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

blob: cca9960d835b7c5f41ad099a29e5d4a8f638ffd7 [file] [log] [blame]

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