpt">
839 MultipleIncludeOpt</a> object, which is used to
840 detect whether the buffer uses the standard "<tt>#ifndef XX</tt> /
841 <tt>#define XX</tt>" idiom to prevent multiple inclusion. If a buffer does,
842 subsequent includes can be ignored if the XX macro is defined.</li>
843</ul>
844
845<!-- ======================================================================= -->
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]846<h3 id="TokenLexer">The TokenLexer class</h3>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]847<!-- ======================================================================= -->
848
Chris Lattner79281252008-03-09 02:27:26 +0000[diff] [blame]849<p>The TokenLexer class is a token provider that returns tokens from a list
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]850of tokens that came from somewhere else. It typically used for two things: 1)
851returning tokens from a macro definition as it is being expanded 2) returning
852tokens from an arbitrary buffer of tokens. The later use is used by _Pragma and
853will most likely be used to handle unbounded look-ahead for the C++ parser.</p>
854
855<!-- ======================================================================= -->
856<h3 id="MultipleIncludeOpt">The MultipleIncludeOpt class</h3>
857<!-- ======================================================================= -->
858
859<p>The MultipleIncludeOpt class implements a really simple little state machine
860that is used to detect the standard "<tt>#ifndef XX</tt> / <tt>#define XX</tt>"
861idiom that people typically use to prevent multiple inclusion of headers. If a
862buffer uses this idiom and is subsequently #include'd, the preprocessor can
863simply check to see whether the guarding condition is defined or not. If so,
864the preprocessor can completely ignore the include of the header.</p>
865
866
867
868<!-- ======================================================================= -->
869<h2 id="libparse">The Parser Library</h2>
870<!-- ======================================================================= -->
871
872<!-- ======================================================================= -->
873<h2 id="libast">The AST Library</h2>
874<!-- ======================================================================= -->
875
876<!-- ======================================================================= -->
877<h3 id="Type">The Type class and its subclasses</h3>
878<!-- ======================================================================= -->
879
880<p>The Type class (and its subclasses) are an important part of the AST. Types
881are accessed through the ASTContext class, which implicitly creates and uniques
882them as they are needed. Types have a couple of non-obvious features: 1) they
883do not capture type qualifiers like const or volatile (See
884<a href="#QualType">QualType</a>), and 2) they implicitly capture typedef
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]885information. Once created, types are immutable (unlike decls).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]886
887<p>Typedefs in C make semantic analysis a bit more complex than it would
888be without them. The issue is that we want to capture typedef information
889and represent it in the AST perfectly, but the semantics of operations need to
890"see through" typedefs. For example, consider this code:</p>
891
892<code>
893void func() {<br>
Bill Wendling30d17752007-10-06 01:56:01 +0000[diff] [blame]894&nbsp;&nbsp;typedef int foo;<br>
895&nbsp;&nbsp;foo X, *Y;<br>
896&nbsp;&nbsp;typedef foo* bar;<br>
897&nbsp;&nbsp;bar Z;<br>
898&nbsp;&nbsp;*X; <i>// error</i><br>
899&nbsp;&nbsp;**Y; <i>// error</i><br>
900&nbsp;&nbsp;**Z; <i>// error</i><br>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]901}<br>
902</code>
903
904<p>The code above is illegal, and thus we expect there to be diagnostics emitted
905on the annotated lines. In this example, we expect to get:</p>
906
907<pre>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]908<b>test.c:6:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]909*X; // error
910<font color="blue">^~</font>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]911<b>test.c:7:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]912**Y; // error
913<font color="blue">^~~</font>
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]914<b>test.c:8:1: error: indirection requires pointer operand ('foo' invalid)</b>
915**Z; // error
916<font color="blue">^~~</font>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]917</pre>
918
919<p>While this example is somewhat silly, it illustrates the point: we want to
920retain typedef information where possible, so that we can emit errors about
921"<tt>std::string</tt>" instead of "<tt>std::basic_string&lt;char, std:...</tt>".
922Doing this requires properly keeping typedef information (for example, the type
923of "X" is "foo", not "int"), and requires properly propagating it through the
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]924various operators (for example, the type of *Y is "foo", not "int"). In order
925to retain this information, the type of these expressions is an instance of the
926TypedefType class, which indicates that the type of these expressions is a
927typedef for foo.
928</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]929
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]930<p>Representing types like this is great for diagnostics, because the
931user-specified type is always immediately available. There are two problems
932with this: first, various semantic checks need to make judgements about the
Chris Lattner33fc68a2007-07-31 18:54:50 +0000[diff] [blame]933<em>actual structure</em> of a type, ignoring typdefs. Second, we need an
934efficient way to query whether two types are structurally identical to each
935other, ignoring typedefs. The solution to both of these problems is the idea of
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]936canonical types.</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]937
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]938<!-- =============== -->
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]939<h4>Canonical Types</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]940<!-- =============== -->
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]941
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]942<p>Every instance of the Type class contains a canonical type pointer. For
943simple types with no typedefs involved (e.g. "<tt>int</tt>", "<tt>int*</tt>",
944"<tt>int**</tt>"), the type just points to itself. For types that have a
945typedef somewhere in their structure (e.g. "<tt>foo</tt>", "<tt>foo*</tt>",
946"<tt>foo**</tt>", "<tt>bar</tt>"), the canonical type pointer points to their
947structurally equivalent type without any typedefs (e.g. "<tt>int</tt>",
948"<tt>int*</tt>", "<tt>int**</tt>", and "<tt>int*</tt>" respectively).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]949
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]950<p>This design provides a constant time operation (dereferencing the canonical
951type pointer) that gives us access to the structure of types. For example,
952we can trivially tell that "bar" and "foo*" are the same type by dereferencing
953their canonical type pointers and doing a pointer comparison (they both point
954to the single "<tt>int*</tt>" type).</p>
955
956<p>Canonical types and typedef types bring up some complexities that must be
957carefully managed. Specifically, the "isa/cast/dyncast" operators generally
958shouldn't be used in code that is inspecting the AST. For example, when type
959checking the indirection operator (unary '*' on a pointer), the type checker
960must verify that the operand has a pointer type. It would not be correct to
961check that with "<tt>isa&lt;PointerType&gt;(SubExpr-&gt;getType())</tt>",
962because this predicate would fail if the subexpression had a typedef type.</p>
963
964<p>The solution to this problem are a set of helper methods on Type, used to
965check their properties. In this case, it would be correct to use
966"<tt>SubExpr-&gt;getType()-&gt;isPointerType()</tt>" to do the check. This
967predicate will return true if the <em>canonical type is a pointer</em>, which is
968true any time the type is structurally a pointer type. The only hard part here
969is remembering not to use the <tt>isa/cast/dyncast</tt> operations.</p>
970
971<p>The second problem we face is how to get access to the pointer type once we
972know it exists. To continue the example, the result type of the indirection
973operator is the pointee type of the subexpression. In order to determine the
974type, we need to get the instance of PointerType that best captures the typedef
975information in the program. If the type of the expression is literally a
976PointerType, we can return that, otherwise we have to dig through the
977typedefs to find the pointer type. For example, if the subexpression had type
978"<tt>foo*</tt>", we could return that type as the result. If the subexpression
979had type "<tt>bar</tt>", we want to return "<tt>foo*</tt>" (note that we do
980<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]981a getAsPointerType() method that checks whether the type is structurally a
Chris Lattner8a2bc622007-07-31 06:37:39 +0000[diff] [blame]982PointerType and, if so, returns the best one. If not, it returns a null
983pointer.</p>
984
985<p>This structure is somewhat mystical, but after meditating on it, it will
986make sense to you :).</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]987
988<!-- ======================================================================= -->
989<h3 id="QualType">The QualType class</h3>
990<!-- ======================================================================= -->
991
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]992<p>The QualType class is designed as a trivial value class that is
993small, passed by-value and is efficient to query. The idea of
994QualType is that it stores the type qualifiers (const, volatile,
995restrict, plus some extended qualifiers required by language
996extensions) separately from the types themselves. QualType is
997conceptually a pair of "Type*" and the bits for these type qualifiers.</p>
Chris Lattner86920d32007-07-31 05:42:17 +0000[diff] [blame]998
999<p>By storing the type qualifiers as bits in the conceptual pair, it is
1000extremely efficient to get the set of qualifiers on a QualType (just return the
1001field of the pair), add a type qualifier (which is a trivial constant-time
1002operation that sets a bit), and remove one or more type qualifiers (just return
1003a QualType with the bitfield set to empty).</p>
1004
1005<p>Further, because the bits are stored outside of the type itself, we do not
1006need to create duplicates of types with different sets of qualifiers (i.e. there
1007is only a single heap allocated "int" type: "const int" and "volatile const int"
1008both point to the same heap allocated "int" type). This reduces the heap size
1009used to represent bits and also means we do not have to consider qualifiers when
1010uniquing types (<a href="#Type">Type</a> does not even contain qualifiers).</p>
1011
John McCall027ac442010-09-03 05:07:55 +0000[diff] [blame]1012<p>In practice, the two most common type qualifiers (const and
1013restrict) are stored in the low bits of the pointer to the Type
1014object, together with a flag indicating whether extended qualifiers
1015are present (which must be heap-allocated). This means that QualType
1016is exactly the same size as a pointer.</p>
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1017
1018<!-- ======================================================================= -->
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1019<h3 id="DeclarationName">Declaration names</h3>
1020<!-- ======================================================================= -->
1021
1022<p>The <tt>DeclarationName</tt> class represents the name of a
1023 declaration in Clang. Declarations in the C family of languages can
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]1024 take several different forms. Most declarations are named by
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1025 simple identifiers, e.g., "<code>f</code>" and "<code>x</code>" in
1026 the function declaration <code>f(int x)</code>. In C++, declaration
1027 names can also name class constructors ("<code>Class</code>"
1028 in <code>struct Class { Class(); }</code>), class destructors
1029 ("<code>~Class</code>"), overloaded operator names ("operator+"),
1030 and conversion functions ("<code>operator void const *</code>"). In
1031 Objective-C, declaration names can refer to the names of Objective-C
1032 methods, which involve the method name and the parameters,
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]1033 collectively called a <i>selector</i>, e.g.,
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1034 "<code>setWidth:height:</code>". Since all of these kinds of
Chris Lattner3fcbb892008-11-23 08:32:53 +0000[diff] [blame]1035 entities - variables, functions, Objective-C methods, C++
1036 constructors, destructors, and operators - are represented as
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1037 subclasses of Clang's common <code>NamedDecl</code>
1038 class, <code>DeclarationName</code> is designed to efficiently
1039 represent any kind of name.</p>
1040
1041<p>Given
1042 a <code>DeclarationName</code> <code>N</code>, <code>N.getNameKind()</code>
Douglas Gregor2def4832008-11-17 20:34:05 +0000[diff] [blame]1043 will produce a value that describes what kind of name <code>N</code>
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1044 stores. There are 8 options (all of the names are inside
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1045 the <code>DeclarationName</code> class)</p>
1046<dl>
1047 <dt>Identifier</dt>
1048 <dd>The name is a simple
1049 identifier. Use <code>N.getAsIdentifierInfo()</code> to retrieve the
1050 corresponding <code>IdentifierInfo*</code> pointing to the actual
1051 identifier. Note that C++ overloaded operators (e.g.,
1052 "<code>operator+</code>") are represented as special kinds of
1053 identifiers. Use <code>IdentifierInfo</code>'s <code>getOverloadedOperatorID</code>
1054 function to determine whether an identifier is an overloaded
1055 operator name.</dd>
1056
1057 <dt>ObjCZeroArgSelector, ObjCOneArgSelector,
1058 ObjCMultiArgSelector</dt>
1059 <dd>The name is an Objective-C selector, which can be retrieved as a
1060 <code>Selector</code> instance
1061 via <code>N.getObjCSelector()</code>. The three possible name
1062 kinds for Objective-C reflect an optimization within
1063 the <code>DeclarationName</code> class: both zero- and
1064 one-argument selectors are stored as a
1065 masked <code>IdentifierInfo</code> pointer, and therefore require
1066 very little space, since zero- and one-argument selectors are far
1067 more common than multi-argument selectors (which use a different
1068 structure).</dd>
1069
1070 <dt>CXXConstructorName</dt>
1071 <dd>The name is a C++ constructor
1072 name. Use <code>N.getCXXNameType()</code> to retrieve
1073 the <a href="#QualType">type</a> that this constructor is meant to
1074 construct. The type is always the canonical type, since all
1075 constructors for a given type have the same name.</dd>
1076
1077 <dt>CXXDestructorName</dt>
1078 <dd>The name is a C++ destructor
1079 name. Use <code>N.getCXXNameType()</code> to retrieve
1080 the <a href="#QualType">type</a> whose destructor is being
1081 named. This type is always a canonical type.</dd>
1082
1083 <dt>CXXConversionFunctionName</dt>
1084 <dd>The name is a C++ conversion function. Conversion functions are
1085 named according to the type they convert to, e.g., "<code>operator void
1086 const *</code>". Use <code>N.getCXXNameType()</code> to retrieve
1087 the type that this conversion function converts to. This type is
1088 always a canonical type.</dd>
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1089
1090 <dt>CXXOperatorName</dt>
1091 <dd>The name is a C++ overloaded operator name. Overloaded operators
1092 are named according to their spelling, e.g.,
1093 "<code>operator+</code>" or "<code>operator new
1094 []</code>". Use <code>N.getCXXOverloadedOperator()</code> to
1095 retrieve the overloaded operator (a value of
1096 type <code>OverloadedOperatorKind</code>).</dd>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1097</dl>
1098
1099<p><code>DeclarationName</code>s are cheap to create, copy, and
1100 compare. They require only a single pointer's worth of storage in
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1101 the common cases (identifiers, zero-
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1102 and one-argument Objective-C selectors) and use dense, uniqued
1103 storage for the other kinds of
1104 names. Two <code>DeclarationName</code>s can be compared for
1105 equality (<code>==</code>, <code>!=</code>) using a simple bitwise
1106 comparison, can be ordered
1107 with <code>&lt;</code>, <code>&gt;</code>, <code>&lt;=</code>,
1108 and <code>&gt;=</code> (which provide a lexicographical ordering for
1109 normal identifiers but an unspecified ordering for other kinds of
1110 names), and can be placed into LLVM <code>DenseMap</code>s
1111 and <code>DenseSet</code>s.</p>
1112
1113<p><code>DeclarationName</code> instances can be created in different
1114 ways depending on what kind of name the instance will store. Normal
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1115 identifiers (<code>IdentifierInfo</code> pointers) and Objective-C selectors
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1116 (<code>Selector</code>) can be implicitly converted
1117 to <code>DeclarationName</code>s. Names for C++ constructors,
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1118 destructors, conversion functions, and overloaded operators can be retrieved from
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1119 the <code>DeclarationNameTable</code>, an instance of which is
1120 available as <code>ASTContext::DeclarationNames</code>. The member
1121 functions <code>getCXXConstructorName</code>, <code>getCXXDestructorName</code>,
Douglas Gregore94ca9e42008-11-18 14:39:36 +0000[diff] [blame]1122 <code>getCXXConversionFunctionName</code>, and <code>getCXXOperatorName</code>, respectively,
1123 return <code>DeclarationName</code> instances for the four kinds of
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]1124 C++ special function names.</p>
1125
1126<!-- ======================================================================= -->
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1127<h3 id="DeclContext">Declaration contexts</h3>
1128<!-- ======================================================================= -->
1129<p>Every declaration in a program exists within some <i>declaration
1130 context</i>, such as a translation unit, namespace, class, or
1131 function. Declaration contexts in Clang are represented by
1132 the <code>DeclContext</code> class, from which the various
1133 declaration-context AST nodes
1134 (<code>TranslationUnitDecl</code>, <code>NamespaceDecl</code>, <code>RecordDecl</code>, <code>FunctionDecl</code>,
1135 etc.) will derive. The <code>DeclContext</code> class provides
1136 several facilities common to each declaration context:</p>
1137<dl>
1138 <dt>Source-centric vs. Semantics-centric View of Declarations</dt>
1139 <dd><code>DeclContext</code> provides two views of the declarations
1140 stored within a declaration context. The source-centric view
1141 accurately represents the program source code as written, including
1142 multiple declarations of entities where present (see the
1143 section <a href="#Redeclarations">Redeclarations and
1144 Overloads</a>), while the semantics-centric view represents the
1145 program semantics. The two views are kept synchronized by semantic
1146 analysis while the ASTs are being constructed.</dd>
1147
1148 <dt>Storage of declarations within that context</dt>
1149 <dd>Every declaration context can contain some number of
1150 declarations. For example, a C++ class (represented
1151 by <code>RecordDecl</code>) contains various member functions,
1152 fields, nested types, and so on. All of these declarations will be
1153 stored within the <code>DeclContext</code>, and one can iterate
1154 over the declarations via
1155 [<code>DeclContext::decls_begin()</code>,
1156 <code>DeclContext::decls_end()</code>). This mechanism provides
1157 the source-centric view of declarations in the context.</dd>
1158
1159 <dt>Lookup of declarations within that context</dt>
1160 <dd>The <code>DeclContext</code> structure provides efficient name
1161 lookup for names within that declaration context. For example,
1162 if <code>N</code> is a namespace we can look for the
1163 name <code>N::f</code>
1164 using <code>DeclContext::lookup</code>. The lookup itself is
1165 based on a lazily-constructed array (for declaration contexts
1166 with a small number of declarations) or hash table (for
1167 declaration contexts with more declarations). The lookup
1168 operation provides the semantics-centric view of the declarations
1169 in the context.</dd>
1170
1171 <dt>Ownership of declarations</dt>
1172 <dd>The <code>DeclContext</code> owns all of the declarations that
1173 were declared within its declaration context, and is responsible
1174 for the management of their memory as well as their
1175 (de-)serialization.</dd>
1176</dl>
1177
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1178<p>All declarations are stored within a declaration context, and one
1179 can query
1180 information about the context in which each declaration lives. One
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1181 can retrieve the <code>DeclContext</code> that contains a
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1182 particular <code>Decl</code>
1183 using <code>Decl::getDeclContext</code>. However, see the
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1184 section <a href="#LexicalAndSemanticContexts">Lexical and Semantic
1185 Contexts</a> for more information about how to interpret this
1186 context information.</p>
1187
1188<h4 id="Redeclarations">Redeclarations and Overloads</h4>
1189<p>Within a translation unit, it is common for an entity to be
1190declared several times. For example, we might declare a function "f"
1191 and then later re-declare it as part of an inlined definition:</p>
1192
1193<pre>
1194void f(int x, int y, int z = 1);
1195
1196inline void f(int x, int y, int z) { /* ... */ }
1197</pre>
1198
1199<p>The representation of "f" differs in the source-centric and
1200 semantics-centric views of a declaration context. In the
1201 source-centric view, all redeclarations will be present, in the
1202 order they occurred in the source code, making
1203 this view suitable for clients that wish to see the structure of
1204 the source code. In the semantics-centric view, only the most recent "f"
1205 will be found by the lookup, since it effectively replaces the first
1206 declaration of "f".</p>
1207
1208<p>In the semantics-centric view, overloading of functions is
1209 represented explicitly. For example, given two declarations of a
1210 function "g" that are overloaded, e.g.,</p>
1211<pre>
1212void g();
1213void g(int);
1214</pre>
1215<p>the <code>DeclContext::lookup</code> operation will return
Jonathan D. Turnerd3224292011-07-06 18:12:36 +0000[diff] [blame]1216 a <code>DeclContext::lookup_result</code> that contains a range of iterators
1217 over declarations of "g". Clients that perform semantic analysis on a
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1218 program that is not concerned with the actual source code will
1219 primarily use this semantics-centric view.</p>
1220
1221<h4 id="LexicalAndSemanticContexts">Lexical and Semantic Contexts</h4>
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1222<p>Each declaration has two potentially different
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1223 declaration contexts: a <i>lexical</i> context, which corresponds to
1224 the source-centric view of the declaration context, and
1225 a <i>semantic</i> context, which corresponds to the
1226 semantics-centric view. The lexical context is accessible
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1227 via <code>Decl::getLexicalDeclContext</code> while the
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1228 semantic context is accessible
Douglas Gregor4afa39d2009-01-20 01:17:11 +0000[diff] [blame]1229 via <code>Decl::getDeclContext</code>, both of which return
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1230 <code>DeclContext</code> pointers. For most declarations, the two
1231 contexts are identical. For example:</p>
1232
1233<pre>
1234class X {
1235public:
1236 void f(int x);
1237};
1238</pre>
1239
1240<p>Here, the semantic and lexical contexts of <code>X::f</code> are
1241 the <code>DeclContext</code> associated with the
1242 class <code>X</code> (itself stored as a <code>RecordDecl</code> AST
1243 node). However, we can now define <code>X::f</code> out-of-line:</p>
1244
1245<pre>
1246void X::f(int x = 17) { /* ... */ }
1247</pre>
1248
1249<p>This definition of has different lexical and semantic
1250 contexts. The lexical context corresponds to the declaration
1251 context in which the actual declaration occurred in the source
1252 code, e.g., the translation unit containing <code>X</code>. Thus,
1253 this declaration of <code>X::f</code> can be found by traversing
1254 the declarations provided by
1255 [<code>decls_begin()</code>, <code>decls_end()</code>) in the
1256 translation unit.</p>
1257
1258<p>The semantic context of <code>X::f</code> corresponds to the
1259 class <code>X</code>, since this member function is (semantically) a
1260 member of <code>X</code>. Lookup of the name <code>f</code> into
1261 the <code>DeclContext</code> associated with <code>X</code> will
1262 then return the definition of <code>X::f</code> (including
1263 information about the default argument).</p>
1264
1265<h4 id="TransparentContexts">Transparent Declaration Contexts</h4>
1266<p>In C and C++, there are several contexts in which names that are
1267 logically declared inside another declaration will actually "leak"
1268 out into the enclosing scope from the perspective of name
1269 lookup. The most obvious instance of this behavior is in
1270 enumeration types, e.g.,</p>
1271<pre>
1272enum Color {
1273 Red,
1274 Green,
1275 Blue
1276};
1277</pre>
1278
1279<p>Here, <code>Color</code> is an enumeration, which is a declaration
1280 context that contains the
1281 enumerators <code>Red</code>, <code>Green</code>,
1282 and <code>Blue</code>. Thus, traversing the list of declarations
1283 contained in the enumeration <code>Color</code> will
1284 yield <code>Red</code>, <code>Green</code>,
1285 and <code>Blue</code>. However, outside of the scope
1286 of <code>Color</code> one can name the enumerator <code>Red</code>
1287 without qualifying the name, e.g.,</p>
1288
1289<pre>
1290Color c = Red;
1291</pre>
1292
1293<p>There are other entities in C++ that provide similar behavior. For
1294 example, linkage specifications that use curly braces:</p>
1295
1296<pre>
1297extern "C" {
1298 void f(int);
1299 void g(int);
1300}
1301// f and g are visible here
1302</pre>
1303
1304<p>For source-level accuracy, we treat the linkage specification and
1305 enumeration type as a
1306 declaration context in which its enclosed declarations ("Red",
1307 "Green", and "Blue"; "f" and "g")
1308 are declared. However, these declarations are visible outside of the
1309 scope of the declaration context.</p>
1310
1311<p>These language features (and several others, described below) have
1312 roughly the same set of
1313 requirements: declarations are declared within a particular lexical
1314 context, but the declarations are also found via name lookup in
1315 scopes enclosing the declaration itself. This feature is implemented
1316 via <i>transparent</i> declaration contexts
1317 (see <code>DeclContext::isTransparentContext()</code>), whose
1318 declarations are visible in the nearest enclosing non-transparent
1319 declaration context. This means that the lexical context of the
1320 declaration (e.g., an enumerator) will be the
1321 transparent <code>DeclContext</code> itself, as will the semantic
1322 context, but the declaration will be visible in every outer context
1323 up to and including the first non-transparent declaration context (since
1324 transparent declaration contexts can be nested).</p>
1325
1326<p>The transparent <code>DeclContexts</code> are:</p>
1327<ul>
1328 <li>Enumerations (but not C++0x "scoped enumerations"):
1329 <pre>
1330enum Color {
1331 Red,
1332 Green,
1333 Blue
1334};
1335// Red, Green, and Blue are in scope
1336 </pre></li>
1337 <li>C++ linkage specifications:
1338 <pre>
1339extern "C" {
1340 void f(int);
1341 void g(int);
1342}
1343// f and g are in scope
1344 </pre></li>
1345 <li>Anonymous unions and structs:
1346 <pre>
1347struct LookupTable {
1348 bool IsVector;
1349 union {
1350 std::vector&lt;Item&gt; *Vector;
1351 std::set&lt;Item&gt; *Set;
1352 };
1353};
1354
1355LookupTable LT;
1356LT.Vector = 0; // Okay: finds Vector inside the unnamed union
1357 </pre>
1358 </li>
1359 <li>C++0x inline namespaces:
1360<pre>
1361namespace mylib {
1362 inline namespace debug {
1363 class X;
1364 }
1365}
1366mylib::X *xp; // okay: mylib::X refers to mylib::debug::X
1367</pre>
1368</li>
1369</ul>
1370
1371
1372<h4 id="MultiDeclContext">Multiply-Defined Declaration Contexts</h4>
1373<p>C++ namespaces have the interesting--and, so far, unique--property that
1374the namespace can be defined multiple times, and the declarations
1375provided by each namespace definition are effectively merged (from
1376the semantic point of view). For example, the following two code
1377snippets are semantically indistinguishable:</p>
1378<pre>
1379// Snippet #1:
1380namespace N {
1381 void f();
1382}
1383namespace N {
1384 void f(int);
1385}
1386
1387// Snippet #2:
1388namespace N {
1389 void f();
1390 void f(int);
1391}
1392</pre>
1393
1394<p>In Clang's representation, the source-centric view of declaration
1395 contexts will actually have two separate <code>NamespaceDecl</code>
1396 nodes in Snippet #1, each of which is a declaration context that
1397 contains a single declaration of "f". However, the semantics-centric
1398 view provided by name lookup into the namespace <code>N</code> for
Jonathan D. Turnerd3224292011-07-06 18:12:36 +0000[diff] [blame]1399 "f" will return a <code>DeclContext::lookup_result</code> that contains
1400 a range of iterators over declarations of "f".</p>
Douglas Gregor074149e2009-01-05 19:45:36 +0000[diff] [blame]1401
1402<p><code>DeclContext</code> manages multiply-defined declaration
1403 contexts internally. The
1404 function <code>DeclContext::getPrimaryContext</code> retrieves the
1405 "primary" context for a given <code>DeclContext</code> instance,
1406 which is the <code>DeclContext</code> responsible for maintaining
1407 the lookup table used for the semantics-centric view. Given the
1408 primary context, one can follow the chain
1409 of <code>DeclContext</code> nodes that define additional
1410 declarations via <code>DeclContext::getNextContext</code>. Note that
1411 these functions are used internally within the lookup and insertion
1412 methods of the <code>DeclContext</code>, so the vast majority of
1413 clients can ignore them.</p>
1414
1415<!-- ======================================================================= -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1416<h3 id="CFG">The <tt>CFG</tt> class</h3>
1417<!-- ======================================================================= -->
1418
1419<p>The <tt>CFG</tt> class is designed to represent a source-level
1420control-flow graph for a single statement (<tt>Stmt*</tt>). Typically
1421instances of <tt>CFG</tt> are constructed for function bodies (usually
1422an instance of <tt>CompoundStmt</tt>), but can also be instantiated to
1423represent the control-flow of any class that subclasses <tt>Stmt</tt>,
1424which includes simple expressions. Control-flow graphs are especially
1425useful for performing
1426<a href="http://en.wikipedia.org/wiki/Data_flow_analysis#Sensitivities">flow-
1427or path-sensitive</a> program analyses on a given function.</p>
1428
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1429<!-- ============ -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1430<h4>Basic Blocks</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1431<!-- ============ -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1432
1433<p>Concretely, an instance of <tt>CFG</tt> is a collection of basic
1434blocks. Each basic block is an instance of <tt>CFGBlock</tt>, which
1435simply contains an ordered sequence of <tt>Stmt*</tt> (each referring
1436to statements in the AST). The ordering of statements within a block
1437indicates unconditional flow of control from one statement to the
1438next. <a href="#ConditionalControlFlow">Conditional control-flow</a>
1439is represented using edges between basic blocks. The statements
1440within a given <tt>CFGBlock</tt> can be traversed using
1441the <tt>CFGBlock::*iterator</tt> interface.</p>
1442
1443<p>
Ted Kremenek18e17e72007-10-18 22:50:52 +0000[diff] [blame]1444A <tt>CFG</tt> object owns the instances of <tt>CFGBlock</tt> within
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1445the control-flow graph it represents. Each <tt>CFGBlock</tt> within a
1446CFG is also uniquely numbered (accessible
1447via <tt>CFGBlock::getBlockID()</tt>). Currently the number is
1448based on the ordering the blocks were created, but no assumptions
1449should be made on how <tt>CFGBlock</tt>s are numbered other than their
1450numbers are unique and that they are numbered from 0..N-1 (where N is
1451the number of basic blocks in the CFG).</p>
1452
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1453<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1454<h4>Entry and Exit Blocks</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1455<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1456
1457Each instance of <tt>CFG</tt> contains two special blocks:
1458an <i>entry</i> block (accessible via <tt>CFG::getEntry()</tt>), which
1459has no incoming edges, and an <i>exit</i> block (accessible
1460via <tt>CFG::getExit()</tt>), which has no outgoing edges. Neither
1461block contains any statements, and they serve the role of providing a
1462clear entrance and exit for a body of code such as a function body.
1463The presence of these empty blocks greatly simplifies the
1464implementation of many analyses built on top of CFGs.
1465
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1466<!-- ===================================================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1467<h4 id ="ConditionalControlFlow">Conditional Control-Flow</h4>
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1468<!-- ===================================================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1469
1470<p>Conditional control-flow (such as those induced by if-statements
1471and loops) is represented as edges between <tt>CFGBlock</tt>s.
1472Because different C language constructs can induce control-flow,
1473each <tt>CFGBlock</tt> also records an extra <tt>Stmt*</tt> that
1474represents the <i>terminator</i> of the block. A terminator is simply
1475the statement that caused the control-flow, and is used to identify
1476the nature of the conditional control-flow between blocks. For
1477example, in the case of an if-statement, the terminator refers to
1478the <tt>IfStmt</tt> object in the AST that represented the given
1479branch.</p>
1480
1481<p>To illustrate, consider the following code example:</p>
1482
1483<code>
1484int foo(int x) {<br>
1485&nbsp;&nbsp;x = x + 1;<br>
1486<br>
1487&nbsp;&nbsp;if (x > 2) x++;<br>
1488&nbsp;&nbsp;else {<br>
1489&nbsp;&nbsp;&nbsp;&nbsp;x += 2;<br>
1490&nbsp;&nbsp;&nbsp;&nbsp;x *= 2;<br>
1491&nbsp;&nbsp;}<br>
1492<br>
1493&nbsp;&nbsp;return x;<br>
1494}
1495</code>
1496
1497<p>After invoking the parser+semantic analyzer on this code fragment,
1498the AST of the body of <tt>foo</tt> is referenced by a
1499single <tt>Stmt*</tt>. We can then construct an instance
1500of <tt>CFG</tt> representing the control-flow graph of this function
1501body by single call to a static class method:</p>
1502
1503<code>
1504&nbsp;&nbsp;Stmt* FooBody = ...<br>
1505&nbsp;&nbsp;CFG* FooCFG = <b>CFG::buildCFG</b>(FooBody);
1506</code>
1507
1508<p>It is the responsibility of the caller of <tt>CFG::buildCFG</tt>
1509to <tt>delete</tt> the returned <tt>CFG*</tt> when the CFG is no
1510longer needed.</p>
1511
1512<p>Along with providing an interface to iterate over
1513its <tt>CFGBlock</tt>s, the <tt>CFG</tt> class also provides methods
1514that are useful for debugging and visualizing CFGs. For example, the
1515method
1516<tt>CFG::dump()</tt> dumps a pretty-printed version of the CFG to
1517standard error. This is especially useful when one is using a
1518debugger such as gdb. For example, here is the output
1519of <tt>FooCFG->dump()</tt>:</p>
1520
1521<code>
1522&nbsp;[ B5 (ENTRY) ]<br>
1523&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (0):<br>
1524&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B4<br>
1525<br>
1526&nbsp;[ B4 ]<br>
1527&nbsp;&nbsp;&nbsp;&nbsp;1: x = x + 1<br>
1528&nbsp;&nbsp;&nbsp;&nbsp;2: (x > 2)<br>
1529&nbsp;&nbsp;&nbsp;&nbsp;<b>T: if [B4.2]</b><br>
1530&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B5<br>
1531&nbsp;&nbsp;&nbsp;&nbsp;Successors (2): B3 B2<br>
1532<br>
1533&nbsp;[ B3 ]<br>
1534&nbsp;&nbsp;&nbsp;&nbsp;1: x++<br>
1535&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B4<br>
1536&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B1<br>
1537<br>
1538&nbsp;[ B2 ]<br>
1539&nbsp;&nbsp;&nbsp;&nbsp;1: x += 2<br>
1540&nbsp;&nbsp;&nbsp;&nbsp;2: x *= 2<br>
1541&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B4<br>
1542&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B1<br>
1543<br>
1544&nbsp;[ B1 ]<br>
1545&nbsp;&nbsp;&nbsp;&nbsp;1: return x;<br>
1546&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (2): B2 B3<br>
1547&nbsp;&nbsp;&nbsp;&nbsp;Successors (1): B0<br>
1548<br>
1549&nbsp;[ B0 (EXIT) ]<br>
1550&nbsp;&nbsp;&nbsp;&nbsp;Predecessors (1): B1<br>
1551&nbsp;&nbsp;&nbsp;&nbsp;Successors (0):
1552</code>
1553
1554<p>For each block, the pretty-printed output displays for each block
1555the number of <i>predecessor</i> blocks (blocks that have outgoing
1556control-flow to the given block) and <i>successor</i> blocks (blocks
1557that have control-flow that have incoming control-flow from the given
1558block). We can also clearly see the special entry and exit blocks at
1559the beginning and end of the pretty-printed output. For the entry
1560block (block B5), the number of predecessor blocks is 0, while for the
1561exit block (block B0) the number of successor blocks is 0.</p>
1562
1563<p>The most interesting block here is B4, whose outgoing control-flow
1564represents the branching caused by the sole if-statement
1565in <tt>foo</tt>. Of particular interest is the second statement in
1566the block, <b><tt>(x > 2)</tt></b>, and the terminator, printed
1567as <b><tt>if [B4.2]</tt></b>. The second statement represents the
1568evaluation of the condition of the if-statement, which occurs before
1569the actual branching of control-flow. Within the <tt>CFGBlock</tt>
1570for B4, the <tt>Stmt*</tt> for the second statement refers to the
1571actual expression in the AST for <b><tt>(x > 2)</tt></b>. Thus
1572pointers to subclasses of <tt>Expr</tt> can appear in the list of
1573statements in a block, and not just subclasses of <tt>Stmt</tt> that
1574refer to proper C statements.</p>
1575
1576<p>The terminator of block B4 is a pointer to the <tt>IfStmt</tt>
1577object in the AST. The pretty-printer outputs <b><tt>if
1578[B4.2]</tt></b> because the condition expression of the if-statement
1579has an actual place in the basic block, and thus the terminator is
1580essentially
1581<i>referring</i> to the expression that is the second statement of
1582block B4 (i.e., B4.2). In this manner, conditions for control-flow
1583(which also includes conditions for loops and switch statements) are
1584hoisted into the actual basic block.</p>
1585
Chris Lattner62fd2782008-11-22 21:41:31 +0000[diff] [blame]1586<!-- ===================== -->
1587<!-- <h4>Implicit Control-Flow</h4> -->
1588<!-- ===================== -->
Ted Kremenek8bc05712007-10-10 23:01:43 +0000[diff] [blame]1589
1590<!--
1591<p>A key design principle of the <tt>CFG</tt> class was to not require
1592any transformations to the AST in order to represent control-flow.
1593Thus the <tt>CFG</tt> does not perform any "lowering" of the
1594statements in an AST: loops are not transformed into guarded gotos,
1595short-circuit operations are not converted to a set of if-statements,
1596and so on.</p>
1597-->
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]1598
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1599
1600<!-- ======================================================================= -->
1601<h3 id="Constants">Constant Folding in the Clang AST</h3>
1602<!-- ======================================================================= -->
1603
1604<p>There are several places where constants and constant folding matter a lot to
1605the Clang front-end. First, in general, we prefer the AST to retain the source
1606code as close to how the user wrote it as possible. This means that if they
1607wrote "5+4", we want to keep the addition and two constants in the AST, we don't
1608want to fold to "9". This means that constant folding in various ways turns
1609into a tree walk that needs to handle the various cases.</p>
1610
1611<p>However, there are places in both C and C++ that require constants to be
1612folded. For example, the C standard defines what an "integer constant
1613expression" (i-c-e) is with very precise and specific requirements. The
1614language then requires i-c-e's in a lot of places (for example, the size of a
1615bitfield, the value for a case statement, etc). For these, we have to be able
1616to constant fold the constants, to do semantic checks (e.g. verify bitfield size
1617is non-negative and that case statements aren't duplicated). We aim for Clang
1618to be very pedantic about this, diagnosing cases when the code does not use an
1619i-c-e where one is required, but accepting the code unless running with
1620<tt>-pedantic-errors</tt>.</p>
1621
1622<p>Things get a little bit more tricky when it comes to compatibility with
1623real-world source code. Specifically, GCC has historically accepted a huge
1624superset of expressions as i-c-e's, and a lot of real world code depends on this
1625unfortuate accident of history (including, e.g., the glibc system headers). GCC
1626accepts anything its "fold" optimizer is capable of reducing to an integer
1627constant, which means that the definition of what it accepts changes as its
1628optimizer does. One example is that GCC accepts things like "case X-X:" even
1629when X is a variable, because it can fold this to 0.</p>
1630
1631<p>Another issue are how constants interact with the extensions we support, such
1632as __builtin_constant_p, __builtin_inf, __extension__ and many others. C99
1633obviously does not specify the semantics of any of these extensions, and the
1634definition of i-c-e does not include them. However, these extensions are often
1635used in real code, and we have to have a way to reason about them.</p>
1636
1637<p>Finally, this is not just a problem for semantic analysis. The code
1638generator and other clients have to be able to fold constants (e.g. to
1639initialize global variables) and has to handle a superset of what C99 allows.
1640Further, these clients can benefit from extended information. For example, we
1641know that "foo()||1" always evaluates to true, but we can't replace the
1642expression with true because it has side effects.</p>
1643
1644<!-- ======================= -->
1645<h4>Implementation Approach</h4>
1646<!-- ======================= -->
1647
1648<p>After trying several different approaches, we've finally converged on a
1649design (Note, at the time of this writing, not all of this has been implemented,
1650consider this a design goal!). Our basic approach is to define a single
1651recursive method evaluation method (<tt>Expr::Evaluate</tt>), which is
1652implemented in <tt>AST/ExprConstant.cpp</tt>. Given an expression with 'scalar'
1653type (integer, fp, complex, or pointer) this method returns the following
1654information:</p>
1655
1656<ul>
1657<li>Whether the expression is an integer constant expression, a general
1658 constant that was folded but has no side effects, a general constant that
1659 was folded but that does have side effects, or an uncomputable/unfoldable
1660 value.
1661</li>
1662<li>If the expression was computable in any way, this method returns the APValue
1663 for the result of the expression.</li>
1664<li>If the expression is not evaluatable at all, this method returns
1665 information on one of the problems with the expression. This includes a
1666 SourceLocation for where the problem is, and a diagnostic ID that explains
1667 the problem. The diagnostic should be have ERROR type.</li>
1668<li>If the expression is not an integer constant expression, this method returns
1669 information on one of the problems with the expression. This includes a
1670 SourceLocation for where the problem is, and a diagnostic ID that explains
1671 the problem. The diagnostic should be have EXTENSION type.</li>
1672</ul>
1673
1674<p>This information gives various clients the flexibility that they want, and we
1675will eventually have some helper methods for various extensions. For example,
1676Sema should have a <tt>Sema::VerifyIntegerConstantExpression</tt> method, which
1677calls Evaluate on the expression. If the expression is not foldable, the error
1678is emitted, and it would return true. If the expression is not an i-c-e, the
1679EXTENSION diagnostic is emitted. Finally it would return false to indicate that
1680the AST is ok.</p>
1681
1682<p>Other clients can use the information in other ways, for example, codegen can
1683just use expressions that are foldable in any way.</p>
1684
1685<!-- ========== -->
1686<h4>Extensions</h4>
1687<!-- ========== -->
1688
Chris Lattner552de0a2008-11-23 08:16:56 +0000[diff] [blame]1689<p>This section describes how some of the various extensions Clang supports
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1690interacts with constant evaluation:</p>
1691
1692<ul>
1693<li><b><tt>__extension__</tt></b>: The expression form of this extension causes
1694 any evaluatable subexpression to be accepted as an integer constant
1695 expression.</li>
1696<li><b><tt>__builtin_constant_p</tt></b>: This returns true (as a integer
Chris Lattner28daa532008-12-12 06:55:44 +0000[diff] [blame]1697 constant expression) if the operand is any evaluatable constant. As a
1698 special case, if <tt>__builtin_constant_p</tt> is the (potentially
1699 parenthesized) condition of a conditional operator expression ("?:"), only
Chris Lattner42b83dd2008-12-12 18:00:51 +0000[diff] [blame]1700 the true side of the conditional operator is considered, and it is evaluated
1701 with full constant folding.</li>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1702<li><b><tt>__builtin_choose_expr</tt></b>: The condition is required to be an
1703 integer constant expression, but we accept any constant as an "extension of
1704 an extension". This only evaluates one operand depending on which way the
1705 condition evaluates.</li>
1706<li><b><tt>__builtin_classify_type</tt></b>: This always returns an integer
1707 constant expression.</li>
1708<li><b><tt>__builtin_inf,nan,..</tt></b>: These are treated just like a
1709 floating-point literal.</li>
1710<li><b><tt>__builtin_abs,copysign,..</tt></b>: These are constant folded as
1711 general constant expressions.</li>
Douglas Gregoreb661ed2010-09-11 18:08:34 +0000[diff] [blame]1712<li><b><tt>__builtin_strlen</tt></b> and <b><tt>strlen</tt></b>: These are constant folded as integer constant expressions if the argument is a string literal.</li>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1713</ul>
1714
1715
Jeffrey Yasskin28dadd62011-01-28 23:41:54 +0000[diff] [blame]1716<!-- ======================================================================= -->
1717<h2 id="Howtos">How to change Clang</h2>
1718<!-- ======================================================================= -->
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1719
Jeffrey Yasskin28dadd62011-01-28 23:41:54 +0000[diff] [blame]1720<!-- ======================================================================= -->
1721<h3 id="AddingAttributes">How to add an attribute</h3>
1722<!-- ======================================================================= -->
1723
1724<p>To add an attribute, you'll have to add it to the list of attributes, add it
1725to the parsing phase, and look for it in the AST scan.
1726<a href="http://llvm.org/viewvc/llvm-project?view=rev&revision=124217">r124217</a>
1727has a good example of adding a warning attribute.</p>
1728
1729<p>(Beware that this hasn't been reviewed/fixed by the people who designed the
1730attributes system yet.)</p>
1731
1732<h4><a
1733href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/Attr.td?view=markup">include/clang/Basic/Attr.td</a></h4>
1734
1735<p>Each attribute gets a <tt>def</tt> inheriting from <tt>Attr</tt> or one of
1736its subclasses. <tt>InheritableAttr</tt> means that the attribute also applies
1737to subsequent declarations of the same name.</p>
1738
1739<p><tt>Spellings</tt> lists the strings that can appear in
1740<tt>__attribute__((here))</tt> or <tt>[[here]]</tt>. All such strings
1741will be synonymous. If you want to allow the <tt>[[]]</tt> C++0x
1742syntax, you have to define a list of <tt>Namespaces</tt>, which will
1743let users write <tt>[[namespace:spelling]]</tt>. Using the empty
1744string for a namespace will allow users to write just the spelling
1745with no "<tt>:</tt>".</p>
1746
1747<p><tt>Subjects</tt> restricts what kinds of AST node to which this attribute
1748can appertain (roughly, attach).</p>
1749
1750<p><tt>Args</tt> names the arguments the attribute takes, in order. If
1751<tt>Args</tt> is <tt>[StringArgument&lt;"Arg1">, IntArgument&lt;"Arg2">]</tt>
1752then <tt>__attribute__((myattribute("Hello", 3)))</tt> will be a valid use.</p>
1753
1754<h4>Boilerplate</h4>
1755
1756<p>Add an element to the <tt>AttributeList::Kind</tt> enum in <a
1757href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Sema/AttributeList.h?view=markup">include/clang/Sema/AttributeList.h</a>
1758named <tt>AT_lower_with_underscores</tt>. That is, a CamelCased
1759<tt>AttributeName</tt> in <tt>Attr.td</tt> name should become
1760<tt>AT_attribute_name</tt>.</p>
1761
1762<p>Add a case to the <tt>StringSwitch</tt> in <tt>AttributeList::getKind()</tt>
1763in <a
1764href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/AttributeList.cpp?view=markup">lib/Sema/AttributeList.cpp</a>
1765for each spelling of your attribute. Less common attributes should come toward
1766the end of that list.</p>
1767
1768<p>Write a new <tt>HandleYourAttr()</tt> function in <a
1769href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaDeclAttr.cpp?view=markup">lib/Sema/SemaDeclAttr.cpp</a>,
1770and add a case to the switch in <tt>ProcessNonInheritableDeclAttr()</tt> or
1771<tt>ProcessInheritableDeclAttr()</tt> forwarding to it.</p>
1772
1773<p>If your attribute causes extra warnings to fire, define a <tt>DiagGroup</tt>
1774in <a
1775href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticGroups.td?view=markup">include/clang/Basic/DiagnosticGroups.td</a>
1776named after the attribute's <tt>Spelling</tt> with "_"s replaced by "-"s. If
1777you're only defining one diagnostic, you can skip <tt>DiagnosticGroups.td</tt>
1778and use <tt>InGroup&lt;DiagGroup&lt;"your-attribute">></tt> directly in <a
1779href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td?view=markup">DiagnosticSemaKinds.td</a></p>
1780
1781<h4>The meat of your attribute</h4>
1782
1783<p>Find an appropriate place in Clang to do whatever your attribute needs to do.
1784Check for the attribute's presence using <tt>Decl::getAttr&lt;YourAttr>()</tt>.</p>
1785
1786<p>Update the <a href="LanguageExtensions.html">Clang Language Extensions</a>
1787document to describe your new attribute.</p>
Chris Lattner7bad1992008-11-16 21:48:07 +0000[diff] [blame]1788
Douglas Gregor1f634c62011-09-30 21:32:37 +0000[diff] [blame]1789<!-- ======================================================================= -->
1790<h3 id="AddingExprStmt">How to add an expression or statement</h3>
1791<!-- ======================================================================= -->
1792
1793<p>Expressions and statements are one of the most fundamental constructs within a
1794compiler, because they interact with many different parts of the AST,
1795semantic analysis, and IR generation. Therefore, adding a new
1796expression or statement kind into Clang requires some care. The following list
1797details the various places in Clang where an expression or statement needs to be
1798introduced, along with patterns to follow to ensure that the new
1799expression or statement works well across all of the C languages. We
1800focus on expressions, but statements are similar.</p>
1801
1802<ol>
1803 <li>Introduce parsing actions into the parser. Recursive-descent
1804 parsing is mostly self-explanatory, but there are a few things that
1805 are worth keeping in mind:
1806 <ul>
1807 <li>Keep as much source location information as possible! You'll
1808 want it later to produce great diagnostics and support Clang's
1809 various features that map between source code and the AST.</li>
1810 <li>Write tests for all of the "bad" parsing cases, to make sure
1811 your recovery is good. If you have matched delimiters (e.g.,
1812 parentheses, square brackets, etc.), use
Douglas Gregor4a8dfb52011-10-12 16:37:45 +0000[diff] [blame^]1813 <tt>Parser::BalancedDelimiterTracker</tt> to give nice diagnostics when
Douglas Gregor1f634c62011-09-30 21:32:37 +0000[diff] [blame]1814 things go wrong.</li>
1815 </ul>
1816 </li>
1817
1818 <li>Introduce semantic analysis actions into <tt>Sema</tt>. Semantic
1819 analysis should always involve two functions: an <tt>ActOnXXX</tt>
1820 function that will be called directly from the parser, and a
1821 <tt>BuildXXX</tt> function that performs the actual semantic
1822 analysis and will (eventually!) build the AST node. It's fairly
1823 common for the <tt>ActOnCXX</tt> function to do very little (often
1824 just some minor translation from the parser's representation to
1825 <tt>Sema</tt>'s representation of the same thing), but the separation
1826 is still important: C++ template instantiation, for example,
1827 should always call the <tt>BuildXXX</tt> variant. Several notes on
1828 semantic analysis before we get into construction of the AST:
1829 <ul>
1830 <li>Your expression probably involves some types and some
1831 subexpressions. Make sure to fully check that those types, and the
1832 types of those subexpressions, meet your expectations. Add
1833 implicit conversions where necessary to make sure that all of the
1834 types line up exactly the way you want them. Write extensive tests
1835 to check that you're getting good diagnostics for mistakes and
1836 that you can use various forms of subexpressions with your
1837 expression.</li>
1838 <li>When type-checking a type or subexpression, make sure to first
1839 check whether the type is "dependent"
1840 (<tt>Type::isDependentType()</tt>) or whether a subexpression is
1841 type-dependent (<tt>Expr::isTypeDependent()</tt>). If any of these
1842 return true, then you're inside a template and you can't do much
1843 type-checking now. That's normal, and your AST node (when you get
1844 there) will have to deal with this case. At this point, you can
1845 write tests that use your expression within templates, but don't
1846 try to instantiate the templates.</li>
1847 <li>For each subexpression, be sure to call
1848 <tt>Sema::CheckPlaceholderExpr()</tt> to deal with "weird"
1849 expressions that don't behave well as subexpressions. Then,
1850 determine whether you need to perform
1851 lvalue-to-rvalue conversions
1852 (<tt>Sema::DefaultLvalueConversion</tt>e) or
1853 the usual unary conversions
1854 (<tt>Sema::UsualUnaryConversions</tt>), for places where the
1855 subexpression is producing a value you intend to use.</li>
1856 <li>Your <tt>BuildXXX</tt> function will probably just return
1857 <tt>ExprError()</tt> at this point, since you don't have an AST.
1858 That's perfectly fine, and shouldn't impact your testing.</li>
1859 </ul>
1860 </li>
1861
1862 <li>Introduce an AST node for your new expression. This starts with
1863 declaring the node in <tt>include/Basic/StmtNodes.td</tt> and
1864 creating a new class for your expression in the appropriate
1865 <tt>include/AST/Expr*.h</tt> header. It's best to look at the class
1866 for a similar expression to get ideas, and there are some specific
1867 things to watch for:
1868 <ul>
1869 <li>If you need to allocate memory, use the <tt>ASTContext</tt>
1870 allocator to allocate memory. Never use raw <tt>malloc</tt> or
1871 <tt>new</tt>, and never hold any resources in an AST node, because
1872 the destructor of an AST node is never called.</li>
1873
1874 <li>Make sure that <tt>getSourceRange()</tt> covers the exact
1875 source range of your expression. This is needed for diagnostics
1876 and for IDE support.</li>
1877
1878 <li>Make sure that <tt>children()</tt> visits all of the
1879 subexpressions. This is important for a number of features (e.g., IDE
1880 support, C++ variadic templates). If you have sub-types, you'll
1881 also need to visit those sub-types in the
1882 <tt>RecursiveASTVisitor</tt>.</li>
1883
1884 <li>Add printing support (<tt>StmtPrinter.cpp</tt>) and dumping
1885 support (<tt>StmtDumper.cpp</tt>) for your expression.</li>
1886
1887 <li>Add profiling support (<tt>StmtProfile.cpp</tt>) for your AST
1888 node, noting the distinguishing (non-source location)
1889 characteristics of an instance of your expression. Omitting this
1890 step will lead to hard-to-diagnose failures regarding matching of
1891 template declarations.</li>
1892 </ul>
1893 </li>
1894
1895 <li>Teach semantic analysis to build your AST node! At this point,
1896 you can wire up your <tt>Sema::BuildXXX</tt> function to actually
1897 create your AST. A few things to check at this point:
1898 <ul>
1899 <li>If your expression can construct a new C++ class or return a
1900 new Objective-C object, be sure to update and then call
1901 <tt>Sema::MaybeBindToTemporary</tt> for your just-created AST node
1902 to be sure that the object gets properly destructed. An easy way
1903 to test this is to return a C++ class with a private destructor:
1904 semantic analysis should flag an error here with the attempt to
1905 call the destructor.</li>
1906 <li>Inspect the generated AST by printing it using <tt>clang -cc1
1907 -ast-print</tt>, to make sure you're capturing all of the
1908 important information about how the AST was written.</li>
1909 <li>Inspect the generated AST under <tt>clang -cc1 -ast-dump</tt>
1910 to verify that all of the types in the generated AST line up the
1911 way you want them. Remember that clients of the AST should never
1912 have to "think" to understand what's going on. For example, all
1913 implicit conversions should show up explicitly in the AST.</li>
1914 <li>Write tests that use your expression as a subexpression of
1915 other, well-known expressions. Can you call a function using your
1916 expression as an argument? Can you use the ternary operator?</li>
1917 </ul>
1918 </li>
1919
1920 <li>Teach code generation to create IR to your AST node. This step
1921 is the first (and only) that requires knowledge of LLVM IR. There
1922 are several things to keep in mind:
1923 <ul>
1924 <li>Code generation is separated into scalar/aggregate/complex and
1925 lvalue/rvalue paths, depending on what kind of result your
1926 expression produces. On occasion, this requires some careful
1927 factoring of code to avoid duplication.</li>
1928
1929 <li><tt>CodeGenFunction</tt> contains functions
1930 <tt>ConvertType</tt> and <tt>ConvertTypeForMem</tt> that convert
1931 Clang's types (<tt>clang::Type*</tt> or <tt>clang::QualType</tt>)
1932 to LLVM types.
1933 Use the former for values, and the later for memory locations:
1934 test with the C++ "bool" type to check this. If you find
1935 that you are having to use LLVM bitcasts to make
1936 the subexpressions of your expression have the type that your
1937 expression expects, STOP! Go fix semantic analysis and the AST so
1938 that you don't need these bitcasts.</li>
1939
1940 <li>The <tt>CodeGenFunction</tt> class has a number of helper
1941 functions to make certain operations easy, such as generating code
1942 to produce an lvalue or an rvalue, or to initialize a memory
1943 location with a given value. Prefer to use these functions rather
1944 than directly writing loads and stores, because these functions
1945 take care of some of the tricky details for you (e.g., for
1946 exceptions).</li>
1947
1948 <li>If your expression requires some special behavior in the event
1949 of an exception, look at the <tt>push*Cleanup</tt> functions in
1950 <tt>CodeGenFunction</tt> to introduce a cleanup. You shouldn't
1951 have to deal with exception-handling directly.</li>
1952
1953 <li>Testing is extremely important in IR generation. Use <tt>clang
1954 -cc1 -emit-llvm</tt> and <a
1955 href="http://llvm.org/cmds/FileCheck.html">FileCheck</a> to verify
1956 that you're generating the right IR.</li>
1957 </ul>
1958 </li>
1959
1960 <li>Teach template instantiation how to cope with your AST
1961 node, which requires some fairly simple code:
1962 <ul>
1963 <li>Make sure that your expression's constructor properly
1964 computes the flags for type dependence (i.e., the type your
1965 expression produces can change from one instantiation to the
1966 next), value dependence (i.e., the constant value your expression
1967 produces can change from one instantiation to the next),
1968 instantiation dependence (i.e., a template parameter or occurs
1969 anywhere in your expression), and whether your expression contains
1970 a parameter pack (for variadic templates). Often, computing these
1971 flags just means combining the results from the various types and
1972 subexpressions.</li>
1973
1974 <li>Add <tt>TransformXXX</tt> and <tt>RebuildXXX</tt> functions to
1975 the
1976 <tt>TreeTransform</tt> class template in <tt>Sema</tt>.
1977 <tt>TransformXXX</tt> should (recursively) transform all of the
1978 subexpressions and types
1979 within your expression, using <tt>getDerived().TransformYYY</tt>.
1980 If all of the subexpressions and types transform without error, it
1981 will then call the <tt>RebuildXXX</tt> function, which will in
1982 turn call <tt>getSema().BuildXXX</tt> to perform semantic analysis
1983 and build your expression.</li>
1984
1985 <li>To test template instantiation, take those tests you wrote to
1986 make sure that you were type checking with type-dependent
1987 expressions and dependent types (from step #2) and instantiate
1988 those templates with various types, some of which type-check and
1989 some that don't, and test the error messages in each case.</li>
1990 </ul>
1991 </li>
1992
1993 <li>There are some "extras" that make other features work better.
1994 It's worth handling these extras to give your expression complete
1995 integration into Clang:
1996 <ul>
1997 <li>Add code completion support for your expression in
1998 <tt>SemaCodeComplete.cpp</tt>.</li>
1999
2000 <li>If your expression has types in it, or has any "interesting"
2001 features other than subexpressions, extend libclang's
2002 <tt>CursorVisitor</tt> to provide proper visitation for your
2003 expression, enabling various IDE features such as syntax
2004 highlighting, cross-referencing, and so on. The
2005 <tt>c-index-test</tt> helper program can be used to test these
2006 features.</li>
2007 </ul>
2008 </li>
2009</ol>
2010
Ted Kremenek17a295d2008-06-11 06:19:49 +0000[diff] [blame]2011</div>
2012</body>
Douglas Gregor2e1cd422008-11-17 14:58:09 +0000[diff] [blame]2013</html>