Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 522fa537033d94bf1ea500e9eb5ec743f702bd36 / . / clang / docs / LanguageExtensions.html
blob: 9f271900cb956c6a78f7fe4d0014d826a70fc883 [file] [log] [blame]
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<!-- Material used from: HTML 4.01 specs: http://www.w3.org/TR/html401/ -->
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]4<html>
5<head>
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]6 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Eli Friedman6b9c41e2011-09-19 23:17:44 +0000[diff] [blame]7 <title>Clang Language Extensions</title>
Alexis Huntefa07162011-06-23 01:22:53 +0000[diff] [blame]8 <link type="text/css" rel="stylesheet" href="../menu.css">
9 <link type="text/css" rel="stylesheet" href="../content.css">
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]10 <style type="text/css">
11 td {
12 vertical-align: top;
13 }
Benjamin Kramer83369b32012-01-15 16:42:14 +0000[diff] [blame]14 th { background-color: #ffddaa; }
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]15 </style>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]16</head>
17<body>
18
19<!--#include virtual="../menu.html.incl"-->
20
21<div id="content">
22
23<h1>Clang Language Extensions</h1>
24
25<ul>
26<li><a href="#intro">Introduction</a></li>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]27<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]28<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner7de64562009-04-13 02:45:46 +0000[diff] [blame]29<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]30<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]31<li><a href="#deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> attributes</a></li>
32<li><a href="#attributes-on-enumerators">Attributes on enumerators</a></li>
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]33<li><a href="#checking_language_features">Checks for Standard Language Features</a>
Ted Kremenek71d74632009-12-03 02:05:57 +0000[diff] [blame]34 <ul>
35 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
36 <li><a href="#cxx_rtti">C++ RTTI</a></li>
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]37 </ul></li>
38<li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]39 <ul>
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]40 <li><a href="#cxx0x">C++11</a>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]41 <ul>
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]42 <li><a href="#cxx_access_control_sfinae">C++11 SFINAE includes
43 access control</a></li>
44 <li><a href="#cxx_alias_templates">C++11 alias templates</a></li>
45 <li><a href="#cxx_alignas">C++11 alignment specifiers</a></li>
46 <li><a href="#cxx_attributes">C++11 attributes</a></li>
47 <li><a href="#cxx_constexpr">C++11 generalized constant expressions</a></li>
48 <li><a href="#cxx_decltype">C++11 <tt>decltype()</tt></a></li>
49 <li><a href="#cxx_default_function_template_args">C++11 default template arguments in function templates</a></li>
Douglas Gregor3c469762011-11-01 01:19:34 +0000[diff] [blame]50 <li><a href="#cxx_defaulted_functions">C++11 defaulted functions</a></li>
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]51 <li><a href="#cxx_delegating_constructor">C++11 delegating constructors</a></li>
52 <li><a href="#cxx_deleted_functions">C++11 deleted functions</a></li>
53 <li><a href="#cxx_explicit_conversions">C++11 explicit conversion functions</a></li>
54 <li><a href="#cxx_generalized_initializers">C++11 generalized initializers</a></li>
55 <li><a href="#cxx_implicit_moves">C++11 implicit move constructors/assignment operators</a></li>
56 <li><a href="#cxx_inheriting_constructors">C++11 inheriting constructors</a></li>
57 <li><a href="#cxx_inline_namespaces">C++11 inline namespaces</a></li>
58 <li><a href="#cxx_lambdas">C++11 lambdas</a></li>
59 <li><a href="#cxx_noexcept">C++11 noexcept specification</a></li>
60 <li><a href="#cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</a></li>
61 <li><a href="#cxx_nullptr">C++11 nullptr</a></li>
62 <li><a href="#cxx_override_control">C++11 override control</a></li>
63 <li><a href="#cxx_range_for">C++11 range-based for loop</a></li>
64 <li><a href="#cxx_raw_string_literals">C++11 raw string literals</a></li>
65 <li><a href="#cxx_rvalue_references">C++11 rvalue references</a></li>
66 <li><a href="#cxx_reference_qualified_functions">C++11 reference-qualified functions</a></li>
67 <li><a href="#cxx_static_assert">C++11 <tt>static_assert()</tt></a></li>
68 <li><a href="#cxx_auto_type">C++11 type inference</a></li>
69 <li><a href="#cxx_strong_enums">C++11 strongly-typed enumerations</a></li>
70 <li><a href="#cxx_trailing_return">C++11 trailing return type</a></li>
71 <li><a href="#cxx_unicode_literals">C++11 Unicode string literals</a></li>
72 <li><a href="#cxx_unrestricted_unions">C++11 unrestricted unions</a></li>
73 <li><a href="#cxx_user_literals">C++11 user-defined literals</a></li>
74 <li><a href="#cxx_variadic_templates">C++11 variadic templates</a></li>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]75 </ul></li>
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]76 <li><a href="#c11">C11</a>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]77 <ul>
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]78 <li><a href="#c_alignas">C11 alignment specifiers</a></li>
79 <li><a href="#c_generic_selections">C11 generic selections</a></li>
80 <li><a href="#c_static_assert">C11 <tt>_Static_assert()</tt></a></li>
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]81 </ul></li>
82 </ul> </li>
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]83<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]84<li><a href="#blocks">Blocks</a></li>
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]85<li><a href="#objc_features">Objective-C Features</a>
86 <ul>
87 <li><a href="#objc_instancetype">Related result types</a></li>
John McCall31168b02011-06-15 23:02:42 +0000[diff] [blame]88 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor8b7d4032011-09-08 17:18:35 +0000[diff] [blame]89 <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li>
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]90 </ul>
91</li>
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]92<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman6b9c41e2011-09-19 23:17:44 +0000[diff] [blame]93<li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]94<li><a href="#builtins">Builtin Functions</a>
95 <ul>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]96 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattnerbf206382009-09-21 03:09:59 +0000[diff] [blame]97 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner9cb59fa2011-04-09 03:57:26 +0000[diff] [blame]98 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]99 </ul>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]100</li>
Chris Lattnerc86ffc32009-04-09 19:58:15 +0000[diff] [blame]101<li><a href="#targetspecific">Target-Specific Extensions</a>
102 <ul>
103 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
104 </ul>
105</li>
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]106<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Benjamin Kramereaa262b2012-01-15 15:26:07 +0000[diff] [blame]107<li><a href="#dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</a>
Kostya Serebryany9751d342011-11-28 20:51:02 +0000[diff] [blame]108 <ul>
109 <li><a href="#address_sanitizer">AddressSanitizer</a></li>
110 </ul>
Benjamin Kramereaa262b2012-01-15 15:26:07 +0000[diff] [blame]111</li>
112<li><a href="#threadsafety">Thread Safety Annotation Checking</a>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]113 <ul>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]114 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
115 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
116 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]117 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
118 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]119 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
120 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
121 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
122 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
123 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
124 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
125 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
126 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
127 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
128 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
129 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
130 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
131 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]132 </ul>
Benjamin Kramereaa262b2012-01-15 15:26:07 +0000[diff] [blame]133</li>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]134</ul>
135
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]136<!-- ======================================================================= -->
137<h2 id="intro">Introduction</h2>
138<!-- ======================================================================= -->
139
140<p>This document describes the language extensions provided by Clang. In
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]141addition to the language extensions listed here, Clang aims to support a broad
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]142range of GCC extensions. Please see the <a
143href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
144more information on these extensions.</p>
145
146<!-- ======================================================================= -->
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]147<h2 id="feature_check">Feature Checking Macros</h2>
148<!-- ======================================================================= -->
149
150<p>Language extensions can be very useful, but only if you know you can depend
Chris Lattner04fe64f2011-03-21 16:25:11 +0000[diff] [blame]151on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]152function-like macros. This allows you to directly test for a feature in your
153code without having to resort to something like autoconf or fragile "compiler
154version checks".</p>
155
156<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]157<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]158<!-- ======================================================================= -->
159
160<p>This function-like macro takes a single identifier argument that is the name
161of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
162not. It can be used like this:</p>
163
164<blockquote>
165<pre>
166#ifndef __has_builtin // Optional of course.
167 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
168#endif
169
170...
171#if __has_builtin(__builtin_trap)
172 __builtin_trap();
173#else
174 abort();
175#endif
176...
177</pre>
178</blockquote>
179
180
181<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]182<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]183<!-- ======================================================================= -->
184
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]185<p>These function-like macros take a single identifier argument that is the
186name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
187is both supported by Clang and standardized in the current language standard
188or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
189<code>__has_extension</code> evaluates to 1 if the feature is supported by
190Clang in the current language (either as a language extension or a standard
191language feature) or 0 if not. They can be used like this:</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]192
193<blockquote>
194<pre>
195#ifndef __has_feature // Optional of course.
196 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
197#endif
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]198#ifndef __has_extension
199 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
200#endif
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]201
202...
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]203#if __has_feature(cxx_rvalue_references)
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]204// This code will only be compiled with the -std=c++11 and -std=gnu++11
205// options, because rvalue references are only standardized in C++11.
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]206#endif
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]207
208#if __has_extension(cxx_rvalue_references)
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]209// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]210// and -std=gnu++98 options, because rvalue references are supported as a
211// language extension in C++98.
212#endif
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]213</pre>
214</blockquote>
215
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]216<p id="has_feature_back_compat">For backwards compatibility reasons,
217<code>__has_feature</code> can also be used to test for support for
218non-standardized features, i.e. features not prefixed <code>c_</code>,
219<code>cxx_</code> or <code>objc_</code>.</p>
220
Kostya Serebryany9751d342011-11-28 20:51:02 +0000[diff] [blame]221<p id="has_feature_for_non_language_features">
222Another use of <code>__has_feature</code> is to check for compiler features
223not related to the language standard, such as e.g.
224<a href="AddressSanitizer.html">AddressSanitizer</a>.
225
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]226<p>If the <code>-pedantic-errors</code> option is given,
227<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
228
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]229<p>The feature tag is described along with the language feature below.</p>
230
Richard Smith2cca7b52012-02-25 10:41:10 +0000[diff] [blame]231<p>The feature name or extension name can also be specified with a preceding and
232following <code>__</code> (double underscore) to avoid interference from a macro
Richard Smithd18cd182012-03-01 02:12:07 +0000[diff] [blame]233with the same name. For instance, <code>__cxx_rvalue_references__</code> can be
234used instead of <code>cxx_rvalue_references</code>.</p>
Richard Smith2cca7b52012-02-25 10:41:10 +0000[diff] [blame]235
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]236<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]237<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlsson274a70e2010-10-20 02:31:43 +0000[diff] [blame]238<!-- ======================================================================= -->
239
240<p>This function-like macro takes a single identifier argument that is the name
241of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
242can be used like this:</p>
243
244<blockquote>
245<pre>
246#ifndef __has_attribute // Optional of course.
247 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
248#endif
249
250...
Anders Carlsson3bbfb7f2011-01-24 03:54:51 +0000[diff] [blame]251#if __has_attribute(always_inline)
252#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlsson274a70e2010-10-20 02:31:43 +0000[diff] [blame]253#else
Anders Carlsson3bbfb7f2011-01-24 03:54:51 +0000[diff] [blame]254#define ALWAYS_INLINE
Anders Carlsson274a70e2010-10-20 02:31:43 +0000[diff] [blame]255#endif
256...
257</pre>
258</blockquote>
259
Jean-Daniel Dupas908f1302012-03-01 14:53:16 +0000[diff] [blame]260<p>The attribute name can also be specified with a preceding and
261following <code>__</code> (double underscore) to avoid interference from a macro
262with the same name. For instance, <code>__always_inline__</code> can be used
263instead of <code>always_inline</code>.</p>
264
Anders Carlsson274a70e2010-10-20 02:31:43 +0000[diff] [blame]265<!-- ======================================================================= -->
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]266<h2 id="has_include">Include File Checking Macros</h2>
267<!-- ======================================================================= -->
268
269<p>Not all developments systems have the same include files.
270The <a href="#__has_include">__has_include</a> and
271<a href="#__has_include_next">__has_include_next</a> macros allow you to
272check for the existence of an include file before doing
273a possibly failing #include directive.</p>
274
275<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]276<h3><a name="__has_include">__has_include</a></h3>
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]277<!-- ======================================================================= -->
278
279<p>This function-like macro takes a single file name string argument that
280is the name of an include file. It evaluates to 1 if the file can
281be found using the include paths, or 0 otherwise:</p>
282
283<blockquote>
284<pre>
285// Note the two possible file name string formats.
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]286#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]287# include "myinclude.h"
288#endif
289
290// To avoid problem with non-clang compilers not having this macro.
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]291#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]292# include "myinclude.h"
293#endif
294</pre>
295</blockquote>
296
297<p>To test for this feature, use #if defined(__has_include).</p>
298
299<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]300<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]301<!-- ======================================================================= -->
302
303<p>This function-like macro takes a single file name string argument that
304is the name of an include file. It is like __has_include except that it
305looks for the second instance of the given file found in the include
306paths. It evaluates to 1 if the second instance of the file can
307be found using the include paths, or 0 otherwise:</p>
308
309<blockquote>
310<pre>
311// Note the two possible file name string formats.
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]312#if __has_include_next("myinclude.h") &amp;&amp; __has_include_next(&lt;stdint.h&gt;)
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]313# include_next "myinclude.h"
314#endif
315
316// To avoid problem with non-clang compilers not having this macro.
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]317#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompsonac0b0982009-11-02 22:28:12 +0000[diff] [blame]318# include_next "myinclude.h"
319#endif
320</pre>
321</blockquote>
322
323<p>Note that __has_include_next, like the GNU extension
324#include_next directive, is intended for use in headers only,
325and will issue a warning if used in the top-level compilation
326file. A warning will also be issued if an absolute path
327is used in the file argument.</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]328
Ted Kremeneka35d67d2011-10-12 19:46:30 +0000[diff] [blame]329
330<!-- ======================================================================= -->
331<h3><a name="__has_warning">__has_warning</a></h3>
332<!-- ======================================================================= -->
333
334<p>This function-like macro takes a string literal that represents a command
335 line option for a warning and returns true if that is a valid warning
336 option.</p>
337
338<blockquote>
339<pre>
340#if __has_warning("-Wformat")
341...
342#endif
343</pre>
344</blockquote>
345
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]346<!-- ======================================================================= -->
Chris Lattner7de64562009-04-13 02:45:46 +0000[diff] [blame]347<h2 id="builtinmacros">Builtin Macros</h2>
348<!-- ======================================================================= -->
349
Douglas Gregor6602c252010-04-30 02:51:06 +0000[diff] [blame]350<dl>
351 <dt><code>__BASE_FILE__</code></dt>
352 <dd>Defined to a string that contains the name of the main input
353 file passed to Clang.</dd>
354
355 <dt><code>__COUNTER__</code></dt>
356 <dd>Defined to an integer value that starts at zero and is
357 incremented each time the <code>__COUNTER__</code> macro is
358 expanded.</dd>
359
360 <dt><code>__INCLUDE_LEVEL__</code></dt>
361 <dd>Defined to an integral value that is the include depth of the
362 file currently being translated. For the main file, this value is
363 zero.</dd>
364
365 <dt><code>__TIMESTAMP__</code></dt>
366 <dd>Defined to the date and time of the last modification of the
367 current source file.</dd>
368
369 <dt><code>__clang__</code></dt>
370 <dd>Defined when compiling with Clang</dd>
371
372 <dt><code>__clang_major__</code></dt>
Chris Lattneree9c24a2011-12-15 19:06:36 +0000[diff] [blame]373 <dd>Defined to the major marketing version number of Clang (e.g., the
374 2 in 2.0.1). Note that marketing version numbers should not be used to
375 check for language features, as different vendors use different numbering
376 schemes. Instead, use the <a href="#feature_check">feature checking
377 macros</a>.</dd>
Douglas Gregor6602c252010-04-30 02:51:06 +0000[diff] [blame]378
379 <dt><code>__clang_minor__</code></dt>
380 <dd>Defined to the minor version number of Clang (e.g., the 0 in
Chris Lattneree9c24a2011-12-15 19:06:36 +0000[diff] [blame]381 2.0.1). Note that marketing version numbers should not be used to
382 check for language features, as different vendors use different numbering
383 schemes. Instead, use the <a href="#feature_check">feature checking
384 macros</a>.</dd>
Douglas Gregor6602c252010-04-30 02:51:06 +0000[diff] [blame]385
386 <dt><code>__clang_patchlevel__</code></dt>
Chris Lattneree9c24a2011-12-15 19:06:36 +0000[diff] [blame]387 <dd>Defined to the marketing patch level of Clang (e.g., the 1 in 2.0.1).</dd>
Douglas Gregor6602c252010-04-30 02:51:06 +0000[diff] [blame]388
389 <dt><code>__clang_version__</code></dt>
Chris Lattneree9c24a2011-12-15 19:06:36 +0000[diff] [blame]390 <dd>Defined to a string that captures the Clang marketing version, including
391 the Subversion tag or revision number, e.g., "1.5 (trunk 102332)".</dd>
Douglas Gregor6602c252010-04-30 02:51:06 +0000[diff] [blame]392</dl>
Chris Lattner7de64562009-04-13 02:45:46 +0000[diff] [blame]393
394<!-- ======================================================================= -->
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]395<h2 id="vectors">Vectors and Extended Vectors</h2>
396<!-- ======================================================================= -->
397
Anton Yartseve8985672012-01-15 16:22:24 +0000[diff] [blame]398<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
Owen Anderson029eb7d2010-01-27 01:22:36 +0000[diff] [blame]399
Benjamin Kramer83369b32012-01-15 16:42:14 +0000[diff] [blame]400<p>OpenCL vector types are created using <tt>ext_vector_type</tt> attribute. It
401support for <tt>V.xyzw</tt> syntax and other tidbits as seen in OpenCL. An
402example is:</p>
Owen Anderson029eb7d2010-01-27 01:22:36 +0000[diff] [blame]403
404<blockquote>
405<pre>
406typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
407typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
408
409float4 foo(float2 a, float2 b) {
410 float4 c;
411 c.xz = a;
412 c.yw = b;
413 return c;
414}
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]415</pre>
Owen Anderson029eb7d2010-01-27 01:22:36 +0000[diff] [blame]416</blockquote>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]417
Benjamin Kramer83369b32012-01-15 16:42:14 +0000[diff] [blame]418<p>Query for this feature with
419<tt>__has_extension(attribute_ext_vector_type)</tt>.</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]420
Benjamin Kramer83369b32012-01-15 16:42:14 +0000[diff] [blame]421<p>Giving <tt>-faltivec</tt> option to clang enables support for AltiVec vector
422syntax and functions. For example:</p>
Anton Yartseve8985672012-01-15 16:22:24 +0000[diff] [blame]423
424<blockquote>
425<pre>
426vector float foo(vector int a) {
427 vector int b;
428 b = vec_add(a, a) + a;
429 return (vector float)b;
430}
431</pre>
432</blockquote>
433
434<p>NEON vector types are created using <tt>neon_vector_type</tt> and
435<tt>neon_polyvector_type</tt> attributes. For example:</p>
436
437<blockquote>
438<pre>
439typedef <b>__attribute__((neon_vector_type(8)))</b> int8_t int8x8_t;
440typedef <b>__attribute__((neon_polyvector_type(16)))</b> poly8_t poly8x16_t;
441
442int8x8_t foo(int8x8_t a) {
443 int8x8_t v;
444 v = a;
445 return v;
446}
447</pre>
448</blockquote>
449
450<!-- ======================================================================= -->
451<h3><a name="vector_literals">Vector Literals</a></h3>
452<!-- ======================================================================= -->
453
454<p>Vector literals can be used to create vectors from a set of scalars, or
455vectors. Either parentheses or braces form can be used. In the parentheses form
456the number of literal values specified must be one, i.e. referring to a scalar
457value, or must match the size of the vector type being created. If a single
458scalar literal value is specified, the scalar literal value will be replicated
459to all the components of the vector type. In the brackets form any number of
460literals can be specified. For example:</p>
461
462<blockquote>
463<pre>
464typedef int v4si __attribute__((__vector_size__(16)));
465typedef float float4 __attribute__((ext_vector_type(4)));
466typedef float float2 __attribute__((ext_vector_type(2)));
467
468v4si vsi = (v4si){1, 2, 3, 4};
469float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
470vector int vi1 = (vector int)(1); // vi1 will be (1, 1, 1, 1).
471vector int vi2 = (vector int){1}; // vi2 will be (1, 0, 0, 0).
472vector int vi3 = (vector int)(1, 2); // error
473vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
474vector int vi5 = (vector int)(1, 2, 3, 4);
475float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));
476</pre>
477</blockquote>
478
479<!-- ======================================================================= -->
480<h3><a name="vector_operations">Vector Operations</a></h3>
481<!-- ======================================================================= -->
482
483<p>The table below shows the support for each operation by vector extension.
484A dash indicates that an operation is not accepted according to a corresponding
485specification.</p>
486
487<table width="500" border="1" cellspacing="0">
488 <tr>
Benjamin Kramer83369b32012-01-15 16:42:14 +0000[diff] [blame]489 <th>Operator</th>
490 <th>OpenCL</th>
491 <th>AltiVec</th>
492 <th>GCC</th>
493 <th>NEON</th>
Anton Yartseve8985672012-01-15 16:22:24 +0000[diff] [blame]494 </tr>
495 <tr>
496 <td>[]</td>
497 <td align="center">yes</td>
498 <td align="center">yes</td>
499 <td align="center">yes</td>
500 <td align="center">-</td>
501 </tr>
502 <tr>
503 <td>unary operators +, -</td>
504 <td align="center">yes</td>
505 <td align="center">yes</td>
506 <td align="center">yes</td>
507 <td align="center">-</td>
508 </tr>
509 <tr>
510 <td>++, --</td>
511 <td align="center">yes</td>
512 <td align="center">yes</td>
513 <td align="center">-</td>
514 <td align="center">-</td>
515 </tr>
516 <tr>
517 <td>+, -, *, /, %</td>
518 <td align="center">yes</td>
519 <td align="center">yes</td>
520 <td align="center">yes</td>
521 <td align="center">-</td>
522 </tr>
523 <tr>
524 <td>bitwise operators &, |, ^, ~</td>
525 <td align="center">yes</td>
526 <td align="center">yes</td>
527 <td align="center">yes</td>
528 <td align="center">-</td>
529 </tr>
530 <tr>
531 <td>&gt&gt, &lt&lt</td>
532 <td align="center">yes</td>
533 <td align="center">yes</td>
534 <td align="center">yes</td>
535 <td align="center">-</td>
536 </tr>
537 <tr>
538 <td>!, &&,||</td>
539 <td align="center">no</td>
540 <td align="center">-</td>
541 <td align="center">-</td>
542 <td align="center">-</td>
543 </tr>
544 <tr>
545 <td>==,!=, >, <, >=, <=</td>
546 <td align="center">yes</td>
547 <td align="center">yes</td>
548 <td align="center">-</td>
549 <td align="center">-</td>
550 </tr>
551 <tr>
552 <td>=</td>
553 <td align="center">yes</td>
554 <td align="center">yes</td>
555 <td align="center">yes</td>
556 <td align="center">yes</td>
557 </tr>
558 <tr>
559 <td>:?</td>
560 <td align="center">yes</td>
561 <td align="center">-</td>
562 <td align="center">-</td>
563 <td align="center">-</td>
564 </tr>
565 <tr>
566 <td>sizeof</td>
567 <td align="center">yes</td>
568 <td align="center">yes</td>
569 <td align="center">yes</td>
570 <td align="center">yes</td>
571 </tr>
572</table>
573
Owen Anderson029eb7d2010-01-27 01:22:36 +0000[diff] [blame]574<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
575
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]576<!-- ======================================================================= -->
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]577<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc74073c2010-10-06 23:12:32 +0000[diff] [blame]578<!-- ======================================================================= -->
579
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]580<p>An optional string message can be added to the <tt>deprecated</tt>
581and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc74073c2010-10-06 23:12:32 +0000[diff] [blame]582
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]583<blockquote>
Chris Lattnere56c7862010-11-09 19:43:35 +0000[diff] [blame]584<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]585</blockquote>
586
587<p>If the deprecated or unavailable declaration is used, the message
588will be incorporated into the appropriate diagnostic:</p>
589
590<blockquote>
Chris Lattnere56c7862010-11-09 19:43:35 +0000[diff] [blame]591<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!! [-Wdeprecated-declarations]
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]592 explode();
593 ^</pre>
594</blockquote>
595
596<p>Query for this feature
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]597with <tt>__has_extension(attribute_deprecated_with_message)</tt>
598and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCallf71b4532010-11-08 19:48:17 +0000[diff] [blame]599
600<!-- ======================================================================= -->
601<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
602<!-- ======================================================================= -->
603
604<p>Clang allows attributes to be written on individual enumerators.
605This allows enumerators to be deprecated, made unavailable, etc. The
606attribute must appear after the enumerator name and before any
607initializer, like so:</p>
608
609<blockquote>
610<pre>enum OperationMode {
611 OM_Invalid,
612 OM_Normal,
613 OM_Terrified __attribute__((deprecated)),
614 OM_AbortOnError __attribute__((deprecated)) = 4
615};</pre>
616</blockquote>
617
618<p>Attributes on the <tt>enum</tt> declaration do not apply to
619individual enumerators.</p>
620
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]621<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc74073c2010-10-06 23:12:32 +0000[diff] [blame]622
623<!-- ======================================================================= -->
Ted Kremenek51540f92009-12-03 02:04:01 +0000[diff] [blame]624<h2 id="checking_language_features">Checks for Standard Language Features</h2>
625<!-- ======================================================================= -->
626
627<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
628enabled. Those features are listed here.</p>
629
Ted Kremenek71d74632009-12-03 02:05:57 +0000[diff] [blame]630<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek51540f92009-12-03 02:04:01 +0000[diff] [blame]631
Ted Kremenek71d74632009-12-03 02:05:57 +0000[diff] [blame]632<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Alexis Huntc0a2ab62011-06-23 00:42:53 +0000[diff] [blame]633example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p>
Ted Kremenek51540f92009-12-03 02:04:01 +0000[diff] [blame]634
Ted Kremenek71d74632009-12-03 02:05:57 +0000[diff] [blame]635<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek51540f92009-12-03 02:04:01 +0000[diff] [blame]636
Ted Kremenek72de23e2009-12-03 02:06:43 +0000[diff] [blame]637<p>Use <tt>__has_feature(cxx_rtti)</tt> to determine if C++ RTTI has been enabled. For example,
Ted Kremenek71d74632009-12-03 02:05:57 +0000[diff] [blame]638compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek51540f92009-12-03 02:04:01 +0000[diff] [blame]639
640<!-- ======================================================================= -->
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]641<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
642<!-- ======================================================================= -->
643
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]644<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
645to query if certain upcoming standard language features are enabled. Those
646features are listed here. Features that are not yet implemented will be
647noted.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]648
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]649<h3 id="cxx0x">C++11</h3>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]650
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]651<p>The features listed below are slated for inclusion in the upcoming
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]652C++11 standard. As a result, all these features are enabled
653with the <tt>-std=c++11</tt> option when compiling C++ code.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]654
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]655<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor347d6262011-05-11 23:45:11 +0000[diff] [blame]656
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]657<p>Use <tt>__has_feature(cxx_access_control_sfinae)</tt> or <tt>__has_extension(cxx_access_control_sfinae)</tt> to determine whether access-control errors (e.g., calling a private constructor) are considered to be template argument deduction errors (aka SFINAE errors), per <a href="http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#1170">C++ DR1170</a>.</p>
Douglas Gregor347d6262011-05-11 23:45:11 +0000[diff] [blame]658
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]659<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3f1b5d02011-05-05 21:57:07 +0000[diff] [blame]660
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]661<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
662<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]663C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3f1b5d02011-05-05 21:57:07 +0000[diff] [blame]664
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]665<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbourne60188eb2011-10-14 23:44:46 +0000[diff] [blame]666
667<p>Use <tt>__has_feature(cxx_alignas)</tt> or
668<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
669specifiers using <tt>alignas</tt> is enabled.</p>
670
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]671<h4 id="cxx_attributes">C++11 attributes</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]672
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]673<p>Use <tt>__has_feature(cxx_attributes)</tt> or
674<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]675parsing with C++11's square bracket notation is enabled.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]676
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]677<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]678
679<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
680for generalized constant expressions (e.g., <tt>constexpr</tt>) is
Richard Smith1cb2af02012-02-14 22:56:17 +0000[diff] [blame]681enabled.</p>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]682
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]683<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]684
685<p>Use <tt>__has_feature(cxx_decltype)</tt> or
686<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
687<tt>decltype()</tt> specifier is enabled.</p>
688
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]689<h4 id="cxx_default_function_template_args">C++11 default template arguments in function templates</h4>
Douglas Gregor7115aee2011-02-05 20:35:30 +0000[diff] [blame]690
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]691<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
692<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
693if support for default template arguments in function templates is enabled.</p>
Douglas Gregor7115aee2011-02-05 20:35:30 +0000[diff] [blame]694
Douglas Gregor3c469762011-11-01 01:19:34 +0000[diff] [blame]695<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
696
697<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
698<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
699defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
700
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]701<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Alexis Huntb0d81992011-06-23 06:11:37 +0000[diff] [blame]702
703<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
704support for delegating constructors is enabled.</p>
705
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]706<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]707
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]708<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
709<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redla93bb5b2010-08-31 23:28:47 +0000[diff] [blame]710deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]711
Benjamin Kramereaa262b2012-01-15 15:26:07 +0000[diff] [blame]712<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]713<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
714
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]715<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Alexis Hunt7d1492a2011-08-07 00:34:32 +0000[diff] [blame]716
717<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
718support for generalized initializers (using braced lists and
Richard Smith2dc867f2012-02-26 07:09:21 +0000[diff] [blame]719<tt>std::initializer_list</tt>) is enabled.</p>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]720
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]721<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]722
Sebastian Redl6ce53ae2011-10-10 18:10:00 +0000[diff] [blame]723<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
724implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]725
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]726<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]727
728<p>Use <tt>__has_feature(cxx_inheriting_constructors)</tt> to determine if support for inheriting constructors is enabled. Clang does not currently implement this feature.</p>
729
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]730<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]731
732<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
733<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
734inline namespaces is enabled.</p>
Alexis Hunt7d1492a2011-08-07 00:34:32 +0000[diff] [blame]735
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]736<h4 id="cxx_lambdas">C++11 lambdas</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]737
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]738<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
739<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregor64bfb2e2012-02-23 05:44:09 +0000[diff] [blame]740is enabled. </p>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]741
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]742<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]743
744<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
745<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
746exception specifications is enabled.</p>
747
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]748<h4 id="cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]749
750<p>Use <tt>__has_feature(cxx_nonstatic_member_init)</tt> to determine whether in-class initialization of non-static data members is enabled.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]751
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]752<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]753
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]754<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
755<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor80af3132011-05-21 23:15:46 +0000[diff] [blame]756<tt>nullptr</tt> is enabled.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]757
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]758<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlsson69f1f932011-03-25 15:04:23 +0000[diff] [blame]759
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]760<p>Use <tt>__has_feature(cxx_override_control)</tt> or
761<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlsson69f1f932011-03-25 15:04:23 +0000[diff] [blame]762the override control keywords is enabled.</p>
763
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]764<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]765<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
766<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
767if support for reference-qualified functions (e.g., member functions with
768<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
769is enabled.</p>
Douglas Gregor56d5f0a2011-01-26 21:25:54 +0000[diff] [blame]770
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]771<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smith9bc6eed2011-04-15 15:14:40 +0000[diff] [blame]772
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]773<p>Use <tt>__has_feature(cxx_range_for)</tt> or
774<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
775range-based for loop is enabled. </p>
Richard Smith9bc6eed2011-04-15 15:14:40 +0000[diff] [blame]776
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]777<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]778<p>Use <tt>__has_feature(cxx_raw_string_literals)</tt> to determine if support for raw string literals (e.g., <tt>R"foo\bar"</tt>) is enabled.</p>
779
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]780<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]781
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]782<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
783<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56d5f0a2011-01-26 21:25:54 +0000[diff] [blame]784rvalue references is enabled. </p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]785
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]786<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]787
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]788<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
789<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]790compile-time assertions using <tt>static_assert</tt> is enabled.</p>
791
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]792<h4 id="cxx_auto_type">C++11 type inference</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]793
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]794<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]795<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]796supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
797will instead be a storage class specifier, as in C or C++98.</p>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]798
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]799<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]800
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]801<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
802<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
803strongly typed, scoped enumerations is enabled.</p>
Sebastian Redla93bb5b2010-08-31 23:28:47 +0000[diff] [blame]804
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]805<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregor7fb25412010-10-01 18:44:50 +0000[diff] [blame]806
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]807<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
808<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
809alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregor7fb25412010-10-01 18:44:50 +0000[diff] [blame]810
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]811<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]812<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
813support for Unicode string literals is enabled.</p>
Sebastian Redlc17003c2011-03-15 21:17:12 +0000[diff] [blame]814
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]815<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redlc17003c2011-03-15 21:17:12 +0000[diff] [blame]816
Richard Smith522fa532012-03-03 23:51:05 +0000[diff] [blame^]817<p>Use <tt>__has_feature(cxx_unrestricted_unions)</tt> to determine if support for unrestricted unions is enabled.</p>
Douglas Gregor0bf31402010-10-08 23:50:27 +0000[diff] [blame]818
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]819<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]820
821<p>Use <tt>__has_feature(cxx_user_literals)</tt> to determine if support for user-defined literals is enabled. Clang does not currently support this feature.</p>
822
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]823<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregor26858b72011-08-29 17:28:38 +0000[diff] [blame]824
825<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
826<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
827for variadic templates is enabled.</p>
Douglas Gregor0bf31402010-10-08 23:50:27 +0000[diff] [blame]828
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]829<h3 id="c11">C11</h3>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]830
831<p>The features listed below are slated for inclusion in the upcoming
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]832C11 standard. As a result, all these features are enabled
833with the <tt>-std=c11</tt> option when compiling C code.</p>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]834
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]835<h4 id="c_alignas">C11 alignment specifiers</h4>
Peter Collingbourne60188eb2011-10-14 23:44:46 +0000[diff] [blame]836
837<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
838to determine if support for alignment specifiers using <tt>_Alignas</tt>
839is enabled.</p>
840
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]841<h4 id="c_generic_selections">C11 generic selections</h4>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]842
843<p>Use <tt>__has_feature(c_generic_selections)</tt> or
844<tt>__has_extension(c_generic_selections)</tt> to determine if support for
845generic selections is enabled.</p>
846
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]847<p>As an extension, the C11 generic selection expression is available in all
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]848languages supported by Clang. The syntax is the same as that given in the
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]849C11 standard.</p>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]850
851<p>In C, type compatibility is decided according to the rules given in the
852appropriate standard, but in C++, which lacks the type compatibility rules
853used in C, types are considered compatible only if they are equivalent.</p>
854
Benjamin Kramere56f3932011-12-23 17:00:35 +0000[diff] [blame]855<h4 id="c_static_assert">C11 <tt>_Static_assert()</tt></h4>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]856
857<p>Use <tt>__has_feature(c_static_assert)</tt> or
858<tt>__has_extension(c_static_assert)</tt> to determine if support for
859compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
860
Alexis Huntaf4e28d2010-01-13 08:31:49 +0000[diff] [blame]861<!-- ======================================================================= -->
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]862<h2 id="checking_type_traits">Checks for Type Traits</h2>
863<!-- ======================================================================= -->
864
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]865<p>Clang supports the <a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html">GNU C++ type traits</a> and a subset of the <a href="http://msdn.microsoft.com/en-us/library/ms177194(v=VS.100).aspx">Microsoft Visual C++ Type traits</a>. For each supported type trait <code>__X</code>, <code>__has_extension(X)</code> indicates the presence of the type trait. For example:
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]866<blockquote>
867<pre>
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]868#if __has_extension(is_convertible_to)
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]869template&lt;typename From, typename To&gt;
870struct is_convertible_to {
871 static const bool value = __is_convertible_to(From, To);
872};
873#else
874// Emulate type trait
875#endif
876</pre>
877</blockquote>
878
879<p>The following type traits are supported by Clang:</p>
880<ul>
881 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
882 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
883 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
884 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
885 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
886 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
887 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
888 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
889 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
890 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
891 <li><code>__is_class</code> (GNU, Microsoft)</li>
892 <li><code>__is_convertible_to</code> (Microsoft)</li>
893 <li><code>__is_empty</code> (GNU, Microsoft)</li>
894 <li><code>__is_enum</code> (GNU, Microsoft)</li>
895 <li><code>__is_pod</code> (GNU, Microsoft)</li>
896 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
897 <li><code>__is_union</code> (GNU, Microsoft)</li>
898 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
Douglas Gregordca70af2011-12-03 18:14:24 +0000[diff] [blame]899 <li><code>__is_final</code>: Determines whether the given type is declared with a <code>final</code> class-virt-specifier.</li>
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]900 <li><code>__underlying_type(type)</code>: Retrieves the underlying type for a given <code>enum</code> type. This trait is required to implement the C++11 standard library.</li>
Douglas Gregor29c42f22012-02-24 07:38:34 +0000[diff] [blame]901 <li><code>__is_trivially_assignable(totype, fromtype)</code>: Determines whether a value of type <tt>totype</tt> can be assigned to from a value of type <tt>fromtype</tt> such that no non-trivial functions are called as part of that assignment. This trait is required to implement the C++11 standard library.</li>
902 <li><code>__is_trivially_constructible(type, argtypes...)</code>: Determines whether a value of type <tt>type</tt> can be direct-initialized with arguments of types <tt>argtypes...</tt> such that no non-trivial functions are called as part of that initialization. This trait is required to implement the C++11 standard library.</li>
Douglas Gregor7c607682011-02-03 21:57:35 +0000[diff] [blame]903</ul>
904
905<!-- ======================================================================= -->
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]906<h2 id="blocks">Blocks</h2>
907<!-- ======================================================================= -->
908
Chris Lattner14aea9a2009-03-09 07:03:22 +0000[diff] [blame]909<p>The syntax and high level language feature description is in <a
910href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
911details for the clang implementation are in <a
Chris Lattner87c5e2f2010-03-16 21:43:03 +0000[diff] [blame]912href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]913
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]914
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]915<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]916
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]917<!-- ======================================================================= -->
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]918<h2 id="objc_features">Objective-C Features</h2>
919<!-- ======================================================================= -->
920
921<h3 id="objc_instancetype">Related result types</h3>
922
923<p>According to Cocoa conventions, Objective-C methods with certain names ("init", "alloc", etc.) always return objects that are an instance of the receiving class's type. Such methods are said to have a "related result type", meaning that a message send to one of these methods will have the same static type as an instance of the receiver class. For example, given the following classes:</p>
924
925<blockquote>
926<pre>
927@interface NSObject
928+ (id)alloc;
929- (id)init;
930@end
931
932@interface NSArray : NSObject
933@end
934</pre>
935</blockquote>
936
937<p>and this common initialization pattern</p>
938
939<blockquote>
940<pre>
941NSArray *array = [[NSArray alloc] init];
942</pre>
943</blockquote>
944
945<p>the type of the expression <code>[NSArray alloc]</code> is
946<code>NSArray*</code> because <code>alloc</code> implicitly has a
947related result type. Similarly, the type of the expression
948<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
949<code>init</code> has a related result type and its receiver is known
950to have the type <code>NSArray *</code>. If neither <code>alloc</code> nor <code>init</code> had a related result type, the expressions would have had type <code>id</code>, as declared in the method signature.</p>
951
Douglas Gregorbab8a962011-09-08 01:46:34 +0000[diff] [blame]952<p>A method with a related result type can be declared by using the
953type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
954is a contextual keyword that is only permitted in the result type of
955an Objective-C method, e.g.</p>
956
957<pre>
958@interface A
959+ (<b>instancetype</b>)constructAnA;
960@end
961</pre>
962
963<p>The related result type can also be inferred for some methods.
964To determine whether a method has an inferred related result type, the first
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]965word in the camel-case selector (e.g., "init" in "initWithObjects") is
Douglas Gregorf8b02ca2011-11-03 18:33:01 +0000[diff] [blame]966considered, and the method will have a related result type if its return
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]967type is compatible with the type of its class and if</p>
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]968
969<ul>
970
971 <li>the first word is "alloc" or "new", and the method is a class
972 method, or</li>
973
974 <li>the first word is "autorelease", "init", "retain", or "self",
975 and the method is an instance method.</li>
976
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]977</ul>
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]978
979<p>If a method with a related result type is overridden by a subclass
980method, the subclass method must also return a type that is compatible
981with the subclass type. For example:</p>
982
983<blockquote>
984<pre>
985@interface NSString : NSObject
986- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
987@end
988</pre>
989</blockquote>
990
991<p>Related result types only affect the type of a message send or
992property access via the given method. In all other respects, a method
Douglas Gregorbab8a962011-09-08 01:46:34 +0000[diff] [blame]993with a related result type is treated the same way as method that
994returns <tt>id</tt>.</p>
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]995
Douglas Gregordc28ba12011-09-08 17:19:31 +0000[diff] [blame]996<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
997the <tt>instancetype</tt> contextual keyword is available.</p>
998
Douglas Gregor33823722011-06-11 01:09:30 +0000[diff] [blame]999<!-- ======================================================================= -->
John McCall31168b02011-06-15 23:02:42 +0000[diff] [blame]1000<h2 id="objc_arc">Automatic reference counting </h2>
1001<!-- ======================================================================= -->
1002
1003<p>Clang provides support for <a href="AutomaticReferenceCounting.html">automated reference counting</a> in Objective-C, which eliminates the need for manual retain/release/autorelease message sends. There are two feature macros associated with automatic reference counting: <code>__has_feature(objc_arc)</code> indicates the availability of automated reference counting in general, while <code>__has_feature(objc_arc_weak)</code> indicates that automated reference counting also includes support for <code>__weak</code> pointers to Objective-C objects.</p>
1004
1005<!-- ======================================================================= -->
Douglas Gregor8b7d4032011-09-08 17:18:35 +0000[diff] [blame]1006<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
1007<!-- ======================================================================= -->
1008
David Blaikie83172452011-10-18 05:49:30 +0000[diff] [blame]1009<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor8b7d4032011-09-08 17:18:35 +0000[diff] [blame]1010underlying type within Objective-C. For example, one can write an
1011enumeration type as:</p>
1012
1013<pre>
1014typedef enum : unsigned char { Red, Green, Blue } Color;
1015</pre>
1016
1017<p>This specifies that the underlying type, which is used to store the
1018enumeration value, is <tt>unsigned char</tt>.</p>
1019
1020<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
1021support for fixed underlying types is available in Objective-C.</p>
1022
1023<!-- ======================================================================= -->
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1024<h2 id="overloading-in-c">Function Overloading in C</h2>
1025<!-- ======================================================================= -->
1026
Chris Lattnerb3ef5332009-02-13 21:51:45 +0000[diff] [blame]1027<p>Clang provides support for C++ function overloading in C. Function
1028overloading in C is introduced using the <tt>overloadable</tt> attribute. For
1029example, one might provide several overloaded versions of a <tt>tgsin</tt>
1030function that invokes the appropriate standard function computing the sine of a
1031value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
1032precision:</p>
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1033
1034<blockquote>
1035<pre>
1036#include &lt;math.h&gt;
1037float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
1038double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
1039long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
1040</pre>
1041</blockquote>
1042
1043<p>Given these declarations, one can call <tt>tgsin</tt> with a
1044<tt>float</tt> value to receive a <tt>float</tt> result, with a
1045<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
1046overloading in C follows the rules of C++ function overloading to pick
1047the best overload given the call arguments, with a few C-specific
1048semantics:</p>
1049<ul>
1050 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
1051 double</tt> is ranked as a floating-point promotion (per C99) rather
1052 than as a floating-point conversion (as in C++).</li>
1053
1054 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
1055 <tt>U*</tt> is considered a pointer conversion (with conversion
1056 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
1057
1058 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
1059 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
1060 conversion is given "conversion" rank.</li>
1061</ul>
1062
1063<p>The declaration of <tt>overloadable</tt> functions is restricted to
1064function declarations and definitions. Most importantly, if any
1065function with a given name is given the <tt>overloadable</tt>
1066attribute, then all function declarations and definitions with that
1067name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerb3ef5332009-02-13 21:51:45 +0000[diff] [blame]1068attribute. This rule even applies to redeclarations of functions whose original
1069declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1070
1071<blockquote>
1072<pre>
1073int f(int) __attribute__((overloadable));
1074float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
1075
1076int g(int) __attribute__((overloadable));
1077int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
1078</pre>
1079</blockquote>
1080
Douglas Gregor94349fd2009-02-18 07:07:28 +0000[diff] [blame]1081<p>Functions marked <tt>overloadable</tt> must have
1082prototypes. Therefore, the following code is ill-formed:</p>
1083
1084<blockquote>
1085<pre>
1086int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
1087</pre>
1088</blockquote>
1089
1090<p>However, <tt>overloadable</tt> functions are allowed to use a
1091ellipsis even if there are no named parameters (as is permitted in C++). This feature is particularly useful when combined with the <tt>unavailable</tt> attribute:</p>
1092
1093<blockquote>
1094<pre>
Chris Lattnerfaca7d92009-02-18 22:27:46 +0000[diff] [blame]1095void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor94349fd2009-02-18 07:07:28 +0000[diff] [blame]1096</pre>
1097</blockquote>
1098
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1099<p>Functions declared with the <tt>overloadable</tt> attribute have
1100their names mangled according to the same rules as C++ function
1101names. For example, the three <tt>tgsin</tt> functions in our
1102motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner19e0fd02010-11-28 18:19:13 +0000[diff] [blame]1103<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1104caveats to this use of name mangling:</p>
1105
1106<ul>
1107
1108 <li>Future versions of Clang may change the name mangling of
1109 functions overloaded in C, so you should not depend on an specific
1110 mangling. To be completely safe, we strongly urge the use of
1111 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
1112
1113 <li>The <tt>overloadable</tt> attribute has almost no meaning when
1114 used in C++, because names will already be mangled and functions are
1115 already overloadable. However, when an <tt>overloadable</tt>
1116 function occurs within an <tt>extern "C"</tt> linkage specification,
1117 it's name <i>will</i> be mangled in the same way as it would in
1118 C.</li>
1119</ul>
1120
Peter Collingbourned5d410f2011-05-13 20:54:45 +0000[diff] [blame]1121<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]1122
Eli Friedman6b9c41e2011-09-19 23:17:44 +0000[diff] [blame]1123<!-- ======================================================================= -->
1124<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
1125<!-- ======================================================================= -->
1126
1127<p>clang supports an extension which allows the following in C:</p>
1128
1129<blockquote>
1130<pre>
1131#include &lt;math.h&gt;
1132#include &lt;complex.h&gt;
1133complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
1134</pre>
1135</blockquote>
1136
1137<p>This construct is useful because there is no way to separately
1138initialize the real and imaginary parts of a complex variable in
1139standard C, given that clang does not support <code>_Imaginary</code>.
1140(clang also supports the <code>__real__</code> and <code>__imag__</code>
1141extensions from gcc, which help in some cases, but are not usable in
1142static initializers.)
1143
1144<p>Note that this extension does not allow eliding the braces; the
1145meaning of the following two lines is different:</p>
1146
1147<blockquote>
1148<pre>
1149complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
1150complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
1151</pre>
1152</blockquote>
1153
1154<p>This extension also works in C++ mode, as far as that goes, but does not
1155 apply to the C++ <code>std::complex</code>. (In C++11, list
1156 initialization allows the same syntax to be used with
1157 <code>std::complex</code> with the same meaning.)
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]1158
Douglas Gregor725de572009-02-13 00:57:04 +0000[diff] [blame]1159<!-- ======================================================================= -->
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1160<h2 id="builtins">Builtin Functions</h2>
1161<!-- ======================================================================= -->
1162
1163<p>Clang supports a number of builtin library functions with the same syntax as
1164GCC, including things like <tt>__builtin_nan</tt>,
1165<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
1166<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
1167addition to the GCC builtins, Clang supports a number of builtins that GCC does
1168not, which are listed here.</p>
1169
1170<p>Please note that Clang does not and will not support all of the GCC builtins
1171for vector operations. Instead of using builtins, you should use the functions
1172defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
1173define portable wrappers for these. Many of the Clang versions of these
1174functions are implemented directly in terms of <a href="#vectors">extended
1175vector support</a> instead of builtins, in order to reduce the number of
1176builtins that we need to implement.</p>
1177
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1178<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]1179<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1180<!-- ======================================================================= -->
1181
Chris Lattnerc15c2522009-09-16 18:56:12 +0000[diff] [blame]1182<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner9631dfd2009-02-13 20:00:20 +0000[diff] [blame]1183permutation/shuffle/swizzle operations. This builtin is also very important for
1184the implementation of various target-specific header files like
1185<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1186</p>
1187
1188<p><b>Syntax:</b></p>
1189
1190<pre>
Chris Lattner9631dfd2009-02-13 20:00:20 +0000[diff] [blame]1191__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1192</pre>
1193
1194<p><b>Examples:</b></p>
1195
1196<pre>
Chris Lattner9631dfd2009-02-13 20:00:20 +0000[diff] [blame]1197 // Identity operation - return 4-element vector V1.
1198 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1199
1200 // "Splat" element 0 of V1 into a 4-element result.
1201 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1202
1203 // Reverse 4-element vector V1.
1204 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1205
1206 // Concatenate every other element of 4-element vectors V1 and V2.
1207 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1208
1209 // Concatenate every other element of 8-element vectors V1 and V2.
1210 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1211</pre>
1212
1213<p><b>Description:</b></p>
1214
Chris Lattner9631dfd2009-02-13 20:00:20 +0000[diff] [blame]1215<p>The first two arguments to __builtin_shufflevector are vectors that have the
1216same element type. The remaining arguments are a list of integers that specify
1217the elements indices of the first two vectors that should be extracted and
1218returned in a new vector. These element indices are numbered sequentially
1219starting with the first vector, continuing into the second vector. Thus, if
1220vec1 is a 4-element vector, index 5 would refer to the second element of vec2.
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1221</p>
1222
Chris Lattner9631dfd2009-02-13 20:00:20 +0000[diff] [blame]1223<p>The result of __builtin_shufflevector is a vector
1224with the same element type as vec1/vec2 but that has an element count equal to
1225the number of indices specified.
1226</p>
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1227
Chris Lattnerbf206382009-09-21 03:09:59 +0000[diff] [blame]1228<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1229
1230<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]1231<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattnerbf206382009-09-21 03:09:59 +0000[diff] [blame]1232<!-- ======================================================================= -->
1233
1234<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1235the program cannot be reached, even if the compiler might otherwise think it
1236can. This is useful to improve optimization and eliminates certain warnings.
1237For example, without the <tt>__builtin_unreachable</tt> in the example below,
1238the compiler assumes that the inline asm can fall through and prints a "function
1239declared 'noreturn' should not return" warning.
1240</p>
1241
1242<p><b>Syntax:</b></p>
1243
1244<pre>
1245__builtin_unreachable()
1246</pre>
1247
1248<p><b>Example of Use:</b></p>
1249
1250<pre>
1251void myabort(void) __attribute__((noreturn));
1252void myabort(void) {
1253 asm("int3");
1254 __builtin_unreachable();
1255}
1256</pre>
1257
1258<p><b>Description:</b></p>
1259
1260<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1261it has undefined behavior, it is a statement that it is never reached and the
1262optimizer can take advantage of this to produce better code. This builtin takes
1263no arguments and produces a void result.
1264</p>
1265
1266<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1267
Chris Lattner9cb59fa2011-04-09 03:57:26 +0000[diff] [blame]1268<!-- ======================================================================= -->
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]1269<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner9cb59fa2011-04-09 03:57:26 +0000[diff] [blame]1270<!-- ======================================================================= -->
1271
1272<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1273memory.
1274</p>
1275
1276<p><b>Syntax:</b></p>
1277
1278<pre>
1279<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1280</pre>
1281
1282<p><b>Example of Use:</b></p>
1283
1284<pre>
Alexis Huntaedbfe42011-06-23 01:21:01 +0000[diff] [blame]1285int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner9cb59fa2011-04-09 03:57:26 +0000[diff] [blame]1286</pre>
1287
1288<p><b>Description:</b></p>
1289
1290<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1291intrinsics to allow code to atomically swap the current value with the new
1292value. More importantly, it helps developers write more efficient and correct
1293code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1294relying on the platform specific implementation details of
1295__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1296</p>
1297
Chris Lattnerbf206382009-09-21 03:09:59 +0000[diff] [blame]1298
Chris Lattnerc86ffc32009-04-09 19:58:15 +0000[diff] [blame]1299<!-- ======================================================================= -->
1300<h2 id="targetspecific">Target-Specific Extensions</h2>
1301<!-- ======================================================================= -->
1302
1303<p>Clang supports some language features conditionally on some targets.</p>
1304
1305<!-- ======================================================================= -->
1306<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1307<!-- ======================================================================= -->
1308
1309<p>The X86 backend has these language extensions:</p>
1310
1311<!-- ======================================================================= -->
1312<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1313<!-- ======================================================================= -->
1314
1315<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnerb7323972009-05-05 18:54:47 +0000[diff] [blame]1316relative to the X86 GS segment register, and address space #257 causes it to be
1317relative to the X86 FS segment. Note that this is a very very low-level
1318feature that should only be used if you know what you're doing (for example in
1319an OS kernel).</p>
Chris Lattnerc86ffc32009-04-09 19:58:15 +0000[diff] [blame]1320
1321<p>Here is an example:</p>
1322
1323<pre>
1324#define GS_RELATIVE __attribute__((address_space(256)))
1325int foo(int GS_RELATIVE *P) {
1326 return *P;
1327}
1328</pre>
1329
1330<p>Which compiles to (on X86-32):</p>
1331
1332<pre>
1333_foo:
1334 movl 4(%esp), %eax
1335 movl %gs:(%eax), %eax
1336 ret
1337</pre>
1338
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1339<!-- ======================================================================= -->
1340<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1341<!-- ======================================================================= -->
1342
1343<p>Clang supports additional attributes that are useful for documenting program
1344invariants and rules for static analysis tools. The extensions documented here
1345are used by the <a
1346href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1347engine</a> that is part of Clang's Analysis library.</p>
1348
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]1349<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1350
1351<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek86ee0592009-04-10 05:04:22 +0000[diff] [blame]1352attribute. This attribute, which is typically affixed to a function prototype,
1353indicates that a call to a given function never returns. Function prototypes for
1354common functions like <tt>exit</tt> are typically annotated with this attribute,
1355as well as a variety of common assertion handlers. Users can educate the static
1356analyzer about their own custom assertion handles (thus cutting down on false
1357positives due to false paths) by marking their own &quot;panic&quot; functions
1358with this attribute.</p>
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1359
1360<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewyckybe42c112009-06-14 04:08:08 +0000[diff] [blame]1361there are special functions that for all intents and purposes should be
1362considered panic functions (i.e., they are only called when an internal program
1363error occurs) but may actually return so that the program can fail gracefully.
1364The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1365as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattnerbb72f772009-04-10 05:54:56 +0000[diff] [blame]1366pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1367<tt>noreturn</tt>).</p>
1368
1369<p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the
Chris Lattnerbb72f772009-04-10 05:54:56 +0000[diff] [blame]1370same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1371placed at the end of function prototypes:</p>
1372
1373<pre>
1374 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]1375</pre>
1376
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]1377<p>Query for this feature with
1378<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattnerb6f77af2009-06-13 07:13:28 +0000[diff] [blame]1379
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]1380<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1381
1382<p>Many methods in Objective-C have conventional meanings determined
1383by their selectors. For the purposes of static analysis, it is
1384sometimes useful to be able to mark a method as having a particular
1385conventional meaning despite not having the right selector, or as not
1386having the conventional meaning that its selector would suggest.
1387For these use cases, we provide an attribute to specifically describe
1388the <q>method family</q> that a method belongs to.</p>
1389
1390<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1391where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1392<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1393can only be placed at the end of a method declaration:</p>
1394
1395<pre>
1396 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1397</pre>
1398
1399<p>Users who do not wish to change the conventional meaning of a
1400method, and who merely want to document its non-standard retain and
1401release semantics, should use the
1402<a href="#attr_retain_release">retaining behavior attributes</a>
1403described below.</p>
1404
1405<p>Query for this feature with
1406<tt>__has_attribute(objc_method_family)</tt>.</p>
1407
1408<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall46669c02011-01-25 04:26:21 +0000[diff] [blame]1409
1410<p>In Objective-C, functions and methods are generally assumed to take
1411and return objects with +0 retain counts, with some exceptions for
1412special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1413there are exceptions, and so Clang provides attributes to allow these
1414exceptions to be documented, which helps the analyzer find leaks (and
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]1415ignore non-leaks). Some exceptions may be better described using
1416the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1417attribute instead.</p>
John McCall46669c02011-01-25 04:26:21 +0000[diff] [blame]1418
1419<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1420<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1421and <tt>cf_returns_not_retained</tt> attributes can be placed on
1422methods and functions that return Objective-C or CoreFoundation
1423objects. They are commonly placed at the end of a function prototype
1424or method declaration:</p>
1425
1426<pre>
1427 id foo() <b>__attribute__((ns_returns_retained))</b>;
1428
1429 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1430</pre>
1431
1432<p>The <tt>*_returns_retained</tt> attributes specify that the
1433returned object has a +1 retain count.
1434The <tt>*_returns_not_retained</tt> attributes specify that the return
1435object has a +0 retain count, even if the normal convention for its
1436selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1437returned object is +0, but is guaranteed to live at least as long as the
1438next flush of an autorelease pool.</p>
1439
1440<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1441attributes can be placed on an parameter declaration; they specify
1442that the argument is expected to have a +1 retain count, which will be
1443balanced in some way by the function or method.
1444The <tt>ns_consumes_self</tt> attribute can only be placed on an
1445Objective-C method; it specifies that the method expects
1446its <tt>self</tt> parameter to have a +1 retain count, which it will
1447balance in some way.</p>
1448
1449<pre>
1450 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1451
1452 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1453 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1454</pre>
Ted Kremeneka9deb972009-04-10 05:03:33 +0000[diff] [blame]1455
John McCall6107fc92011-03-18 03:51:49 +0000[diff] [blame]1456<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1457<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1458
Kostya Serebryany9751d342011-11-28 20:51:02 +0000[diff] [blame]1459<!-- ======================================================================= -->
1460<h2 id="dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</h2>
1461<!-- ======================================================================= -->
1462<h3 id="address_sanitizer">AddressSanitizer</h3>
1463<p> Use <code>__has_feature(address_sanitizer)</code>
1464to check if the code is being built with <a
1465 href="AddressSanitizer.html">AddressSanitizer</a>.
1466</p>
Kostya Serebryany588d6ab2012-01-24 19:25:38 +0000[diff] [blame]1467<p>Use <tt>__attribute__((no_address_safety_analysis))</tt> on a function
1468declaration to specify that address safety instrumentation (e.g.
1469AddressSanitizer) should not be applied to that function.
1470</p>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1471
1472<!-- ======================================================================= -->
Caitlin Sadowski13d44442011-07-28 18:38:36 +0000[diff] [blame]1473<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1474<!-- ======================================================================= -->
1475
1476<p>Clang supports additional attributes for checking basic locking policies in
1477multithreaded programs.
1478Clang currently parses the following list of attributes, although
1479<b>the implementation for these annotations is currently in development.</b>
1480For more details, see the
1481<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1482</p>
1483
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1484<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1485
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1486<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1487declaration to specify that the thread safety analysis should not be run on that
1488function. This attribute provides an escape hatch (e.g. for situations when it
1489is difficult to annotate the locking policy). </p>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1490
1491<h4 id="ts_lockable">lockable</h4>
1492
1493<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1494that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1495used to check consistency.</p>
1496
1497<h4 id="ts_scopedlockable">scoped_lockable</h4>
1498
1499<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1500specify that it has a "scoped" lockable type. Objects of this type will acquire
1501the lock upon construction and release it upon going out of scope.
1502 This annotation is primarily used to check
1503consistency.</p>
1504
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1505<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1506
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1507<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1508specify that the variable must be accessed while holding some lock.</p>
1509
1510<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1511
1512<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1513specify that the pointer must be dereferenced while holding some lock.</p>
1514
1515<h4 id="ts_guardedby">guarded_by(l)</h4>
1516
1517<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1518specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1519
1520<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1521
1522<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1523specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1524
1525<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1526
1527<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1528of a lockable variable to specify that the lock must be acquired before all
1529attribute arguments. Arguments must be lockable type, and there must be at
1530least one argument.</p>
1531
1532<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1533
1534<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1535of a lockable variable to specify that the lock must be acquired after all
1536attribute arguments. Arguments must be lockable type, and there must be at
1537least one argument.</p>
1538
1539<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1540
1541<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1542declaration to specify that the function acquires all listed locks
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1543exclusively. This attribute takes zero or more arguments: either of lockable
1544type or integers indexing into function parameters of lockable type. If no
1545arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1546enclosing object.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1547
1548<h4 id="ts_slf">shared_lock_function(...)</h4>
1549
1550<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1551declaration to specify that the function acquires all listed locks, although
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1552 the locks may be shared (e.g. read locks). This attribute takes zero or more
1553arguments: either of lockable type or integers indexing into function
1554parameters of lockable type. If no arguments are given, the acquired lock is
1555implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1556
1557<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1558
1559<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1560declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1561all listed locks exclusively. This attribute takes one or more arguments. The
1562first argument is an integer or boolean value specifying the return value of a
1563successful lock acquisition. The remaining arugments are either of lockable type
1564or integers indexing into function parameters of lockable type. If only one
1565argument is given, the acquired lock is implicitly <tt>this</tt> of the
1566enclosing object.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1567
1568<h4 id="ts_stf">shared_trylock_function(...)</h4>
1569
1570<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1571declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1572all listed locks, although the locks may be shared (e.g. read locks). This
1573attribute takes one or more arguments. The first argument is an integer or
1574boolean value specifying the return value of a successful lock acquisition. The
1575remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1576function parameters of lockable type. If only one argument is given, the
1577acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1578
1579<h4 id="ts_uf">unlock_function(...)</h4>
1580
1581<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1582declaration to specify that the function release all listed locks. This
1583attribute takes zero or more arguments: either of lockable type or integers
1584indexing into function parameters of lockable type. If no arguments are given,
1585the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1586
1587<h4 id="ts_lr">lock_returned(l)</h4>
1588
1589<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1590declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1591must be of lockable type). This annotation is used to aid in resolving lock
1592expressions.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1593
1594<h4 id="ts_le">locks_excluded(...)</h4>
1595
1596<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowski4b1e8392011-08-09 17:59:31 +0000[diff] [blame]1597to specify that the function must not be called with the listed locks. Arguments
1598must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowski63fa6672011-07-28 20:12:35 +0000[diff] [blame]1599
1600<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1601
1602<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1603declaration to specify that the function must be called while holding the listed
1604exclusive locks. Arguments must be lockable type, and there must be at
1605least one argument.</p>
1606
1607<h4 id="ts_slr">shared_locks_required(...)</h4>
1608
1609<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1610declaration to specify that the function must be called while holding the listed
1611shared locks. Arguments must be lockable type, and there must be at
1612least one argument.</p>
Caitlin Sadowskiaac4d212011-07-28 17:21:07 +0000[diff] [blame]1613
Chris Lattnerb8ec6bb2009-02-09 08:46:11 +0000[diff] [blame]1614</div>
1615</body>
1616</html>