Blame - docs/InternalsManual.html - platform/external/clang

blob: 9e8a21e708ef92011b62b27e73353c786d8a089a [file] [log] [blame]

Ted Kremenek	17a295d	2008-06-11 06:19:49 +0000	[diff] [blame]	1	<html>
				2	<head>
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	3	<title>"Clang" CFE Internals Manual</title>
Ted Kremenek	17a295d	2008-06-11 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 Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	6	<style type="text/css">
				7	td {
				8	vertical-align: top;
				9	}
				10	</style>
Ted Kremenek	17a295d	2008-06-11 06:19:49 +0000	[diff] [blame]	11	</head>
				12	<body>
				13
				14	<!--#include virtual="../menu.html.incl"-->
				15
				16	<div id="content">
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	17
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	18	<h1>"Clang" CFE Internals Manual</h1>
Chris Lattner	86920d3	2007-07-31 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 Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	23	<li><a href="#libbasic">The Clang 'Basic' Library</a>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	24	<ul>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	25	<li><a href="#Diagnostics">The Diagnostics Subsystem</a></li>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	26	<li><a href="#SourceLocation">The SourceLocation and SourceManager
				27	classes</a></li>
				28	</ul>
				29	</li>
				30	<li><a href="#liblex">The Lexer and Preprocessor Library</a>
				31	<ul>
				32	<li><a href="#Token">The Token class</a></li>
				33	<li><a href="#Lexer">The Lexer class</a></li>
Chris Lattner	7928125	2008-03-09 02:27:26 +0000	[diff] [blame]	34	<li><a href="#TokenLexer">The TokenLexer class</a></li>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	35	<li><a href="#MultipleIncludeOpt">The MultipleIncludeOpt class</a></li>
				36	</ul>
				37	</li>
				38	<li><a href="#libparse">The Parser Library</a>
				39	<ul>
				40	</ul>
				41	</li>
				42	<li><a href="#libast">The AST Library</a>
				43	<ul>
				44	<li><a href="#Type">The Type class and its subclasses</a></li>
				45	<li><a href="#QualType">The QualType class</a></li>
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	46	<li><a href="#DeclarationName">Declaration names</a></li>
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	47	<li><a href="#CFG">The CFG class</a></li>
Chris Lattner	7bad199	2008-11-16 21:48:07 +0000	[diff] [blame]	48	<li><a href="#Constants">Constant Folding in the Clang AST</a></li>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	49	</ul>
				50	</li>
				51	</ul>
				52
				53
				54	<!-- ======================================================================= -->
				55	<h2 id="intro">Introduction</h2>
				56	<!-- ======================================================================= -->
				57
				58	<p>This document describes some of the more important APIs and internal design
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	59	decisions made in the Clang C front-end. The purpose of this document is to
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	60	both capture some of this high level information and also describe some of the
				61	design decisions behind it. This is meant for people interested in hacking on
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	62	Clang, not for end-users. The description below is categorized by
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	63	libraries, and does not describe any of the clients of the libraries.</p>
				64
				65	<!-- ======================================================================= -->
				66	<h2 id="libsystem">LLVM System and Support Libraries</h2>
				67	<!-- ======================================================================= -->
				68
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	69	<p>The LLVM libsystem library provides the basic Clang system abstraction layer,
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	70	which is used for file system access. The LLVM libsupport library provides many
				71	underlying libraries and <a
				72	href="http://llvm.org/docs/ProgrammersManual.html">data-structures</a>,
				73	including command line option
				74	processing and various containers.</p>
				75
				76	<!-- ======================================================================= -->
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	77	<h2 id="libbasic">The Clang 'Basic' Library</h2>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	78	<!-- ======================================================================= -->
				79
				80	<p>This library certainly needs a better name. The 'basic' library contains a
				81	number of low-level utilities for tracking and manipulating source buffers,
				82	locations within the source buffers, diagnostics, tokens, target abstraction,
				83	and information about the subset of the language being compiled for.</p>
				84
				85	<p>Part of this infrastructure is specific to C (such as the TargetInfo class),
				86	other parts could be reused for other non-C-based languages (SourceLocation,
				87	SourceManager, Diagnostics, FileManager). When and if there is future demand
				88	we can figure out if it makes sense to introduce a new library, move the general
				89	classes somewhere else, or introduce some other solution.</p>
				90
				91	<p>We describe the roles of these classes in order of their dependencies.</p>
				92
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	93
				94	<!-- ======================================================================= -->
				95	<h3 id="Diagnostics">The Diagnostics Subsystem</h3>
				96	<!-- ======================================================================= -->
				97
				98	<p>The Clang Diagnostics subsystem is an important part of how the compiler
				99	communicates with the human. Diagnostics are the warnings and errors produced
				100	when the code is incorrect or dubious. In Clang, each diagnostic produced has
				101	(at the minimum) a unique ID, a <a href="#SourceLocation">SourceLocation</a> to
				102	"put the caret", an English translation associated with it, and a severity (e.g.
				103	<tt>WARNING</tt> or <tt>ERROR</tt>). They can also optionally include a number
				104	of arguments to the dianostic (which fill in "%0"'s in the string) as well as a
				105	number of source ranges that related to the diagnostic.</p>
				106
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	107	<p>In this section, we'll be giving examples produced by the Clang command line
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	108	driver, but diagnostics can be <a href="#DiagnosticClient">rendered in many
				109	different ways</a> depending on how the DiagnosticClient interface is
				110	implemented. A representative example of a diagonstic is:</p>
				111
				112	<pre>
				113	t.c:38:15: error: invalid operands to binary expression ('int *' and '_Complex float')
				114	<font color="darkgreen">P = (P-42) + Gamma*4;</font>
				115	<font color="blue">~~~~~~ ^ ~~~~~~~</font>
				116	</pre>
				117
				118	<p>In this example, you can see the English translation, the severity (error),
				119	you can see the source location (the caret ("^") and file/line/column info),
				120	the source ranges "~~~~", arguments to the diagnostic ("int*" and "_Complex
				121	float"). You'll have to believe me that there is a unique ID backing the
				122	diagnostic :).</p>
				123
				124	<p>Getting all of this to happen has several steps and involves many moving
				125	pieces, this section describes them and talks about best practices when adding
				126	a new diagnostic.</p>
				127
				128	<!-- ============================ -->
				129	<h4>The DiagnosticKinds.def file</h4>
				130	<!-- ============================ -->
				131
				132	<p>Diagnostics are created by adding an entry to the <tt><a
				133	href="http://llvm.org/svn/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticKinds.def"
				134	>DiagnosticKinds.def</a></tt> file. This file encodes the unique ID of the
				135	diagnostic (as an enum, the first argument), the severity of the diagnostic
				136	(second argument) and the English translation + format string.</p>
				137
				138	<p>There is little sanity with the naming of the unique ID's right now. Some
				139	start with err_, warn_, ext_ to encode the severity into the name. Since the
				140	enum is referenced in the C++ code that produces the diagnostic, it is somewhat
				141	useful for it to be reasonably short.</p>
				142
				143	<p>The severity of the diagnostic comes from the set {<tt>NOTE</tt>,
				144	<tt>WARNING</tt>, <tt>EXTENSION</tt>, <tt>EXTWARN</tt>, <tt>ERROR</tt>}. The
				145	<tt>ERROR</tt> severity is used for diagnostics indicating the program is never
				146	acceptable under any circumstances. When an error is emitted, the AST for the
				147	input code may not be fully built. The <tt>EXTENSION</tt> and <tt>EXTWARN</tt>
				148	severities are used for extensions to the language that Clang accepts. This
				149	means that Clang fully understands and can represent them in the AST, but we
				150	produce diagnostics to tell the user their code is non-portable. The difference
				151	is that the former are ignored by default, and the later warn by default. The
				152	<tt>WARNING</tt> severity is used for constructs that are valid in the currently
				153	selected source language but that are dubious in some way. The <tt>NOTE</tt>
				154	level is used to staple more information onto a previous diagnostics.</p>
				155
				156	<p>These <em>severities</em> are mapped into a smaller set (the
				157	Diagnostic::Level enum, {<tt>Ignored</tt>, <tt>Note</tt>, <tt>Warning</tt>,
				158	<tt>Error</tt> }) of output <em>levels</em> by the diagnostics subsystem based
				159	on various configuration options. For example, if the user specifies
				160	<tt>-pedantic</tt>, <tt>EXTENSION</tt> maps to <tt>Warning</tt>, if they specify
				161	<tt>-pedantic-errors</tt>, it turns into <tt>Error</tt>. Clang also internally
				162	supports a fully fine grained mapping mechanism that allows you to map any
				163	diagnostic that doesn't have <tt>ERRROR</tt> severity to any output level that
				164	you want. This is used to implement options like <tt>-Wunused_macros</tt>,
				165	<tt>-Wundef</tt> etc.</p>
				166
				167	<!-- ================= -->
				168	<h4>The Format String</h4>
				169	<!-- ================= -->
				170
				171	<p>The format string for the diagnostic is very simple, but it has some power.
				172	It takes the form of a string in English with markers that indicate where and
				173	how arguments to the diagnostic are inserted and formatted. For example, here
				174	are some simple format strings:</p>
				175
				176	<pre>
				177	"binary integer literals are an extension"
				178	"format string contains '\\0' within the string body"
				179	"more '<b>%%</b>' conversions than data arguments"
Chris Lattner	545b368	2008-11-23 20:27:13 +0000	[diff] [blame]	180	"invalid operands to binary expression (<b>%0</b> and <b>%1</b>)"
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	181	"overloaded '<b>%0</b>' must be a <b>%select{unary\|binary\|unary or binary}2</b> operator"
				182	" (has <b>%1</b> parameter<b>%s1</b>)"
				183	</pre>
				184
				185	<p>These examples show some important points of format strings. You can use any
				186	plain ASCII character in the diagnostic string except "%" without a problem,
				187	but these are C strings, so you have to use and be aware of all the C escape
				188	sequences (as in the second example). If you want to produce a "%" in the
				189	output, use the "%%" escape sequence, like the third diagnostic. Finally,
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	190	Clang uses the "%...[digit]" sequences to specify where and how arguments to
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	191	the diagnostic are formatted.</p>
				192
				193	<p>Arguments to the diagnostic are numbered according to how they are specified
				194	by the C++ code that <a href="#producingdiag">produces them</a>, and are
				195	referenced by <tt>%0</tt> .. <tt>%9</tt>. If you have more than 10 arguments
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	196	to your diagnostic, you are doing something wrong :). Unlike printf, there
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	197	is no requirement that arguments to the diagnostic end up in the output in
				198	the same order as they are specified, you could have a format string with
				199	<tt>"%1 %0"</tt> that swaps them, for example. The text in between the
				200	percent and digit are formatting instructions. If there are no instructions,
				201	the argument is just turned into a string and substituted in.</p>
				202
				203	<p>Here are some "best practices" for writing the English format string:</p>
				204
				205	<ul>
				206	<li>Keep the string short. It should ideally fit in the 80 column limit of the
				207	<tt>DiagnosticKinds.def</tt> file. This avoids the diagnostic wrapping when
				208	printed, and forces you to think about the important point you are conveying
				209	with the diagnostic.</li>
				210	<li>Take advantage of location information. The user will be able to see the
				211	line and location of the caret, so you don't need to tell them that the
				212	problem is with the 4th argument to the function: just point to it.</li>
				213	<li>Do not capitalize the diagnostic string, and do not end it with a
				214	period.</li>
				215	<li>If you need to quote something in the diagnostic string, use single
				216	quotes.</li>
				217	</ul>
				218
				219	<p>Diagnostics should never take random English strings as arguments: you
				220	shouldn't use <tt>"you have a problem with %0"</tt> and pass in things like
				221	<tt>"your argument"</tt> or <tt>"your return value"</tt> as arguments. Doing
				222	this prevents <a href="translation">translating</a> the Clang diagnostics to
				223	other languages (because they'll get random English words in their otherwise
				224	localized diagnostic). The exceptions to this are C/C++ language keywords
				225	(e.g. auto, const, mutable, etc) and C/C++ operators (<tt>/=</tt>). Note
				226	that things like "pointer" and "reference" are not keywords. On the other
				227	hand, you <em>can</em> include anything that comes from the user's source code,
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	228	including variable names, types, labels, etc. The 'select' format can be
				229	used to achieve this sort of thing in a localizable way, see below.</p>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	230
				231	<!-- ==================================== -->
				232	<h4>Formatting a Diagnostic Argument</a></h4>
				233	<!-- ==================================== -->
				234
				235	<p>Arguments to diagnostics are fully typed internally, and come from a couple
				236	different classes: integers, types, names, and random strings. Depending on
				237	the class of the argument, it can be optionally formatted in different ways.
				238	This gives the DiagnosticClient information about what the argument means
				239	without requiring it to use a specific presentation (consider this MVC for
				240	Clang :).</p>
				241
				242	<p>Here are the different diagnostic argument formats currently supported by
				243	Clang:</p>
				244
				245	<table>
				246	<tr><td colspan="2"><b>"s" format</b></td></tr>
				247	<tr><td>Example:</td><td><tt>"requires %1 parameter%s1"</tt></td></tr>
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	248	<tr><td>Class:</td><td>Integers</td></tr>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	249	<tr><td>Description:</td><td>This is a simple formatter for integers that is
				250	useful when producing English diagnostics. When the integer is 1, it prints
				251	as nothing. When the integer is not 1, it prints as "s". This allows some
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	252	simple grammatical forms to be to be handled correctly, and eliminates the
				253	need to use gross things like <tt>"requires %1 parameter(s)"</tt>.</td></tr>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	254
				255	<tr><td colspan="2"><b>"select" format</b></td></tr>
				256	<tr><td>Example:</td><td><tt>"must be a %select{unary\|binary\|unary or binary}2
				257	operator"</tt></td></tr>
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	258	<tr><td>Class:</td><td>Integers</td></tr>
Chris Lattner	cc54334	2008-11-22 23:50:47 +0000	[diff] [blame]	259	<tr><td>Description:</td><td>This format specifier is used to merge multiple
				260	related diagnostics together into one common one, without requiring the
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	261	difference to be specified as an English string argument. Instead of
Chris Lattner	cc54334	2008-11-22 23:50:47 +0000	[diff] [blame]	262	specifying the string, the diagnostic gets an integer argument and the
				263	format string selects the numbered option. In this case, the "%2" value
				264	must be an integer in the range [0..2]. If it is 0, it prints 'unary', if
				265	it is 1 it prints 'binary' if it is 2, it prints 'unary or binary'. This
				266	allows other language translations to substitute reasonable words (or entire
				267	phrases) based on the semantics of the diagnostic instead of having to do
				268	things textually.</td></tr>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	269
				270	<tr><td colspan="2"><b>"plural" format</b></td></tr>
Sebastian Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	271	<tr><td>Example:</td><td><tt>"you have %1 %plural{1:mouse\|:mice}1 connected to
				272	your computer"</tt></td></tr>
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	273	<tr><td>Class:</td><td>Integers</td></tr>
Sebastian Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	274	<tr><td>Description:</td><td><p>This is a formatter for complex plural forms.
				275	It is designed to handle even the requirements of languages with very
				276	complex plural forms, as many Baltic languages have. The argument consists
				277	of a series of expression/form pairs, separated by ':', where the first form
				278	whose expression evaluates to true is the result of the modifier.</p>
				279	<p>An expression can be empty, in which case it is always true. See the
				280	example at the top. Otherwise, it is a series of one or more numeric
				281	conditions, separated by ','. If any condition matches, the expression
				282	matches. Each numeric condition can take one of three forms.</p>
				283	<ul>
				284	<li>number: A simple decimal number matches if the argument is the same
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	285	as the number. Example: <tt>"%plural{1:mouse\|:mice}4"</tt></li>
Sebastian Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	286	<li>range: A range in square brackets matches if the argument is within
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	287	the range. Then range is inclusive on both ends. Example:
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	288	<tt>"%plural{0:none\|1:one\|[2,5]:some\|:many}2"</tt></li>
				289	<li>modulo: A modulo operator is followed by a number, and
				290	equals sign and either a number or a range. The tests are the
				291	same as for plain
Sebastian Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	292	numbers and ranges, but the argument is taken modulo the number first.
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	293	Example: <tt>"%plural{%100=0:even hundred\|%100=[1,50]:lower half\|:everything
				294	else}1"</tt></li>
Sebastian Redl	6816856	2008-11-22 22:16:45 +0000	[diff] [blame]	295	</ul>
				296	<p>The parser is very unforgiving. A syntax error, even whitespace, will
				297	abort, as will a failure to match the argument against any
				298	expression.</p></td></tr>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	299
Chris Lattner	077bf5e	2008-11-24 03:33:13 +0000	[diff] [blame^]	300	<tr><td colspan="2"><b>"objcclass" format</b></td></tr>
				301	<tr><td>Example:</td><td><tt>"method %objcclass0 not found"</tt></td></tr>
				302	<tr><td>Class:</td><td>DeclarationName</td></tr>
				303	<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
				304	DeclarationName corresponds to an Objective-C class method selector. As
				305	such, it prints the selector with a leading '+'.</p></td></tr>
				306
				307	<tr><td colspan="2"><b>"objcinstance" format</b></td></tr>
				308	<tr><td>Example:</td><td><tt>"method %objcinstance0 not found"</tt></td></tr>
				309	<tr><td>Class:</td><td>DeclarationName</td></tr>
				310	<tr><td>Description:</td><td><p>This is a simple formatter that indicates the
				311	DeclarationName corresponds to an Objective-C instance method selector. As
				312	such, it prints the selector with a leading '-'.</p></td></tr>
				313
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	314	</table>
				315
Chris Lattner	cc54334	2008-11-22 23:50:47 +0000	[diff] [blame]	316	<p>It is really easy to add format specifiers to the Clang diagnostics system,
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	317	but they should be discussed before they are added. If you are creating a lot
				318	of repetitive diagnostics and/or have an idea for a useful formatter, please
				319	bring it up on the cfe-dev mailing list.</p>
Chris Lattner	cc54334	2008-11-22 23:50:47 +0000	[diff] [blame]	320
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	321	<!-- ===================================================== -->
				322	<h4><a name="#producingdiag">Producing the Diagnostic</a></h4>
				323	<!-- ===================================================== -->
				324
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	325	<p>Now that you've created the diagnostic in the DiagnosticKinds.def file, you
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	326	need to write the code that detects the condition in question and emits the
				327	new diagnostic. Various components of Clang (e.g. the preprocessor, Sema,
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	328	etc) provide a helper function named "Diag". It creates a diagnostic and
				329	accepts the arguments, ranges, and other information that goes along with
				330	it.</p>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	331
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	332	<p>For example, the binary expression error comes from code like this:</p>
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	333
				334	<pre>
				335	if (various things that are bad)
				336	Diag(Loc, diag::err_typecheck_invalid_operands)
				337	<< lex->getType() << rex->getType()
				338	<< lex->getSourceRange() << rex->getSourceRange();
				339	</pre>
				340
				341	<p>This shows that use of the Diag method: they take a location (a <a
				342	href="#SourceLocation">SourceLocation</a> object) and a diagnostic enum value
				343	(which matches the name from DiagnosticKinds.def). If the diagnostic takes
				344	arguments, they are specified with the << operator: the first argument
				345	becomes %0, the second becomes %1, etc. The diagnostic interface allows you to
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	346	specify arguments of many different types, including <tt>int</tt> and
				347	<tt>unsigned</tt> for integer arguments, <tt>const char*</tt> and
				348	<tt>std::string</tt> for string arguments, <tt>DeclarationName</tt> and
				349	<tt>const IdentifierInfo*</tt> for names, <tt>QualType</tt> for types, etc.
				350	SourceRanges are also specified with the << operator, but do not have a
				351	specific ordering requirement.</p>
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	352
				353	<p>As you can see, adding and producing a diagnostic is pretty straightforward.
				354	The hard part is deciding exactly what you need to say to help the user, picking
				355	a suitable wording, and providing the information needed to format it correctly.
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	356	The good news is that the call site that issues a diagnostic should be
				357	completely independent of how the diagnostic is formatted and in what language
				358	it is rendered.
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	359	</p>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	360
				361	<!-- ============================================================= -->
				362	<h4><a name="DiagnosticClient">The DiagnosticClient Interface</a></h4>
				363	<!-- ============================================================= -->
				364
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	365	<p>Once code generates a diagnostic with all of the arguments and the rest of
				366	the relevant information, Clang needs to know what to do with it. As previously
				367	mentioned, the diagnostic machinery goes through some filtering to map a
				368	severity onto a diagnostic level, then (assuming the diagnostic is not mapped to
				369	"<tt>Ignore</tt>") it invokes an object that implements the DiagnosticClient
				370	interface with the information.</p>
				371
				372	<p>It is possible to implement this interface in many different ways. For
				373	example, the normal Clang DiagnosticClient (named 'TextDiagnosticPrinter') turns
				374	the arguments into strings (according to the various formatting rules), prints
				375	out the file/line/column information and the string, then prints out the line of
				376	code, the source ranges, and the caret. However, this behavior isn't required.
				377	</p>
				378
				379	<p>Another implementation of the DiagnosticClient interface is the
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	380	'TextDiagnosticBuffer' class, which is used when Clang is in -verify mode.
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	381	Instead of formatting and printing out the diagnostics, this implementation just
				382	captures and remembers the diagnostics as they fly by. Then -verify compares
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	383	the list of produced diagnostics to the list of expected ones. If they disagree,
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	384	it prints out its own output.
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	385	</p>
				386
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	387	<p>There are many other possible implementations of this interface, and this is
				388	why we prefer diagnostics to pass down rich structured information in arguments.
				389	For example, an HTML output might want declaration names be linkified to where
				390	they come from in the source. Another example is that a GUI might let you click
				391	on typedefs to expand them. This application would want to pass significantly
				392	more information about types through to the GUI than a simple flat string. The
				393	interface allows this to happen.</p>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	394
				395	<!-- ====================================================== -->
				396	<h4><a name="translation">Adding Translations to Clang</a></h4>
				397	<!-- ====================================================== -->
				398
Chris Lattner	627b705	2008-11-23 00:28:33 +0000	[diff] [blame]	399	<p>Not possible yet! Diagnostic strings should be written in UTF-8, the client
Chris Lattner	fd408ea	2008-11-23 00:42:53 +0000	[diff] [blame]	400	can translate to the relevant code page if needed. Each translation completely
				401	replaces the format string for the diagnostic.</p>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	402
				403
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	404	<!-- ======================================================================= -->
				405	<h3 id="SourceLocation">The SourceLocation and SourceManager classes</h3>
				406	<!-- ======================================================================= -->
				407
				408	<p>Strangely enough, the SourceLocation class represents a location within the
				409	source code of the program. Important design points include:</p>
				410
				411	<ol>
				412	<li>sizeof(SourceLocation) must be extremely small, as these are embedded into
				413	many AST nodes and are passed around often. Currently it is 32 bits.</li>
				414	<li>SourceLocation must be a simple value object that can be efficiently
				415	copied.</li>
				416	<li>We should be able to represent a source location for any byte of any input
				417	file. This includes in the middle of tokens, in whitespace, in trigraphs,
				418	etc.</li>
				419	<li>A SourceLocation must encode the current #include stack that was active when
				420	the location was processed. For example, if the location corresponds to a
				421	token, it should contain the set of #includes active when the token was
				422	lexed. This allows us to print the #include stack for a diagnostic.</li>
				423	<li>SourceLocation must be able to describe macro expansions, capturing both
				424	the ultimate instantiation point and the source of the original character
				425	data.</li>
				426	</ol>
				427
				428	<p>In practice, the SourceLocation works together with the SourceManager class
				429	to encode two pieces of information about a location: it's physical location
				430	and it's virtual location. For most tokens, these will be the same. However,
				431	for a macro expansion (or tokens that came from a _Pragma directive) these will
				432	describe the location of the characters corresponding to the token and the
				433	location where the token was used (i.e. the macro instantiation point or the
				434	location of the _Pragma itself).</p>
				435
Chris Lattner	3fcbb89	2008-11-23 08:32:53 +0000	[diff] [blame]	436	<p>For efficiency, we only track one level of macro instantiations: if a token was
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	437	produced by multiple instantiations, we only track the source and ultimate
				438	destination. Though we could track the intermediate instantiation points, this
				439	would require extra bookkeeping and no known client would benefit substantially
				440	from this.</p>
				441
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	442	<p>The Clang front-end inherently depends on the location of a token being
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	443	tracked correctly. If it is ever incorrect, the front-end may get confused and
				444	die. The reason for this is that the notion of the 'spelling' of a Token in
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	445	Clang depends on being able to find the original input characters for the token.
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	446	This concept maps directly to the "physical" location for the token.</p>
				447
				448	<!-- ======================================================================= -->
				449	<h2 id="liblex">The Lexer and Preprocessor Library</h2>
				450	<!-- ======================================================================= -->
				451
				452	<p>The Lexer library contains several tightly-connected classes that are involved
				453	with the nasty process of lexing and preprocessing C source code. The main
				454	interface to this library for outside clients is the large <a
				455	href="#Preprocessor">Preprocessor</a> class.
				456	It contains the various pieces of state that are required to coherently read
				457	tokens out of a translation unit.</p>
				458
				459	<p>The core interface to the Preprocessor object (once it is set up) is the
				460	Preprocessor::Lex method, which returns the next <a href="#Token">Token</a> from
				461	the preprocessor stream. There are two types of token providers that the
				462	preprocessor is capable of reading from: a buffer lexer (provided by the <a
				463	href="#Lexer">Lexer</a> class) and a buffered token stream (provided by the <a
Chris Lattner	7928125	2008-03-09 02:27:26 +0000	[diff] [blame]	464	href="#TokenLexer">TokenLexer</a> class).
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	465
				466
				467	<!-- ======================================================================= -->
				468	<h3 id="Token">The Token class</h3>
				469	<!-- ======================================================================= -->
				470
				471	<p>The Token class is used to represent a single lexed token. Tokens are
				472	intended to be used by the lexer/preprocess and parser libraries, but are not
				473	intended to live beyond them (for example, they should not live in the ASTs).<p>
				474
				475	<p>Tokens most often live on the stack (or some other location that is efficient
				476	to access) as the parser is running, but occasionally do get buffered up. For
				477	example, macro definitions are stored as a series of tokens, and the C++
Chris Lattner	3fcbb89	2008-11-23 08:32:53 +0000	[diff] [blame]	478	front-end periodically needs to buffer tokens up for tentative parsing and
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	479	various pieces of look-ahead. As such, the size of a Token matter. On a 32-bit
				480	system, sizeof(Token) is currently 16 bytes.</p>
				481
				482	<p>Tokens contain the following information:</p>
				483
				484	<ul>
				485	<li><b>A SourceLocation</b> - This indicates the location of the start of the
				486	token.</li>
				487
				488	<li><b>A length</b> - This stores the length of the token as stored in the
				489	SourceBuffer. For tokens that include them, this length includes trigraphs and
				490	escaped newlines which are ignored by later phases of the compiler. By pointing
				491	into the original source buffer, it is always possible to get the original
				492	spelling of a token completely accurately.</li>
				493
				494	<li><b>IdentifierInfo</b> - If a token takes the form of an identifier, and if
				495	identifier lookup was enabled when the token was lexed (e.g. the lexer was not
				496	reading in 'raw' mode) this contains a pointer to the unique hash value for the
				497	identifier. Because the lookup happens before keyword identification, this
				498	field is set even for language keywords like 'for'.</li>
				499
				500	<li><b>TokenKind</b> - This indicates the kind of token as classified by the
				501	lexer. This includes things like <tt>tok::starequal</tt> (for the "*="
				502	operator), <tt>tok::ampamp</tt> for the "&&" token, and keyword values
				503	(e.g. <tt>tok::kw_for</tt>) for identifiers that correspond to keywords. Note
				504	that some tokens can be spelled multiple ways. For example, C++ supports
				505	"operator keywords", where things like "and" are treated exactly like the
				506	"&&" operator. In these cases, the kind value is set to
				507	<tt>tok::ampamp</tt>, which is good for the parser, which doesn't have to
				508	consider both forms. For something that cares about which form is used (e.g.
				509	the preprocessor 'stringize' operator) the spelling indicates the original
				510	form.</li>
				511
				512	<li><b>Flags</b> - There are currently four flags tracked by the
				513	lexer/preprocessor system on a per-token basis:
				514
				515	<ol>
				516	<li><b>StartOfLine</b> - This was the first token that occurred on its input
				517	source line.</li>
				518	<li><b>LeadingSpace</b> - There was a space character either immediately
				519	before the token or transitively before the token as it was expanded
				520	through a macro. The definition of this flag is very closely defined by
				521	the stringizing requirements of the preprocessor.</li>
				522	<li><b>DisableExpand</b> - This flag is used internally to the preprocessor to
				523	represent identifier tokens which have macro expansion disabled. This
				524	prevents them from being considered as candidates for macro expansion ever
				525	in the future.</li>
				526	<li><b>NeedsCleaning</b> - This flag is set if the original spelling for the
				527	token includes a trigraph or escaped newline. Since this is uncommon,
				528	many pieces of code can fast-path on tokens that did not need cleaning.
				529	</p>
				530	</ol>
				531	</li>
				532	</ul>
				533
				534	<p>One interesting (and somewhat unusual) aspect of tokens is that they don't
				535	contain any semantic information about the lexed value. For example, if the
				536	token was a pp-number token, we do not represent the value of the number that
				537	was lexed (this is left for later pieces of code to decide). Additionally, the
				538	lexer library has no notion of typedef names vs variable names: both are
				539	returned as identifiers, and the parser is left to decide whether a specific
				540	identifier is a typedef or a variable (tracking this requires scope information
				541	among other things).</p>
				542
				543	<!-- ======================================================================= -->
				544	<h3 id="Lexer">The Lexer class</h3>
				545	<!-- ======================================================================= -->
				546
				547	<p>The Lexer class provides the mechanics of lexing tokens out of a source
				548	buffer and deciding what they mean. The Lexer is complicated by the fact that
				549	it operates on raw buffers that have not had spelling eliminated (this is a
				550	necessity to get decent performance), but this is countered with careful coding
				551	as well as standard performance techniques (for example, the comment handling
				552	code is vectorized on X86 and PowerPC hosts).</p>
				553
				554	<p>The lexer has a couple of interesting modal features:</p>
				555
				556	<ul>
				557	<li>The lexer can operate in 'raw' mode. This mode has several features that
				558	make it possible to quickly lex the file (e.g. it stops identifier lookup,
				559	doesn't specially handle preprocessor tokens, handles EOF differently, etc).
				560	This mode is used for lexing within an "<tt>#if 0</tt>" block, for
				561	example.</li>
				562	<li>The lexer can capture and return comments as tokens. This is required to
				563	support the -C preprocessor mode, which passes comments through, and is
				564	used by the diagnostic checker to identifier expect-error annotations.</li>
				565	<li>The lexer can be in ParsingFilename mode, which happens when preprocessing
Chris Lattner	8438624	2007-09-16 19:25:23 +0000	[diff] [blame]	566	after reading a #include directive. This mode changes the parsing of '<'
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	567	to return an "angled string" instead of a bunch of tokens for each thing
				568	within the filename.</li>
				569	<li>When parsing a preprocessor directive (after "<tt>#</tt>") the
				570	ParsingPreprocessorDirective mode is entered. This changes the parser to
				571	return EOM at a newline.</li>
				572	<li>The Lexer uses a LangOptions object to know whether trigraphs are enabled,
				573	whether C++ or ObjC keywords are recognized, etc.</li>
				574	</ul>
				575
				576	<p>In addition to these modes, the lexer keeps track of a couple of other
				577	features that are local to a lexed buffer, which change as the buffer is
				578	lexed:</p>
				579
				580	<ul>
				581	<li>The Lexer uses BufferPtr to keep track of the current character being
				582	lexed.</li>
				583	<li>The Lexer uses IsAtStartOfLine to keep track of whether the next lexed token
				584	will start with its "start of line" bit set.</li>
				585	<li>The Lexer keeps track of the current #if directives that are active (which
				586	can be nested).</li>
				587	<li>The Lexer keeps track of an <a href="#MultipleIncludeOpt">
				588	MultipleIncludeOpt</a> object, which is used to
				589	detect whether the buffer uses the standard "<tt>#ifndef XX</tt> /
				590	<tt>#define XX</tt>" idiom to prevent multiple inclusion. If a buffer does,
				591	subsequent includes can be ignored if the XX macro is defined.</li>
				592	</ul>
				593
				594	<!-- ======================================================================= -->
Chris Lattner	7928125	2008-03-09 02:27:26 +0000	[diff] [blame]	595	<h3 id="TokenLexer">The TokenLexer class</h3>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	596	<!-- ======================================================================= -->
				597
Chris Lattner	7928125	2008-03-09 02:27:26 +0000	[diff] [blame]	598	<p>The TokenLexer class is a token provider that returns tokens from a list
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	599	of tokens that came from somewhere else. It typically used for two things: 1)
				600	returning tokens from a macro definition as it is being expanded 2) returning
				601	tokens from an arbitrary buffer of tokens. The later use is used by _Pragma and
				602	will most likely be used to handle unbounded look-ahead for the C++ parser.</p>
				603
				604	<!-- ======================================================================= -->
				605	<h3 id="MultipleIncludeOpt">The MultipleIncludeOpt class</h3>
				606	<!-- ======================================================================= -->
				607
				608	<p>The MultipleIncludeOpt class implements a really simple little state machine
				609	that is used to detect the standard "<tt>#ifndef XX</tt> / <tt>#define XX</tt>"
				610	idiom that people typically use to prevent multiple inclusion of headers. If a
				611	buffer uses this idiom and is subsequently #include'd, the preprocessor can
				612	simply check to see whether the guarding condition is defined or not. If so,
				613	the preprocessor can completely ignore the include of the header.</p>
				614
				615
				616
				617	<!-- ======================================================================= -->
				618	<h2 id="libparse">The Parser Library</h2>
				619	<!-- ======================================================================= -->
				620
				621	<!-- ======================================================================= -->
				622	<h2 id="libast">The AST Library</h2>
				623	<!-- ======================================================================= -->
				624
				625	<!-- ======================================================================= -->
				626	<h3 id="Type">The Type class and its subclasses</h3>
				627	<!-- ======================================================================= -->
				628
				629	<p>The Type class (and its subclasses) are an important part of the AST. Types
				630	are accessed through the ASTContext class, which implicitly creates and uniques
				631	them as they are needed. Types have a couple of non-obvious features: 1) they
				632	do not capture type qualifiers like const or volatile (See
				633	<a href="#QualType">QualType</a>), and 2) they implicitly capture typedef
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	634	information. Once created, types are immutable (unlike decls).</p>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	635
				636	<p>Typedefs in C make semantic analysis a bit more complex than it would
				637	be without them. The issue is that we want to capture typedef information
				638	and represent it in the AST perfectly, but the semantics of operations need to
				639	"see through" typedefs. For example, consider this code:</p>
				640
				641	<code>
				642	void func() {<br>
Bill Wendling	30d1775	2007-10-06 01:56:01 +0000	[diff] [blame]	643	typedef int foo;<br>
				644	foo X, *Y;<br>
				645	typedef foo* bar;<br>
				646	bar Z;<br>
				647	*X; <i>// error</i><br>
				648	**Y; <i>// error</i><br>
				649	**Z; <i>// error</i><br>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	650	}<br>
				651	</code>
				652
				653	<p>The code above is illegal, and thus we expect there to be diagnostics emitted
				654	on the annotated lines. In this example, we expect to get:</p>
				655
				656	<pre>
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	657	<b>test.c:6:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	658	*X; // error
				659	<font color="blue">^~</font>
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	660	<b>test.c:7:1: error: indirection requires pointer operand ('foo' invalid)</b>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	661	**Y; // error
				662	<font color="blue">^~~</font>
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	663	<b>test.c:8:1: error: indirection requires pointer operand ('foo' invalid)</b>
				664	**Z; // error
				665	<font color="blue">^~~</font>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	666	</pre>
				667
				668	<p>While this example is somewhat silly, it illustrates the point: we want to
				669	retain typedef information where possible, so that we can emit errors about
				670	"<tt>std::string</tt>" instead of "<tt>std::basic_string<char, std:...</tt>".
				671	Doing this requires properly keeping typedef information (for example, the type
				672	of "X" is "foo", not "int"), and requires properly propagating it through the
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	673	various operators (for example, the type of *Y is "foo", not "int"). In order
				674	to retain this information, the type of these expressions is an instance of the
				675	TypedefType class, which indicates that the type of these expressions is a
				676	typedef for foo.
				677	</p>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	678
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	679	<p>Representing types like this is great for diagnostics, because the
				680	user-specified type is always immediately available. There are two problems
				681	with this: first, various semantic checks need to make judgements about the
Chris Lattner	33fc68a	2007-07-31 18:54:50 +0000	[diff] [blame]	682	<em>actual structure</em> of a type, ignoring typdefs. Second, we need an
				683	efficient way to query whether two types are structurally identical to each
				684	other, ignoring typedefs. The solution to both of these problems is the idea of
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	685	canonical types.</p>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	686
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	687	<!-- =============== -->
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	688	<h4>Canonical Types</h4>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	689	<!-- =============== -->
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	690
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	691	<p>Every instance of the Type class contains a canonical type pointer. For
				692	simple types with no typedefs involved (e.g. "<tt>int</tt>", "<tt>int*</tt>",
				693	"<tt>int**</tt>"), the type just points to itself. For types that have a
				694	typedef somewhere in their structure (e.g. "<tt>foo</tt>", "<tt>foo*</tt>",
				695	"<tt>foo**</tt>", "<tt>bar</tt>"), the canonical type pointer points to their
				696	structurally equivalent type without any typedefs (e.g. "<tt>int</tt>",
				697	"<tt>int</tt>", "<tt>int</tt>", and "<tt>int</tt>" respectively).</p>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	698
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	699	<p>This design provides a constant time operation (dereferencing the canonical
				700	type pointer) that gives us access to the structure of types. For example,
				701	we can trivially tell that "bar" and "foo*" are the same type by dereferencing
				702	their canonical type pointers and doing a pointer comparison (they both point
				703	to the single "<tt>int*</tt>" type).</p>
				704
				705	<p>Canonical types and typedef types bring up some complexities that must be
				706	carefully managed. Specifically, the "isa/cast/dyncast" operators generally
				707	shouldn't be used in code that is inspecting the AST. For example, when type
				708	checking the indirection operator (unary '*' on a pointer), the type checker
				709	must verify that the operand has a pointer type. It would not be correct to
				710	check that with "<tt>isa<PointerType>(SubExpr->getType())</tt>",
				711	because this predicate would fail if the subexpression had a typedef type.</p>
				712
				713	<p>The solution to this problem are a set of helper methods on Type, used to
				714	check their properties. In this case, it would be correct to use
				715	"<tt>SubExpr->getType()->isPointerType()</tt>" to do the check. This
				716	predicate will return true if the <em>canonical type is a pointer</em>, which is
				717	true any time the type is structurally a pointer type. The only hard part here
				718	is remembering not to use the <tt>isa/cast/dyncast</tt> operations.</p>
				719
				720	<p>The second problem we face is how to get access to the pointer type once we
				721	know it exists. To continue the example, the result type of the indirection
				722	operator is the pointee type of the subexpression. In order to determine the
				723	type, we need to get the instance of PointerType that best captures the typedef
				724	information in the program. If the type of the expression is literally a
				725	PointerType, we can return that, otherwise we have to dig through the
				726	typedefs to find the pointer type. For example, if the subexpression had type
				727	"<tt>foo*</tt>", we could return that type as the result. If the subexpression
				728	had type "<tt>bar</tt>", we want to return "<tt>foo*</tt>" (note that we do
				729	<em>not</em> want "<tt>int*</tt>"). In order to provide all of this, Type has
Chris Lattner	11406c1	2007-07-31 16:50:51 +0000	[diff] [blame]	730	a getAsPointerType() method that checks whether the type is structurally a
Chris Lattner	8a2bc62	2007-07-31 06:37:39 +0000	[diff] [blame]	731	PointerType and, if so, returns the best one. If not, it returns a null
				732	pointer.</p>
				733
				734	<p>This structure is somewhat mystical, but after meditating on it, it will
				735	make sense to you :).</p>
Chris Lattner	86920d3	2007-07-31 05:42:17 +0000	[diff] [blame]	736
				737	<!-- ======================================================================= -->
				738	<h3 id="QualType">The QualType class</h3>
				739	<!-- ======================================================================= -->
				740
				741	<p>The QualType class is designed as a trivial value class that is small,
				742	passed by-value and is efficient to query. The idea of QualType is that it
				743	stores the type qualifiers (const, volatile, restrict) separately from the types
				744	themselves: QualType is conceptually a pair of "Type*" and bits for the type
				745	qualifiers.</p>
				746
				747	<p>By storing the type qualifiers as bits in the conceptual pair, it is
				748	extremely efficient to get the set of qualifiers on a QualType (just return the
				749	field of the pair), add a type qualifier (which is a trivial constant-time
				750	operation that sets a bit), and remove one or more type qualifiers (just return
				751	a QualType with the bitfield set to empty).</p>
				752
				753	<p>Further, because the bits are stored outside of the type itself, we do not
				754	need to create duplicates of types with different sets of qualifiers (i.e. there
				755	is only a single heap allocated "int" type: "const int" and "volatile const int"
				756	both point to the same heap allocated "int" type). This reduces the heap size
				757	used to represent bits and also means we do not have to consider qualifiers when
				758	uniquing types (<a href="#Type">Type</a> does not even contain qualifiers).</p>
				759
				760	<p>In practice, on hosts where it is safe, the 3 type qualifiers are stored in
				761	the low bit of the pointer to the Type object. This means that QualType is
				762	exactly the same size as a pointer, and this works fine on any system where
				763	malloc'd objects are at least 8 byte aligned.</p>
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	764
				765	<!-- ======================================================================= -->
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	766	<h3 id="DeclarationName">Declaration names</h3>
				767	<!-- ======================================================================= -->
				768
				769	<p>The <tt>DeclarationName</tt> class represents the name of a
				770	declaration in Clang. Declarations in the C family of languages can
Chris Lattner	3fcbb89	2008-11-23 08:32:53 +0000	[diff] [blame]	771	take several different forms. Most declarations are named by
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	772	simple identifiers, e.g., "<code>f</code>" and "<code>x</code>" in
				773	the function declaration <code>f(int x)</code>. In C++, declaration
				774	names can also name class constructors ("<code>Class</code>"
				775	in <code>struct Class { Class(); }</code>), class destructors
				776	("<code>~Class</code>"), overloaded operator names ("operator+"),
				777	and conversion functions ("<code>operator void const *</code>"). In
				778	Objective-C, declaration names can refer to the names of Objective-C
				779	methods, which involve the method name and the parameters,
Chris Lattner	3fcbb89	2008-11-23 08:32:53 +0000	[diff] [blame]	780	collectively called a <i>selector</i>, e.g.,
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	781	"<code>setWidth:height:</code>". Since all of these kinds of
Chris Lattner	3fcbb89	2008-11-23 08:32:53 +0000	[diff] [blame]	782	entities - variables, functions, Objective-C methods, C++
				783	constructors, destructors, and operators - are represented as
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	784	subclasses of Clang's common <code>NamedDecl</code>
				785	class, <code>DeclarationName</code> is designed to efficiently
				786	represent any kind of name.</p>
				787
				788	<p>Given
				789	a <code>DeclarationName</code> <code>N</code>, <code>N.getNameKind()</code>
Douglas Gregor	2def483	2008-11-17 20:34:05 +0000	[diff] [blame]	790	will produce a value that describes what kind of name <code>N</code>
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	791	stores. There are 8 options (all of the names are inside
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	792	the <code>DeclarationName</code> class)</p>
				793	<dl>
				794	<dt>Identifier</dt>
				795	<dd>The name is a simple
				796	identifier. Use <code>N.getAsIdentifierInfo()</code> to retrieve the
				797	corresponding <code>IdentifierInfo*</code> pointing to the actual
				798	identifier. Note that C++ overloaded operators (e.g.,
				799	"<code>operator+</code>") are represented as special kinds of
				800	identifiers. Use <code>IdentifierInfo</code>'s <code>getOverloadedOperatorID</code>
				801	function to determine whether an identifier is an overloaded
				802	operator name.</dd>
				803
				804	<dt>ObjCZeroArgSelector, ObjCOneArgSelector,
				805	ObjCMultiArgSelector</dt>
				806	<dd>The name is an Objective-C selector, which can be retrieved as a
				807	<code>Selector</code> instance
				808	via <code>N.getObjCSelector()</code>. The three possible name
				809	kinds for Objective-C reflect an optimization within
				810	the <code>DeclarationName</code> class: both zero- and
				811	one-argument selectors are stored as a
				812	masked <code>IdentifierInfo</code> pointer, and therefore require
				813	very little space, since zero- and one-argument selectors are far
				814	more common than multi-argument selectors (which use a different
				815	structure).</dd>
				816
				817	<dt>CXXConstructorName</dt>
				818	<dd>The name is a C++ constructor
				819	name. Use <code>N.getCXXNameType()</code> to retrieve
				820	the <a href="#QualType">type</a> that this constructor is meant to
				821	construct. The type is always the canonical type, since all
				822	constructors for a given type have the same name.</dd>
				823
				824	<dt>CXXDestructorName</dt>
				825	<dd>The name is a C++ destructor
				826	name. Use <code>N.getCXXNameType()</code> to retrieve
				827	the <a href="#QualType">type</a> whose destructor is being
				828	named. This type is always a canonical type.</dd>
				829
				830	<dt>CXXConversionFunctionName</dt>
				831	<dd>The name is a C++ conversion function. Conversion functions are
				832	named according to the type they convert to, e.g., "<code>operator void
				833	const *</code>". Use <code>N.getCXXNameType()</code> to retrieve
				834	the type that this conversion function converts to. This type is
				835	always a canonical type.</dd>
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	836
				837	<dt>CXXOperatorName</dt>
				838	<dd>The name is a C++ overloaded operator name. Overloaded operators
				839	are named according to their spelling, e.g.,
				840	"<code>operator+</code>" or "<code>operator new
				841	[]</code>". Use <code>N.getCXXOverloadedOperator()</code> to
				842	retrieve the overloaded operator (a value of
				843	type <code>OverloadedOperatorKind</code>).</dd>
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	844	</dl>
				845
				846	<p><code>DeclarationName</code>s are cheap to create, copy, and
				847	compare. They require only a single pointer's worth of storage in
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	848	the common cases (identifiers, zero-
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	849	and one-argument Objective-C selectors) and use dense, uniqued
				850	storage for the other kinds of
				851	names. Two <code>DeclarationName</code>s can be compared for
				852	equality (<code>==</code>, <code>!=</code>) using a simple bitwise
				853	comparison, can be ordered
				854	with <code><</code>, <code>></code>, <code><=</code>,
				855	and <code>>=</code> (which provide a lexicographical ordering for
				856	normal identifiers but an unspecified ordering for other kinds of
				857	names), and can be placed into LLVM <code>DenseMap</code>s
				858	and <code>DenseSet</code>s.</p>
				859
				860	<p><code>DeclarationName</code> instances can be created in different
				861	ways depending on what kind of name the instance will store. Normal
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	862	identifiers (<code>IdentifierInfo</code> pointers) and Objective-C selectors
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	863	(<code>Selector</code>) can be implicitly converted
				864	to <code>DeclarationName</code>s. Names for C++ constructors,
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	865	destructors, conversion functions, and overloaded operators can be retrieved from
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	866	the <code>DeclarationNameTable</code>, an instance of which is
				867	available as <code>ASTContext::DeclarationNames</code>. The member
				868	functions <code>getCXXConstructorName</code>, <code>getCXXDestructorName</code>,
Douglas Gregor	e94ca9e4	2008-11-18 14:39:36 +0000	[diff] [blame]	869	<code>getCXXConversionFunctionName</code>, and <code>getCXXOperatorName</code>, respectively,
				870	return <code>DeclarationName</code> instances for the four kinds of
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	871	C++ special function names.</p>
				872
				873	<!-- ======================================================================= -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	874	<h3 id="CFG">The <tt>CFG</tt> class</h3>
				875	<!-- ======================================================================= -->
				876
				877	<p>The <tt>CFG</tt> class is designed to represent a source-level
				878	control-flow graph for a single statement (<tt>Stmt*</tt>). Typically
				879	instances of <tt>CFG</tt> are constructed for function bodies (usually
				880	an instance of <tt>CompoundStmt</tt>), but can also be instantiated to
				881	represent the control-flow of any class that subclasses <tt>Stmt</tt>,
				882	which includes simple expressions. Control-flow graphs are especially
				883	useful for performing
				884	<a href="http://en.wikipedia.org/wiki/Data_flow_analysis#Sensitivities">flow-
				885	or path-sensitive</a> program analyses on a given function.</p>
				886
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	887	<!-- ============ -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	888	<h4>Basic Blocks</h4>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	889	<!-- ============ -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	890
				891	<p>Concretely, an instance of <tt>CFG</tt> is a collection of basic
				892	blocks. Each basic block is an instance of <tt>CFGBlock</tt>, which
				893	simply contains an ordered sequence of <tt>Stmt*</tt> (each referring
				894	to statements in the AST). The ordering of statements within a block
				895	indicates unconditional flow of control from one statement to the
				896	next. <a href="#ConditionalControlFlow">Conditional control-flow</a>
				897	is represented using edges between basic blocks. The statements
				898	within a given <tt>CFGBlock</tt> can be traversed using
				899	the <tt>CFGBlock::*iterator</tt> interface.</p>
				900
				901	<p>
Ted Kremenek	18e17e7	2007-10-18 22:50:52 +0000	[diff] [blame]	902	A <tt>CFG</tt> object owns the instances of <tt>CFGBlock</tt> within
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	903	the control-flow graph it represents. Each <tt>CFGBlock</tt> within a
				904	CFG is also uniquely numbered (accessible
				905	via <tt>CFGBlock::getBlockID()</tt>). Currently the number is
				906	based on the ordering the blocks were created, but no assumptions
				907	should be made on how <tt>CFGBlock</tt>s are numbered other than their
				908	numbers are unique and that they are numbered from 0..N-1 (where N is
				909	the number of basic blocks in the CFG).</p>
				910
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	911	<!-- ===================== -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	912	<h4>Entry and Exit Blocks</h4>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	913	<!-- ===================== -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	914
				915	Each instance of <tt>CFG</tt> contains two special blocks:
				916	an <i>entry</i> block (accessible via <tt>CFG::getEntry()</tt>), which
				917	has no incoming edges, and an <i>exit</i> block (accessible
				918	via <tt>CFG::getExit()</tt>), which has no outgoing edges. Neither
				919	block contains any statements, and they serve the role of providing a
				920	clear entrance and exit for a body of code such as a function body.
				921	The presence of these empty blocks greatly simplifies the
				922	implementation of many analyses built on top of CFGs.
				923
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	924	<!-- ===================================================== -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	925	<h4 id ="ConditionalControlFlow">Conditional Control-Flow</h4>
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	926	<!-- ===================================================== -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	927
				928	<p>Conditional control-flow (such as those induced by if-statements
				929	and loops) is represented as edges between <tt>CFGBlock</tt>s.
				930	Because different C language constructs can induce control-flow,
				931	each <tt>CFGBlock</tt> also records an extra <tt>Stmt*</tt> that
				932	represents the <i>terminator</i> of the block. A terminator is simply
				933	the statement that caused the control-flow, and is used to identify
				934	the nature of the conditional control-flow between blocks. For
				935	example, in the case of an if-statement, the terminator refers to
				936	the <tt>IfStmt</tt> object in the AST that represented the given
				937	branch.</p>
				938
				939	<p>To illustrate, consider the following code example:</p>
				940
				941	<code>
				942	int foo(int x) {<br>
				943	x = x + 1;<br>
				944	<br>
				945	if (x > 2) x++;<br>
				946	else {<br>
				947	x += 2;<br>
				948	x *= 2;<br>
				949	}<br>
				950	<br>
				951	return x;<br>
				952	}
				953	</code>
				954
				955	<p>After invoking the parser+semantic analyzer on this code fragment,
				956	the AST of the body of <tt>foo</tt> is referenced by a
				957	single <tt>Stmt*</tt>. We can then construct an instance
				958	of <tt>CFG</tt> representing the control-flow graph of this function
				959	body by single call to a static class method:</p>
				960
				961	<code>
				962	Stmt* FooBody = ...<br>
				963	CFG* FooCFG = <b>CFG::buildCFG</b>(FooBody);
				964	</code>
				965
				966	<p>It is the responsibility of the caller of <tt>CFG::buildCFG</tt>
				967	to <tt>delete</tt> the returned <tt>CFG*</tt> when the CFG is no
				968	longer needed.</p>
				969
				970	<p>Along with providing an interface to iterate over
				971	its <tt>CFGBlock</tt>s, the <tt>CFG</tt> class also provides methods
				972	that are useful for debugging and visualizing CFGs. For example, the
				973	method
				974	<tt>CFG::dump()</tt> dumps a pretty-printed version of the CFG to
				975	standard error. This is especially useful when one is using a
				976	debugger such as gdb. For example, here is the output
				977	of <tt>FooCFG->dump()</tt>:</p>
				978
				979	<code>
				980	[ B5 (ENTRY) ]<br>
				981	Predecessors (0):<br>
				982	Successors (1): B4<br>
				983	<br>
				984	[ B4 ]<br>
				985	1: x = x + 1<br>
				986	2: (x > 2)<br>
				987	<b>T: if [B4.2]</b><br>
				988	Predecessors (1): B5<br>
				989	Successors (2): B3 B2<br>
				990	<br>
				991	[ B3 ]<br>
				992	1: x++<br>
				993	Predecessors (1): B4<br>
				994	Successors (1): B1<br>
				995	<br>
				996	[ B2 ]<br>
				997	1: x += 2<br>
				998	2: x *= 2<br>
				999	Predecessors (1): B4<br>
				1000	Successors (1): B1<br>
				1001	<br>
				1002	[ B1 ]<br>
				1003	1: return x;<br>
				1004	Predecessors (2): B2 B3<br>
				1005	Successors (1): B0<br>
				1006	<br>
				1007	[ B0 (EXIT) ]<br>
				1008	Predecessors (1): B1<br>
				1009	Successors (0):
				1010	</code>
				1011
				1012	<p>For each block, the pretty-printed output displays for each block
				1013	the number of <i>predecessor</i> blocks (blocks that have outgoing
				1014	control-flow to the given block) and <i>successor</i> blocks (blocks
				1015	that have control-flow that have incoming control-flow from the given
				1016	block). We can also clearly see the special entry and exit blocks at
				1017	the beginning and end of the pretty-printed output. For the entry
				1018	block (block B5), the number of predecessor blocks is 0, while for the
				1019	exit block (block B0) the number of successor blocks is 0.</p>
				1020
				1021	<p>The most interesting block here is B4, whose outgoing control-flow
				1022	represents the branching caused by the sole if-statement
				1023	in <tt>foo</tt>. Of particular interest is the second statement in
				1024	the block, <b><tt>(x > 2)</tt></b>, and the terminator, printed
				1025	as <b><tt>if [B4.2]</tt></b>. The second statement represents the
				1026	evaluation of the condition of the if-statement, which occurs before
				1027	the actual branching of control-flow. Within the <tt>CFGBlock</tt>
				1028	for B4, the <tt>Stmt*</tt> for the second statement refers to the
				1029	actual expression in the AST for <b><tt>(x > 2)</tt></b>. Thus
				1030	pointers to subclasses of <tt>Expr</tt> can appear in the list of
				1031	statements in a block, and not just subclasses of <tt>Stmt</tt> that
				1032	refer to proper C statements.</p>
				1033
				1034	<p>The terminator of block B4 is a pointer to the <tt>IfStmt</tt>
				1035	object in the AST. The pretty-printer outputs <b><tt>if
				1036	[B4.2]</tt></b> because the condition expression of the if-statement
				1037	has an actual place in the basic block, and thus the terminator is
				1038	essentially
				1039	<i>referring</i> to the expression that is the second statement of
				1040	block B4 (i.e., B4.2). In this manner, conditions for control-flow
				1041	(which also includes conditions for loops and switch statements) are
				1042	hoisted into the actual basic block.</p>
				1043
Chris Lattner	62fd278	2008-11-22 21:41:31 +0000	[diff] [blame]	1044	<!-- ===================== -->
				1045	<!-- <h4>Implicit Control-Flow</h4> -->
				1046	<!-- ===================== -->
Ted Kremenek	8bc0571	2007-10-10 23:01:43 +0000	[diff] [blame]	1047
				1048	<!--
				1049	<p>A key design principle of the <tt>CFG</tt> class was to not require
				1050	any transformations to the AST in order to represent control-flow.
				1051	Thus the <tt>CFG</tt> does not perform any "lowering" of the
				1052	statements in an AST: loops are not transformed into guarded gotos,
				1053	short-circuit operations are not converted to a set of if-statements,
				1054	and so on.</p>
				1055	-->
Ted Kremenek	17a295d	2008-06-11 06:19:49 +0000	[diff] [blame]	1056
Chris Lattner	7bad199	2008-11-16 21:48:07 +0000	[diff] [blame]	1057
				1058	<!-- ======================================================================= -->
				1059	<h3 id="Constants">Constant Folding in the Clang AST</h3>
				1060	<!-- ======================================================================= -->
				1061
				1062	<p>There are several places where constants and constant folding matter a lot to
				1063	the Clang front-end. First, in general, we prefer the AST to retain the source
				1064	code as close to how the user wrote it as possible. This means that if they
				1065	wrote "5+4", we want to keep the addition and two constants in the AST, we don't
				1066	want to fold to "9". This means that constant folding in various ways turns
				1067	into a tree walk that needs to handle the various cases.</p>
				1068
				1069	<p>However, there are places in both C and C++ that require constants to be
				1070	folded. For example, the C standard defines what an "integer constant
				1071	expression" (i-c-e) is with very precise and specific requirements. The
				1072	language then requires i-c-e's in a lot of places (for example, the size of a
				1073	bitfield, the value for a case statement, etc). For these, we have to be able
				1074	to constant fold the constants, to do semantic checks (e.g. verify bitfield size
				1075	is non-negative and that case statements aren't duplicated). We aim for Clang
				1076	to be very pedantic about this, diagnosing cases when the code does not use an
				1077	i-c-e where one is required, but accepting the code unless running with
				1078	<tt>-pedantic-errors</tt>.</p>
				1079
				1080	<p>Things get a little bit more tricky when it comes to compatibility with
				1081	real-world source code. Specifically, GCC has historically accepted a huge
				1082	superset of expressions as i-c-e's, and a lot of real world code depends on this
				1083	unfortuate accident of history (including, e.g., the glibc system headers). GCC
				1084	accepts anything its "fold" optimizer is capable of reducing to an integer
				1085	constant, which means that the definition of what it accepts changes as its
				1086	optimizer does. One example is that GCC accepts things like "case X-X:" even
				1087	when X is a variable, because it can fold this to 0.</p>
				1088
				1089	<p>Another issue are how constants interact with the extensions we support, such
				1090	as __builtin_constant_p, __builtin_inf, __extension__ and many others. C99
				1091	obviously does not specify the semantics of any of these extensions, and the
				1092	definition of i-c-e does not include them. However, these extensions are often
				1093	used in real code, and we have to have a way to reason about them.</p>
				1094
				1095	<p>Finally, this is not just a problem for semantic analysis. The code
				1096	generator and other clients have to be able to fold constants (e.g. to
				1097	initialize global variables) and has to handle a superset of what C99 allows.
				1098	Further, these clients can benefit from extended information. For example, we
				1099	know that "foo()\|\|1" always evaluates to true, but we can't replace the
				1100	expression with true because it has side effects.</p>
				1101
				1102	<!-- ======================= -->
				1103	<h4>Implementation Approach</h4>
				1104	<!-- ======================= -->
				1105
				1106	<p>After trying several different approaches, we've finally converged on a
				1107	design (Note, at the time of this writing, not all of this has been implemented,
				1108	consider this a design goal!). Our basic approach is to define a single
				1109	recursive method evaluation method (<tt>Expr::Evaluate</tt>), which is
				1110	implemented in <tt>AST/ExprConstant.cpp</tt>. Given an expression with 'scalar'
				1111	type (integer, fp, complex, or pointer) this method returns the following
				1112	information:</p>
				1113
				1114	<ul>
				1115	<li>Whether the expression is an integer constant expression, a general
				1116	constant that was folded but has no side effects, a general constant that
				1117	was folded but that does have side effects, or an uncomputable/unfoldable
				1118	value.
				1119	</li>
				1120	<li>If the expression was computable in any way, this method returns the APValue
				1121	for the result of the expression.</li>
				1122	<li>If the expression is not evaluatable at all, this method returns
				1123	information on one of the problems with the expression. This includes a
				1124	SourceLocation for where the problem is, and a diagnostic ID that explains
				1125	the problem. The diagnostic should be have ERROR type.</li>
				1126	<li>If the expression is not an integer constant expression, this method returns
				1127	information on one of the problems with the expression. This includes a
				1128	SourceLocation for where the problem is, and a diagnostic ID that explains
				1129	the problem. The diagnostic should be have EXTENSION type.</li>
				1130	</ul>
				1131
				1132	<p>This information gives various clients the flexibility that they want, and we
				1133	will eventually have some helper methods for various extensions. For example,
				1134	Sema should have a <tt>Sema::VerifyIntegerConstantExpression</tt> method, which
				1135	calls Evaluate on the expression. If the expression is not foldable, the error
				1136	is emitted, and it would return true. If the expression is not an i-c-e, the
				1137	EXTENSION diagnostic is emitted. Finally it would return false to indicate that
				1138	the AST is ok.</p>
				1139
				1140	<p>Other clients can use the information in other ways, for example, codegen can
				1141	just use expressions that are foldable in any way.</p>
				1142
				1143	<!-- ========== -->
				1144	<h4>Extensions</h4>
				1145	<!-- ========== -->
				1146
Chris Lattner	552de0a	2008-11-23 08:16:56 +0000	[diff] [blame]	1147	<p>This section describes how some of the various extensions Clang supports
Chris Lattner	7bad199	2008-11-16 21:48:07 +0000	[diff] [blame]	1148	interacts with constant evaluation:</p>
				1149
				1150	<ul>
				1151	<li><b><tt>__extension__</tt></b>: The expression form of this extension causes
				1152	any evaluatable subexpression to be accepted as an integer constant
				1153	expression.</li>
				1154	<li><b><tt>__builtin_constant_p</tt></b>: This returns true (as a integer
				1155	constant expression) if the operand is any evaluatable constant.</li>
				1156	<li><b><tt>__builtin_choose_expr</tt></b>: The condition is required to be an
				1157	integer constant expression, but we accept any constant as an "extension of
				1158	an extension". This only evaluates one operand depending on which way the
				1159	condition evaluates.</li>
				1160	<li><b><tt>__builtin_classify_type</tt></b>: This always returns an integer
				1161	constant expression.</li>
				1162	<li><b><tt>__builtin_inf,nan,..</tt></b>: These are treated just like a
				1163	floating-point literal.</li>
				1164	<li><b><tt>__builtin_abs,copysign,..</tt></b>: These are constant folded as
				1165	general constant expressions.</li>
				1166	</ul>
				1167
				1168
				1169
				1170
Ted Kremenek	17a295d	2008-06-11 06:19:49 +0000	[diff] [blame]	1171	</div>
				1172	</body>
Douglas Gregor	2e1cd42	2008-11-17 14:58:09 +0000	[diff] [blame]	1173	</html>