Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 62f123445722414127bffc10865cc2e34acfc608 / . / docs / LanguageExtensions.html
blob: d7e4931b8c8e834474f96fec3681ac441a3a9781 [file] [log] [blame]
Sean Hunt7e98b472011-06-23 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 Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]4<html>
5<head>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]6 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]7 <title>Clang Language Extensions</title>
Sean Hunt64f857b2011-06-23 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">
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]10 <style type="text/css">
11 td {
12 vertical-align: top;
13 }
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]14 th { background-color: #ffddaa; }
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]15 </style>
Chris Lattner5ce933f2009-02-09 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 Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]27<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]28<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]29<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]30<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCall48209082010-11-08 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>
Douglas Gregor93a70672012-03-11 04:53:21 +0000[diff] [blame]33<li><a href="#availability">Availability attribute</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]34<li><a href="#checking_language_features">Checks for Standard Language Features</a>
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]35 <ul>
36 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
37 <li><a href="#cxx_rtti">C++ RTTI</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]38 </ul></li>
39<li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]40 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]41 <li><a href="#cxx0x">C++11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]42 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]43 <li><a href="#cxx_access_control_sfinae">C++11 SFINAE includes
44 access control</a></li>
45 <li><a href="#cxx_alias_templates">C++11 alias templates</a></li>
46 <li><a href="#cxx_alignas">C++11 alignment specifiers</a></li>
47 <li><a href="#cxx_attributes">C++11 attributes</a></li>
48 <li><a href="#cxx_constexpr">C++11 generalized constant expressions</a></li>
49 <li><a href="#cxx_decltype">C++11 <tt>decltype()</tt></a></li>
50 <li><a href="#cxx_default_function_template_args">C++11 default template arguments in function templates</a></li>
Douglas Gregorf695a692011-11-01 01:19:34 +0000[diff] [blame]51 <li><a href="#cxx_defaulted_functions">C++11 defaulted functions</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]52 <li><a href="#cxx_delegating_constructor">C++11 delegating constructors</a></li>
53 <li><a href="#cxx_deleted_functions">C++11 deleted functions</a></li>
54 <li><a href="#cxx_explicit_conversions">C++11 explicit conversion functions</a></li>
55 <li><a href="#cxx_generalized_initializers">C++11 generalized initializers</a></li>
56 <li><a href="#cxx_implicit_moves">C++11 implicit move constructors/assignment operators</a></li>
57 <li><a href="#cxx_inheriting_constructors">C++11 inheriting constructors</a></li>
58 <li><a href="#cxx_inline_namespaces">C++11 inline namespaces</a></li>
59 <li><a href="#cxx_lambdas">C++11 lambdas</a></li>
60 <li><a href="#cxx_noexcept">C++11 noexcept specification</a></li>
61 <li><a href="#cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</a></li>
62 <li><a href="#cxx_nullptr">C++11 nullptr</a></li>
63 <li><a href="#cxx_override_control">C++11 override control</a></li>
64 <li><a href="#cxx_range_for">C++11 range-based for loop</a></li>
65 <li><a href="#cxx_raw_string_literals">C++11 raw string literals</a></li>
66 <li><a href="#cxx_rvalue_references">C++11 rvalue references</a></li>
67 <li><a href="#cxx_reference_qualified_functions">C++11 reference-qualified functions</a></li>
68 <li><a href="#cxx_static_assert">C++11 <tt>static_assert()</tt></a></li>
69 <li><a href="#cxx_auto_type">C++11 type inference</a></li>
70 <li><a href="#cxx_strong_enums">C++11 strongly-typed enumerations</a></li>
71 <li><a href="#cxx_trailing_return">C++11 trailing return type</a></li>
72 <li><a href="#cxx_unicode_literals">C++11 Unicode string literals</a></li>
73 <li><a href="#cxx_unrestricted_unions">C++11 unrestricted unions</a></li>
74 <li><a href="#cxx_user_literals">C++11 user-defined literals</a></li>
75 <li><a href="#cxx_variadic_templates">C++11 variadic templates</a></li>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]76 </ul></li>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]77 <li><a href="#c11">C11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]78 <ul>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]79 <li><a href="#c_alignas">C11 alignment specifiers</a></li>
80 <li><a href="#c_generic_selections">C11 generic selections</a></li>
81 <li><a href="#c_static_assert">C11 <tt>_Static_assert()</tt></a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]82 </ul></li>
83 </ul> </li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]84<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]85<li><a href="#blocks">Blocks</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]86<li><a href="#objc_features">Objective-C Features</a>
87 <ul>
88 <li><a href="#objc_instancetype">Related result types</a></li>
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]89 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]90 <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li>
Douglas Gregor8a4e1822012-03-09 23:24:48 +0000[diff] [blame]91 <li><a href="#objc_lambdas">Interoperability with C++11 lambdas</a></li>
Patrick Beard62f12342012-03-20 21:51:03 +0000[diff] [blame^]92 <li><a href="#object-literals-subscripting">Object Literals and Subscripting</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]93 </ul>
94</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]95<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]96<li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]97<li><a href="#builtins">Builtin Functions</a>
98 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]99 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]100 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]101 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]102 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]103</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]104<li><a href="#targetspecific">Target-Specific Extensions</a>
105 <ul>
106 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
107 </ul>
108</li>
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]109<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]110<li><a href="#dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</a>
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]111 <ul>
112 <li><a href="#address_sanitizer">AddressSanitizer</a></li>
113 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]114</li>
115<li><a href="#threadsafety">Thread Safety Annotation Checking</a>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]116 <ul>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]117 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
118 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
119 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]120 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
121 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]122 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
123 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
124 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
125 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
126 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
127 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
128 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
129 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
130 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
131 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
132 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
133 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
134 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]135 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]136</li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]137</ul>
138
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]139<!-- ======================================================================= -->
140<h2 id="intro">Introduction</h2>
141<!-- ======================================================================= -->
142
143<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]144addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]145range of GCC extensions. Please see the <a
146href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
147more information on these extensions.</p>
148
149<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]150<h2 id="feature_check">Feature Checking Macros</h2>
151<!-- ======================================================================= -->
152
153<p>Language extensions can be very useful, but only if you know you can depend
Chris Lattnerc70e1932011-03-21 16:25:11 +0000[diff] [blame]154on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]155function-like macros. This allows you to directly test for a feature in your
156code without having to resort to something like autoconf or fragile "compiler
157version checks".</p>
158
159<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]160<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]161<!-- ======================================================================= -->
162
163<p>This function-like macro takes a single identifier argument that is the name
164of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
165not. It can be used like this:</p>
166
167<blockquote>
168<pre>
169#ifndef __has_builtin // Optional of course.
170 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
171#endif
172
173...
174#if __has_builtin(__builtin_trap)
175 __builtin_trap();
176#else
177 abort();
178#endif
179...
180</pre>
181</blockquote>
182
183
184<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]185<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]186<!-- ======================================================================= -->
187
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]188<p>These function-like macros take a single identifier argument that is the
189name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
190is both supported by Clang and standardized in the current language standard
191or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
192<code>__has_extension</code> evaluates to 1 if the feature is supported by
193Clang in the current language (either as a language extension or a standard
194language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]195
196<blockquote>
197<pre>
198#ifndef __has_feature // Optional of course.
199 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
200#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]201#ifndef __has_extension
202 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
203#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]204
205...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]206#if __has_feature(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]207// This code will only be compiled with the -std=c++11 and -std=gnu++11
208// options, because rvalue references are only standardized in C++11.
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]209#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]210
211#if __has_extension(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]212// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]213// and -std=gnu++98 options, because rvalue references are supported as a
214// language extension in C++98.
215#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]216</pre>
217</blockquote>
218
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]219<p id="has_feature_back_compat">For backwards compatibility reasons,
220<code>__has_feature</code> can also be used to test for support for
221non-standardized features, i.e. features not prefixed <code>c_</code>,
222<code>cxx_</code> or <code>objc_</code>.</p>
223
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]224<p id="has_feature_for_non_language_features">
225Another use of <code>__has_feature</code> is to check for compiler features
226not related to the language standard, such as e.g.
227<a href="AddressSanitizer.html">AddressSanitizer</a>.
228
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]229<p>If the <code>-pedantic-errors</code> option is given,
230<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
231
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]232<p>The feature tag is described along with the language feature below.</p>
233
Richard Smith5297d712012-02-25 10:41:10 +0000[diff] [blame]234<p>The feature name or extension name can also be specified with a preceding and
235following <code>__</code> (double underscore) to avoid interference from a macro
Richard Smith1d9f4c12012-03-01 02:12:07 +0000[diff] [blame]236with the same name. For instance, <code>__cxx_rvalue_references__</code> can be
237used instead of <code>cxx_rvalue_references</code>.</p>
Richard Smith5297d712012-02-25 10:41:10 +0000[diff] [blame]238
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]239<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]240<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]241<!-- ======================================================================= -->
242
243<p>This function-like macro takes a single identifier argument that is the name
244of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
245can be used like this:</p>
246
247<blockquote>
248<pre>
249#ifndef __has_attribute // Optional of course.
250 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
251#endif
252
253...
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]254#if __has_attribute(always_inline)
255#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]256#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]257#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]258#endif
259...
260</pre>
261</blockquote>
262
Jean-Daniel Dupas8a5e7fd2012-03-01 14:53:16 +0000[diff] [blame]263<p>The attribute name can also be specified with a preceding and
264following <code>__</code> (double underscore) to avoid interference from a macro
265with the same name. For instance, <code>__always_inline__</code> can be used
266instead of <code>always_inline</code>.</p>
267
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]268<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]269<h2 id="has_include">Include File Checking Macros</h2>
270<!-- ======================================================================= -->
271
272<p>Not all developments systems have the same include files.
273The <a href="#__has_include">__has_include</a> and
274<a href="#__has_include_next">__has_include_next</a> macros allow you to
275check for the existence of an include file before doing
276a possibly failing #include directive.</p>
277
278<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]279<h3><a name="__has_include">__has_include</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]280<!-- ======================================================================= -->
281
282<p>This function-like macro takes a single file name string argument that
283is the name of an include file. It evaluates to 1 if the file can
284be found using the include paths, or 0 otherwise:</p>
285
286<blockquote>
287<pre>
288// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]289#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]290# include "myinclude.h"
291#endif
292
293// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]294#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]295# include "myinclude.h"
296#endif
297</pre>
298</blockquote>
299
300<p>To test for this feature, use #if defined(__has_include).</p>
301
302<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]303<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]304<!-- ======================================================================= -->
305
306<p>This function-like macro takes a single file name string argument that
307is the name of an include file. It is like __has_include except that it
308looks for the second instance of the given file found in the include
309paths. It evaluates to 1 if the second instance of the file can
310be found using the include paths, or 0 otherwise:</p>
311
312<blockquote>
313<pre>
314// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]315#if __has_include_next("myinclude.h") &amp;&amp; __has_include_next(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]316# include_next "myinclude.h"
317#endif
318
319// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]320#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]321# include_next "myinclude.h"
322#endif
323</pre>
324</blockquote>
325
326<p>Note that __has_include_next, like the GNU extension
327#include_next directive, is intended for use in headers only,
328and will issue a warning if used in the top-level compilation
329file. A warning will also be issued if an absolute path
330is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]331
Ted Kremenekd7681502011-10-12 19:46:30 +0000[diff] [blame]332
333<!-- ======================================================================= -->
334<h3><a name="__has_warning">__has_warning</a></h3>
335<!-- ======================================================================= -->
336
337<p>This function-like macro takes a string literal that represents a command
338 line option for a warning and returns true if that is a valid warning
339 option.</p>
340
341<blockquote>
342<pre>
343#if __has_warning("-Wformat")
344...
345#endif
346</pre>
347</blockquote>
348
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]349<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]350<h2 id="builtinmacros">Builtin Macros</h2>
351<!-- ======================================================================= -->
352
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]353<dl>
354 <dt><code>__BASE_FILE__</code></dt>
355 <dd>Defined to a string that contains the name of the main input
356 file passed to Clang.</dd>
357
358 <dt><code>__COUNTER__</code></dt>
359 <dd>Defined to an integer value that starts at zero and is
360 incremented each time the <code>__COUNTER__</code> macro is
361 expanded.</dd>
362
363 <dt><code>__INCLUDE_LEVEL__</code></dt>
364 <dd>Defined to an integral value that is the include depth of the
365 file currently being translated. For the main file, this value is
366 zero.</dd>
367
368 <dt><code>__TIMESTAMP__</code></dt>
369 <dd>Defined to the date and time of the last modification of the
370 current source file.</dd>
371
372 <dt><code>__clang__</code></dt>
373 <dd>Defined when compiling with Clang</dd>
374
375 <dt><code>__clang_major__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]376 <dd>Defined to the major marketing version number of Clang (e.g., the
377 2 in 2.0.1). Note that marketing version numbers should not be used to
378 check for language features, as different vendors use different numbering
379 schemes. Instead, use the <a href="#feature_check">feature checking
380 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]381
382 <dt><code>__clang_minor__</code></dt>
383 <dd>Defined to the minor version number of Clang (e.g., the 0 in
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]384 2.0.1). Note that marketing version numbers should not be used to
385 check for language features, as different vendors use different numbering
386 schemes. Instead, use the <a href="#feature_check">feature checking
387 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]388
389 <dt><code>__clang_patchlevel__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]390 <dd>Defined to the marketing patch level of Clang (e.g., the 1 in 2.0.1).</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]391
392 <dt><code>__clang_version__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]393 <dd>Defined to a string that captures the Clang marketing version, including
394 the Subversion tag or revision number, e.g., "1.5 (trunk 102332)".</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]395</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]396
397<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]398<h2 id="vectors">Vectors and Extended Vectors</h2>
399<!-- ======================================================================= -->
400
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]401<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]402
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]403<p>OpenCL vector types are created using <tt>ext_vector_type</tt> attribute. It
404support for <tt>V.xyzw</tt> syntax and other tidbits as seen in OpenCL. An
405example is:</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]406
407<blockquote>
408<pre>
409typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
410typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
411
412float4 foo(float2 a, float2 b) {
413 float4 c;
414 c.xz = a;
415 c.yw = b;
416 return c;
417}
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]418</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]419</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]420
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]421<p>Query for this feature with
422<tt>__has_extension(attribute_ext_vector_type)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]423
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]424<p>Giving <tt>-faltivec</tt> option to clang enables support for AltiVec vector
425syntax and functions. For example:</p>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]426
427<blockquote>
428<pre>
429vector float foo(vector int a) {
430 vector int b;
431 b = vec_add(a, a) + a;
432 return (vector float)b;
433}
434</pre>
435</blockquote>
436
437<p>NEON vector types are created using <tt>neon_vector_type</tt> and
438<tt>neon_polyvector_type</tt> attributes. For example:</p>
439
440<blockquote>
441<pre>
442typedef <b>__attribute__((neon_vector_type(8)))</b> int8_t int8x8_t;
443typedef <b>__attribute__((neon_polyvector_type(16)))</b> poly8_t poly8x16_t;
444
445int8x8_t foo(int8x8_t a) {
446 int8x8_t v;
447 v = a;
448 return v;
449}
450</pre>
451</blockquote>
452
453<!-- ======================================================================= -->
454<h3><a name="vector_literals">Vector Literals</a></h3>
455<!-- ======================================================================= -->
456
457<p>Vector literals can be used to create vectors from a set of scalars, or
458vectors. Either parentheses or braces form can be used. In the parentheses form
459the number of literal values specified must be one, i.e. referring to a scalar
460value, or must match the size of the vector type being created. If a single
461scalar literal value is specified, the scalar literal value will be replicated
462to all the components of the vector type. In the brackets form any number of
463literals can be specified. For example:</p>
464
465<blockquote>
466<pre>
467typedef int v4si __attribute__((__vector_size__(16)));
468typedef float float4 __attribute__((ext_vector_type(4)));
469typedef float float2 __attribute__((ext_vector_type(2)));
470
471v4si vsi = (v4si){1, 2, 3, 4};
472float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
473vector int vi1 = (vector int)(1); // vi1 will be (1, 1, 1, 1).
474vector int vi2 = (vector int){1}; // vi2 will be (1, 0, 0, 0).
475vector int vi3 = (vector int)(1, 2); // error
476vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
477vector int vi5 = (vector int)(1, 2, 3, 4);
478float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));
479</pre>
480</blockquote>
481
482<!-- ======================================================================= -->
483<h3><a name="vector_operations">Vector Operations</a></h3>
484<!-- ======================================================================= -->
485
486<p>The table below shows the support for each operation by vector extension.
487A dash indicates that an operation is not accepted according to a corresponding
488specification.</p>
489
490<table width="500" border="1" cellspacing="0">
491 <tr>
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]492 <th>Operator</th>
493 <th>OpenCL</th>
494 <th>AltiVec</th>
495 <th>GCC</th>
496 <th>NEON</th>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]497 </tr>
498 <tr>
499 <td>[]</td>
500 <td align="center">yes</td>
501 <td align="center">yes</td>
502 <td align="center">yes</td>
503 <td align="center">-</td>
504 </tr>
505 <tr>
506 <td>unary operators +, -</td>
507 <td align="center">yes</td>
508 <td align="center">yes</td>
509 <td align="center">yes</td>
510 <td align="center">-</td>
511 </tr>
512 <tr>
513 <td>++, --</td>
514 <td align="center">yes</td>
515 <td align="center">yes</td>
516 <td align="center">-</td>
517 <td align="center">-</td>
518 </tr>
519 <tr>
520 <td>+, -, *, /, %</td>
521 <td align="center">yes</td>
522 <td align="center">yes</td>
523 <td align="center">yes</td>
524 <td align="center">-</td>
525 </tr>
526 <tr>
527 <td>bitwise operators &, |, ^, ~</td>
528 <td align="center">yes</td>
529 <td align="center">yes</td>
530 <td align="center">yes</td>
531 <td align="center">-</td>
532 </tr>
533 <tr>
534 <td>&gt&gt, &lt&lt</td>
535 <td align="center">yes</td>
536 <td align="center">yes</td>
537 <td align="center">yes</td>
538 <td align="center">-</td>
539 </tr>
540 <tr>
541 <td>!, &&,||</td>
542 <td align="center">no</td>
543 <td align="center">-</td>
544 <td align="center">-</td>
545 <td align="center">-</td>
546 </tr>
547 <tr>
548 <td>==,!=, >, <, >=, <=</td>
549 <td align="center">yes</td>
550 <td align="center">yes</td>
551 <td align="center">-</td>
552 <td align="center">-</td>
553 </tr>
554 <tr>
555 <td>=</td>
556 <td align="center">yes</td>
557 <td align="center">yes</td>
558 <td align="center">yes</td>
559 <td align="center">yes</td>
560 </tr>
561 <tr>
562 <td>:?</td>
563 <td align="center">yes</td>
564 <td align="center">-</td>
565 <td align="center">-</td>
566 <td align="center">-</td>
567 </tr>
568 <tr>
569 <td>sizeof</td>
570 <td align="center">yes</td>
571 <td align="center">yes</td>
572 <td align="center">yes</td>
573 <td align="center">yes</td>
574 </tr>
575</table>
576
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]577<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
578
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]579<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]580<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]581<!-- ======================================================================= -->
582
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]583<p>An optional string message can be added to the <tt>deprecated</tt>
584and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]585
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]586<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]587<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]588</blockquote>
589
590<p>If the deprecated or unavailable declaration is used, the message
591will be incorporated into the appropriate diagnostic:</p>
592
593<blockquote>
Benjamin Kramerb4556862012-03-19 19:12:30 +0000[diff] [blame]594<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!!
595 [-Wdeprecated-declarations]
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]596 explode();
597 ^</pre>
598</blockquote>
599
600<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]601with <tt>__has_extension(attribute_deprecated_with_message)</tt>
602and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]603
604<!-- ======================================================================= -->
605<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
606<!-- ======================================================================= -->
607
608<p>Clang allows attributes to be written on individual enumerators.
609This allows enumerators to be deprecated, made unavailable, etc. The
610attribute must appear after the enumerator name and before any
611initializer, like so:</p>
612
613<blockquote>
614<pre>enum OperationMode {
615 OM_Invalid,
616 OM_Normal,
617 OM_Terrified __attribute__((deprecated)),
618 OM_AbortOnError __attribute__((deprecated)) = 4
619};</pre>
620</blockquote>
621
622<p>Attributes on the <tt>enum</tt> declaration do not apply to
623individual enumerators.</p>
624
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]625<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]626
627<!-- ======================================================================= -->
Douglas Gregor93a70672012-03-11 04:53:21 +0000[diff] [blame]628<h2 id="availability">Availability attribute</h2
629<!-- ======================================================================= -->
630
631<p>Clang introduces the <code>availability</code> attribute, which can
632be placed on declarations to describe the lifecycle of that
633declaration relative to operating system versions. Consider the function declaration for a hypothetical function <code>f</code>:</p>
634
635<pre>
636void f(void) __attribute__((availability(macosx,introduced=10.4,deprecated=10.6,obsoleted=10.7)));
637</pre>
638
639<p>The availability attribute states that <code>f</code> was introduced in Mac OS X 10.4, deprecated in Mac OS X 10.6, and obsoleted in Mac OS X 10.7. This information is used by Clang to determine when it is safe to use <code>f</code>: for example, if Clang is instructed to compile code for Mac OS X 10.5, a call to <code>f()</code> succeeds. If Clang is instructed to compile code for Mac OS X 10.6, the call succeeds but Clang emits a warning specifying that the function is deprecated. Finally, if Clang is instructed to compile code for Mac OS X 10.7, the call fails because <code>f()</code> is no longer available.</p>
640
641<p>The availablility attribute is a comma-separated list starting with the platform name and then including clauses specifying important milestones in the declaration's lifetime (in any order) along with additional information. Those clauses can be:</p>
642
643<dl>
644 <dt>introduced=<i>version</i></dt>
645 <dd>The first version in which this declaration was introduced.</dd>
646
647 <dt>deprecated=<i>version</i></dt>
648 <dd>The first version in which this declaration was deprecated, meaning that users should migrate away from this API.</dd>
649
650 <dt>obsoleted=<i>version</i></dt>
651 <dd>The first version in which this declaration was obsoleted, meaning that it was removed completely and can no longer be used.</dd>
652
653 <dt>unavailable</dt>
654 <dd>This declaration is never available on this platform.</dd>
655
656 <dt>message=<i>string-literal</i></dt>
657 <dd>Additional message text that Clang will provide when emitting a warning or error about use of a deprecated or obsoleted declaration. Useful to direct users to replacement APIs.</dd>
658</dl>
659
660<p>Multiple availability attributes can be placed on a declaration, which may correspond to different platforms. Only the availability attribute with the platform corresponding to the target platform will be used; any others will be ignored. If no availability attribute specifies availability for the current target platform, the availability attributes are ignored. Supported platforms are:</p>
661
662<dl>
663 <dt>ios</dt>
664 <dd>Apple's iOS operating system. The minimum deployment target is specified by the <code>-mios-version-min=<i>version</i></code> or <code>-miphoneos-version-min=<i>version</i></code> command-line arguments.</dd>
665
666 <dt>macosx</dt>
667 <dd>Apple's Mac OS X operating system. The minimum deployment target is specified by the <code>-mmacosx-version-min=<i>version</i></code> command-line argument.</dd>
668</dl>
669
Douglas Gregor594f8412012-03-11 17:21:03 +0000[diff] [blame]670<p>A declaration can be used even when deploying back to a platform
671version prior to when the declaration was introduced. When this
672happens, the declaration is <a
673 href="https://developer.apple.com/library/mac/#documentation/MacOSX/Conceptual/BPFrameworks/Concepts/WeakLinking.html">weakly
674linked</a>, as if the <code>weak_import</code> attribute were added to the declaration. A weakly-linked declaration may or may not be present a run-time, and a program can determine whether the declaration is present by checking whether the address of that declaration is non-NULL.</p>
675
Douglas Gregor93a70672012-03-11 04:53:21 +0000[diff] [blame]676<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]677<h2 id="checking_language_features">Checks for Standard Language Features</h2>
678<!-- ======================================================================= -->
679
680<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
681enabled. Those features are listed here.</p>
682
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]683<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]684
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]685<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Sean Hunt647ba1b2011-06-23 00:42:53 +0000[diff] [blame]686example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]687
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]688<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]689
Ted Kremenek0eb95602009-12-03 02:06:43 +0000[diff] [blame]690<p>Use <tt>__has_feature(cxx_rtti)</tt> to determine if C++ RTTI has been enabled. For example,
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]691compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]692
693<!-- ======================================================================= -->
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]694<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
695<!-- ======================================================================= -->
696
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]697<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
698to query if certain upcoming standard language features are enabled. Those
699features are listed here. Features that are not yet implemented will be
700noted.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]701
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]702<h3 id="cxx0x">C++11</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]703
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]704<p>The features listed below are slated for inclusion in the upcoming
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]705C++11 standard. As a result, all these features are enabled
706with the <tt>-std=c++11</tt> option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]707
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]708<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]709
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]710<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 Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]711
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]712<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]713
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]714<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
715<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]716C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]717
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]718<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]719
720<p>Use <tt>__has_feature(cxx_alignas)</tt> or
721<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
722specifiers using <tt>alignas</tt> is enabled.</p>
723
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]724<h4 id="cxx_attributes">C++11 attributes</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]725
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]726<p>Use <tt>__has_feature(cxx_attributes)</tt> or
727<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]728parsing with C++11's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]729
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]730<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]731
732<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
733for generalized constant expressions (e.g., <tt>constexpr</tt>) is
Richard Smithb5216aa2012-02-14 22:56:17 +0000[diff] [blame]734enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]735
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]736<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]737
738<p>Use <tt>__has_feature(cxx_decltype)</tt> or
739<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
740<tt>decltype()</tt> specifier is enabled.</p>
741
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]742<h4 id="cxx_default_function_template_args">C++11 default template arguments in function templates</h4>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]743
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]744<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
745<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
746if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]747
Douglas Gregorf695a692011-11-01 01:19:34 +0000[diff] [blame]748<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
749
750<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
751<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
752defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
753
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]754<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Sean Huntd9624992011-06-23 06:11:37 +0000[diff] [blame]755
756<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
757support for delegating constructors is enabled.</p>
758
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]759<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]760
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]761<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
762<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]763deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]764
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]765<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]766<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
767
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]768<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]769
770<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
771support for generalized initializers (using braced lists and
Richard Smith88189552012-02-26 07:09:21 +0000[diff] [blame]772<tt>std::initializer_list</tt>) is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]773
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]774<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]775
Sebastian Redl72a81d22011-10-10 18:10:00 +0000[diff] [blame]776<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
777implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]778
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]779<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]780
781<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>
782
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]783<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]784
785<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
786<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
787inline namespaces is enabled.</p>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]788
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]789<h4 id="cxx_lambdas">C++11 lambdas</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]790
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]791<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
792<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregor46e021e2012-02-23 05:44:09 +0000[diff] [blame]793is enabled. </p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]794
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]795<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]796
797<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
798<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
799exception specifications is enabled.</p>
800
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]801<h4 id="cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]802
803<p>Use <tt>__has_feature(cxx_nonstatic_member_init)</tt> to determine whether in-class initialization of non-static data members is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]804
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]805<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]806
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]807<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
808<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000[diff] [blame]809<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]810
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]811<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]812
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]813<p>Use <tt>__has_feature(cxx_override_control)</tt> or
814<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]815the override control keywords is enabled.</p>
816
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]817<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]818<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
819<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
820if support for reference-qualified functions (e.g., member functions with
821<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
822is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]823
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]824<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]825
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]826<p>Use <tt>__has_feature(cxx_range_for)</tt> or
827<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
828range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]829
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]830<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Richard Smith80134582012-03-07 08:57:31 +0000[diff] [blame]831<p>Use <tt>__has_feature(cxx_raw_string_literals)</tt> to determine if support
832for raw string literals (e.g., <tt>R"x(foo\bar)x"</tt>) is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]833
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]834<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]835
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]836<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
837<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]838rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]839
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]840<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]841
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]842<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
843<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]844compile-time assertions using <tt>static_assert</tt> is enabled.</p>
845
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]846<h4 id="cxx_auto_type">C++11 type inference</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]847
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]848<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]849<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]850supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
851will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]852
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]853<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]854
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]855<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
856<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
857strongly typed, scoped enumerations is enabled.</p>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]858
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]859<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]860
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]861<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
862<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
863alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]864
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]865<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]866<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
867support for Unicode string literals is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]868
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]869<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]870
Richard Smithec92bc72012-03-03 23:51:05 +0000[diff] [blame]871<p>Use <tt>__has_feature(cxx_unrestricted_unions)</tt> to determine if support for unrestricted unions is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]872
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]873<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]874
Richard Smith9c1dda72012-03-09 08:41:27 +0000[diff] [blame]875<p>Use <tt>__has_feature(cxx_user_literals)</tt> to determine if support for user-defined literals is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]876
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]877<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]878
879<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
880<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
881for variadic templates is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]882
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]883<h3 id="c11">C11</h3>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]884
885<p>The features listed below are slated for inclusion in the upcoming
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]886C11 standard. As a result, all these features are enabled
887with the <tt>-std=c11</tt> option when compiling C code.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]888
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]889<h4 id="c_alignas">C11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]890
891<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
892to determine if support for alignment specifiers using <tt>_Alignas</tt>
893is enabled.</p>
894
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]895<h4 id="c_generic_selections">C11 generic selections</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]896
897<p>Use <tt>__has_feature(c_generic_selections)</tt> or
898<tt>__has_extension(c_generic_selections)</tt> to determine if support for
899generic selections is enabled.</p>
900
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]901<p>As an extension, the C11 generic selection expression is available in all
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]902languages supported by Clang. The syntax is the same as that given in the
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]903C11 standard.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]904
905<p>In C, type compatibility is decided according to the rules given in the
906appropriate standard, but in C++, which lacks the type compatibility rules
907used in C, types are considered compatible only if they are equivalent.</p>
908
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]909<h4 id="c_static_assert">C11 <tt>_Static_assert()</tt></h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]910
911<p>Use <tt>__has_feature(c_static_assert)</tt> or
912<tt>__has_extension(c_static_assert)</tt> to determine if support for
913compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
914
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]915<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]916<h2 id="checking_type_traits">Checks for Type Traits</h2>
917<!-- ======================================================================= -->
918
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]919<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 Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]920<blockquote>
921<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]922#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]923template&lt;typename From, typename To&gt;
924struct is_convertible_to {
925 static const bool value = __is_convertible_to(From, To);
926};
927#else
928// Emulate type trait
929#endif
930</pre>
931</blockquote>
932
933<p>The following type traits are supported by Clang:</p>
934<ul>
935 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
936 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
937 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
938 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
939 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
940 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
941 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
942 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
943 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
944 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
945 <li><code>__is_class</code> (GNU, Microsoft)</li>
946 <li><code>__is_convertible_to</code> (Microsoft)</li>
947 <li><code>__is_empty</code> (GNU, Microsoft)</li>
948 <li><code>__is_enum</code> (GNU, Microsoft)</li>
949 <li><code>__is_pod</code> (GNU, Microsoft)</li>
950 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
951 <li><code>__is_union</code> (GNU, Microsoft)</li>
952 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
Douglas Gregor5e9392b2011-12-03 18:14:24 +0000[diff] [blame]953 <li><code>__is_final</code>: Determines whether the given type is declared with a <code>final</code> class-virt-specifier.</li>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]954 <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 Gregor4ca8ac22012-02-24 07:38:34 +0000[diff] [blame]955 <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>
956 <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 Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]957</ul>
958
959<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]960<h2 id="blocks">Blocks</h2>
961<!-- ======================================================================= -->
962
Chris Lattnera7dbdf52009-03-09 07:03:22 +0000[diff] [blame]963<p>The syntax and high level language feature description is in <a
964href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
965details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +0000[diff] [blame]966href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]967
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]968
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]969<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]970
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]971<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]972<h2 id="objc_features">Objective-C Features</h2>
973<!-- ======================================================================= -->
974
975<h3 id="objc_instancetype">Related result types</h3>
976
977<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>
978
979<blockquote>
980<pre>
981@interface NSObject
982+ (id)alloc;
983- (id)init;
984@end
985
986@interface NSArray : NSObject
987@end
988</pre>
989</blockquote>
990
991<p>and this common initialization pattern</p>
992
993<blockquote>
994<pre>
995NSArray *array = [[NSArray alloc] init];
996</pre>
997</blockquote>
998
999<p>the type of the expression <code>[NSArray alloc]</code> is
1000<code>NSArray*</code> because <code>alloc</code> implicitly has a
1001related result type. Similarly, the type of the expression
1002<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
1003<code>init</code> has a related result type and its receiver is known
1004to 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>
1005
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]1006<p>A method with a related result type can be declared by using the
1007type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
1008is a contextual keyword that is only permitted in the result type of
1009an Objective-C method, e.g.</p>
1010
1011<pre>
1012@interface A
1013+ (<b>instancetype</b>)constructAnA;
1014@end
1015</pre>
1016
1017<p>The related result type can also be inferred for some methods.
1018To determine whether a method has an inferred related result type, the first
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1019word in the camel-case selector (e.g., "init" in "initWithObjects") is
Douglas Gregor8a0ace62011-11-03 18:33:01 +0000[diff] [blame]1020considered, and the method will have a related result type if its return
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1021type is compatible with the type of its class and if</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1022
1023<ul>
1024
1025 <li>the first word is "alloc" or "new", and the method is a class
1026 method, or</li>
1027
1028 <li>the first word is "autorelease", "init", "retain", or "self",
1029 and the method is an instance method.</li>
1030
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1031</ul>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1032
1033<p>If a method with a related result type is overridden by a subclass
1034method, the subclass method must also return a type that is compatible
1035with the subclass type. For example:</p>
1036
1037<blockquote>
1038<pre>
1039@interface NSString : NSObject
1040- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
1041@end
1042</pre>
1043</blockquote>
1044
1045<p>Related result types only affect the type of a message send or
1046property access via the given method. In all other respects, a method
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]1047with a related result type is treated the same way as method that
1048returns <tt>id</tt>.</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1049
Douglas Gregoraebb6532011-09-08 17:19:31 +0000[diff] [blame]1050<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
1051the <tt>instancetype</tt> contextual keyword is available.</p>
1052
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1053<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]1054<h2 id="objc_arc">Automatic reference counting </h2>
1055<!-- ======================================================================= -->
1056
1057<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>
1058
1059<!-- ======================================================================= -->
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1060<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
1061<!-- ======================================================================= -->
1062
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]1063<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1064underlying type within Objective-C. For example, one can write an
1065enumeration type as:</p>
1066
1067<pre>
1068typedef enum : unsigned char { Red, Green, Blue } Color;
1069</pre>
1070
1071<p>This specifies that the underlying type, which is used to store the
1072enumeration value, is <tt>unsigned char</tt>.</p>
1073
1074<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
1075support for fixed underlying types is available in Objective-C.</p>
1076
1077<!-- ======================================================================= -->
Douglas Gregor8a4e1822012-03-09 23:24:48 +0000[diff] [blame]1078<h2 id="objc_lambdas">Interoperability with C++11 lambdas</h2>
1079<!-- ======================================================================= -->
1080
1081<p>Clang provides interoperability between C++11 lambdas and
1082blocks-based APIs, by permitting a lambda to be implicitly converted
1083to a block pointer with the corresponding signature. For example,
1084consider an API such as <code>NSArray</code>'s array-sorting
1085method:</p>
1086
1087<pre> - (NSArray *)sortedArrayUsingComparator:(NSComparator)cmptr; </pre>
1088
1089<p><code>NSComparator</code> is simply a typedef for the block pointer
1090<code>NSComparisonResult (^)(id, id)</code>, and parameters of this
1091type are generally provided with block literals as arguments. However,
1092one can also use a C++11 lambda so long as it provides the same
1093signature (in this case, accepting two parameters of type
1094<code>id</code> and returning an <code>NSComparisonResult</code>):</p>
1095
1096<pre>
1097 NSArray *array = @[@"string 1", @"string 21", @"string 12", @"String 11",
1098 @"String 02"];
1099 const NSStringCompareOptions comparisonOptions
1100 = NSCaseInsensitiveSearch | NSNumericSearch |
1101 NSWidthInsensitiveSearch | NSForcedOrderingSearch;
1102 NSLocale *currentLocale = [NSLocale currentLocale];
1103 NSArray *sorted
1104 = [array sortedArrayUsingComparator:<b>[=](id s1, id s2) -&gt; NSComparisonResult {
1105 NSRange string1Range = NSMakeRange(0, [s1 length]);
1106 return [s1 compare:s2 options:comparisonOptions
1107 range:string1Range locale:currentLocale];
1108 }</b>];
1109 NSLog(@"sorted: %@", sorted);
1110</pre>
1111
1112<p>This code relies on an implicit conversion from the type of the
1113lambda expression (an unnamed, local class type called the <i>closure
1114type</i>) to the corresponding block pointer type. The conversion
1115itself is expressed by a conversion operator in that closure type
1116that produces a block pointer with the same signature as the lambda
1117itself, e.g.,</p>
1118
1119<pre>
1120 operator NSComparisonResult (^)(id, id)() const;
1121</pre>
1122
1123<p>This conversion function returns a new block that simply forwards
1124the two parameters to the lambda object (which it captures by copy),
1125then returns the result. The returned block is first copied (with
1126<tt>Block_copy</tt>) and then autoreleased. As an optimization, if a
1127lambda expression is immediately converted to a block pointer (as in
1128the first example, above), then the block is not copied and
1129autoreleased: rather, it is given the same lifetime as a block literal
1130written at that point in the program, which avoids the overhead of
1131copying a block to the heap in the common case.</p>
1132
Douglas Gregorbccda482012-03-10 22:20:11 +0000[diff] [blame]1133<p>The conversion from a lambda to a block pointer is only available
1134in Objective-C++, and not in C++ with blocks, due to its use of
1135Objective-C memory management (autorelease).</p>
1136
Douglas Gregor8a4e1822012-03-09 23:24:48 +0000[diff] [blame]1137<!-- ======================================================================= -->
Patrick Beard62f12342012-03-20 21:51:03 +0000[diff] [blame^]1138<h2 id="object-literals-subscripting">Object Literals and Subscripting</h2>
1139<!-- ======================================================================= -->
1140
1141<p>Clang provides support for <a href="ObjectiveCLiterals.html">Object Literals and Subscripting</a> in Objective-C, which simplifies common Objective-C programming patterns, make programs more concise, and improves the safety of container creation. There are several feature macros associated with object literals and subscripting: <code>__has_feature(objc_array_literals)</code> tests the availability of array literals; <code>__has_feature(objc_dictionary_literals)</code> tests the availability of dictionary literals; <code>__has_feature(objc_subscripting)</code> tests the availability of of object subscripting.</p>
1142
1143<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1144<h2 id="overloading-in-c">Function Overloading in C</h2>
1145<!-- ======================================================================= -->
1146
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1147<p>Clang provides support for C++ function overloading in C. Function
1148overloading in C is introduced using the <tt>overloadable</tt> attribute. For
1149example, one might provide several overloaded versions of a <tt>tgsin</tt>
1150function that invokes the appropriate standard function computing the sine of a
1151value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
1152precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1153
1154<blockquote>
1155<pre>
1156#include &lt;math.h&gt;
1157float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
1158double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
1159long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
1160</pre>
1161</blockquote>
1162
1163<p>Given these declarations, one can call <tt>tgsin</tt> with a
1164<tt>float</tt> value to receive a <tt>float</tt> result, with a
1165<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
1166overloading in C follows the rules of C++ function overloading to pick
1167the best overload given the call arguments, with a few C-specific
1168semantics:</p>
1169<ul>
1170 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
1171 double</tt> is ranked as a floating-point promotion (per C99) rather
1172 than as a floating-point conversion (as in C++).</li>
1173
1174 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
1175 <tt>U*</tt> is considered a pointer conversion (with conversion
1176 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
1177
1178 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
1179 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
1180 conversion is given "conversion" rank.</li>
1181</ul>
1182
1183<p>The declaration of <tt>overloadable</tt> functions is restricted to
1184function declarations and definitions. Most importantly, if any
1185function with a given name is given the <tt>overloadable</tt>
1186attribute, then all function declarations and definitions with that
1187name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1188attribute. This rule even applies to redeclarations of functions whose original
1189declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1190
1191<blockquote>
1192<pre>
1193int f(int) __attribute__((overloadable));
1194float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
1195
1196int g(int) __attribute__((overloadable));
1197int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
1198</pre>
1199</blockquote>
1200
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1201<p>Functions marked <tt>overloadable</tt> must have
1202prototypes. Therefore, the following code is ill-formed:</p>
1203
1204<blockquote>
1205<pre>
1206int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
1207</pre>
1208</blockquote>
1209
1210<p>However, <tt>overloadable</tt> functions are allowed to use a
1211ellipsis 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>
1212
1213<blockquote>
1214<pre>
Chris Lattner02246802009-02-18 22:27:46 +0000[diff] [blame]1215void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1216</pre>
1217</blockquote>
1218
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1219<p>Functions declared with the <tt>overloadable</tt> attribute have
1220their names mangled according to the same rules as C++ function
1221names. For example, the three <tt>tgsin</tt> functions in our
1222motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +0000[diff] [blame]1223<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1224caveats to this use of name mangling:</p>
1225
1226<ul>
1227
1228 <li>Future versions of Clang may change the name mangling of
1229 functions overloaded in C, so you should not depend on an specific
1230 mangling. To be completely safe, we strongly urge the use of
1231 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
1232
1233 <li>The <tt>overloadable</tt> attribute has almost no meaning when
1234 used in C++, because names will already be mangled and functions are
1235 already overloadable. However, when an <tt>overloadable</tt>
1236 function occurs within an <tt>extern "C"</tt> linkage specification,
1237 it's name <i>will</i> be mangled in the same way as it would in
1238 C.</li>
1239</ul>
1240
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]1241<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1242
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]1243<!-- ======================================================================= -->
1244<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
1245<!-- ======================================================================= -->
1246
1247<p>clang supports an extension which allows the following in C:</p>
1248
1249<blockquote>
1250<pre>
1251#include &lt;math.h&gt;
1252#include &lt;complex.h&gt;
1253complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
1254</pre>
1255</blockquote>
1256
1257<p>This construct is useful because there is no way to separately
1258initialize the real and imaginary parts of a complex variable in
1259standard C, given that clang does not support <code>_Imaginary</code>.
1260(clang also supports the <code>__real__</code> and <code>__imag__</code>
1261extensions from gcc, which help in some cases, but are not usable in
1262static initializers.)
1263
1264<p>Note that this extension does not allow eliding the braces; the
1265meaning of the following two lines is different:</p>
1266
1267<blockquote>
1268<pre>
1269complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
1270complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
1271</pre>
1272</blockquote>
1273
1274<p>This extension also works in C++ mode, as far as that goes, but does not
1275 apply to the C++ <code>std::complex</code>. (In C++11, list
1276 initialization allows the same syntax to be used with
1277 <code>std::complex</code> with the same meaning.)
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1278
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1279<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1280<h2 id="builtins">Builtin Functions</h2>
1281<!-- ======================================================================= -->
1282
1283<p>Clang supports a number of builtin library functions with the same syntax as
1284GCC, including things like <tt>__builtin_nan</tt>,
1285<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
1286<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
1287addition to the GCC builtins, Clang supports a number of builtins that GCC does
1288not, which are listed here.</p>
1289
1290<p>Please note that Clang does not and will not support all of the GCC builtins
1291for vector operations. Instead of using builtins, you should use the functions
1292defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
1293define portable wrappers for these. Many of the Clang versions of these
1294functions are implemented directly in terms of <a href="#vectors">extended
1295vector support</a> instead of builtins, in order to reduce the number of
1296builtins that we need to implement.</p>
1297
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1298<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1299<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1300<!-- ======================================================================= -->
1301
Chris Lattneraad826b2009-09-16 18:56:12 +0000[diff] [blame]1302<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1303permutation/shuffle/swizzle operations. This builtin is also very important for
1304the implementation of various target-specific header files like
1305<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1306</p>
1307
1308<p><b>Syntax:</b></p>
1309
1310<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1311__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1312</pre>
1313
1314<p><b>Examples:</b></p>
1315
1316<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1317 // Identity operation - return 4-element vector V1.
1318 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1319
1320 // "Splat" element 0 of V1 into a 4-element result.
1321 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1322
1323 // Reverse 4-element vector V1.
1324 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1325
1326 // Concatenate every other element of 4-element vectors V1 and V2.
1327 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1328
1329 // Concatenate every other element of 8-element vectors V1 and V2.
1330 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1331</pre>
1332
1333<p><b>Description:</b></p>
1334
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1335<p>The first two arguments to __builtin_shufflevector are vectors that have the
1336same element type. The remaining arguments are a list of integers that specify
1337the elements indices of the first two vectors that should be extracted and
1338returned in a new vector. These element indices are numbered sequentially
1339starting with the first vector, continuing into the second vector. Thus, if
1340vec1 is a 4-element vector, index 5 would refer to the second element of vec2.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1341</p>
1342
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1343<p>The result of __builtin_shufflevector is a vector
1344with the same element type as vec1/vec2 but that has an element count equal to
1345the number of indices specified.
1346</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1347
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1348<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1349
1350<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1351<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1352<!-- ======================================================================= -->
1353
1354<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1355the program cannot be reached, even if the compiler might otherwise think it
1356can. This is useful to improve optimization and eliminates certain warnings.
1357For example, without the <tt>__builtin_unreachable</tt> in the example below,
1358the compiler assumes that the inline asm can fall through and prints a "function
1359declared 'noreturn' should not return" warning.
1360</p>
1361
1362<p><b>Syntax:</b></p>
1363
1364<pre>
1365__builtin_unreachable()
1366</pre>
1367
1368<p><b>Example of Use:</b></p>
1369
1370<pre>
1371void myabort(void) __attribute__((noreturn));
1372void myabort(void) {
1373 asm("int3");
1374 __builtin_unreachable();
1375}
1376</pre>
1377
1378<p><b>Description:</b></p>
1379
1380<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1381it has undefined behavior, it is a statement that it is never reached and the
1382optimizer can take advantage of this to produce better code. This builtin takes
1383no arguments and produces a void result.
1384</p>
1385
1386<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1387
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1388<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1389<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1390<!-- ======================================================================= -->
1391
1392<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1393memory.
1394</p>
1395
1396<p><b>Syntax:</b></p>
1397
1398<pre>
1399<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1400</pre>
1401
1402<p><b>Example of Use:</b></p>
1403
1404<pre>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1405int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1406</pre>
1407
1408<p><b>Description:</b></p>
1409
1410<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1411intrinsics to allow code to atomically swap the current value with the new
1412value. More importantly, it helps developers write more efficient and correct
1413code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1414relying on the platform specific implementation details of
1415__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1416</p>
1417
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1418
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1419<!-- ======================================================================= -->
1420<h2 id="targetspecific">Target-Specific Extensions</h2>
1421<!-- ======================================================================= -->
1422
1423<p>Clang supports some language features conditionally on some targets.</p>
1424
1425<!-- ======================================================================= -->
1426<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1427<!-- ======================================================================= -->
1428
1429<p>The X86 backend has these language extensions:</p>
1430
1431<!-- ======================================================================= -->
1432<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1433<!-- ======================================================================= -->
1434
1435<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +0000[diff] [blame]1436relative to the X86 GS segment register, and address space #257 causes it to be
1437relative to the X86 FS segment. Note that this is a very very low-level
1438feature that should only be used if you know what you're doing (for example in
1439an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1440
1441<p>Here is an example:</p>
1442
1443<pre>
1444#define GS_RELATIVE __attribute__((address_space(256)))
1445int foo(int GS_RELATIVE *P) {
1446 return *P;
1447}
1448</pre>
1449
1450<p>Which compiles to (on X86-32):</p>
1451
1452<pre>
1453_foo:
1454 movl 4(%esp), %eax
1455 movl %gs:(%eax), %eax
1456 ret
1457</pre>
1458
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1459<!-- ======================================================================= -->
1460<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1461<!-- ======================================================================= -->
1462
1463<p>Clang supports additional attributes that are useful for documenting program
1464invariants and rules for static analysis tools. The extensions documented here
1465are used by the <a
1466href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1467engine</a> that is part of Clang's Analysis library.</p>
1468
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1469<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1470
1471<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +0000[diff] [blame]1472attribute. This attribute, which is typically affixed to a function prototype,
1473indicates that a call to a given function never returns. Function prototypes for
1474common functions like <tt>exit</tt> are typically annotated with this attribute,
1475as well as a variety of common assertion handlers. Users can educate the static
1476analyzer about their own custom assertion handles (thus cutting down on false
1477positives due to false paths) by marking their own &quot;panic&quot; functions
1478with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1479
1480<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +0000[diff] [blame]1481there are special functions that for all intents and purposes should be
1482considered panic functions (i.e., they are only called when an internal program
1483error occurs) but may actually return so that the program can fail gracefully.
1484The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1485as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]1486pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1487<tt>noreturn</tt>).</p>
1488
1489<p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]1490same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1491placed at the end of function prototypes:</p>
1492
1493<pre>
1494 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1495</pre>
1496
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1497<p>Query for this feature with
1498<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1499
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1500<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1501
1502<p>Many methods in Objective-C have conventional meanings determined
1503by their selectors. For the purposes of static analysis, it is
1504sometimes useful to be able to mark a method as having a particular
1505conventional meaning despite not having the right selector, or as not
1506having the conventional meaning that its selector would suggest.
1507For these use cases, we provide an attribute to specifically describe
1508the <q>method family</q> that a method belongs to.</p>
1509
1510<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1511where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1512<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1513can only be placed at the end of a method declaration:</p>
1514
1515<pre>
1516 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1517</pre>
1518
1519<p>Users who do not wish to change the conventional meaning of a
1520method, and who merely want to document its non-standard retain and
1521release semantics, should use the
1522<a href="#attr_retain_release">retaining behavior attributes</a>
1523described below.</p>
1524
1525<p>Query for this feature with
1526<tt>__has_attribute(objc_method_family)</tt>.</p>
1527
1528<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1529
1530<p>In Objective-C, functions and methods are generally assumed to take
1531and return objects with +0 retain counts, with some exceptions for
1532special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1533there are exceptions, and so Clang provides attributes to allow these
1534exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1535ignore non-leaks). Some exceptions may be better described using
1536the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1537attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1538
1539<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1540<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1541and <tt>cf_returns_not_retained</tt> attributes can be placed on
1542methods and functions that return Objective-C or CoreFoundation
1543objects. They are commonly placed at the end of a function prototype
1544or method declaration:</p>
1545
1546<pre>
1547 id foo() <b>__attribute__((ns_returns_retained))</b>;
1548
1549 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1550</pre>
1551
1552<p>The <tt>*_returns_retained</tt> attributes specify that the
1553returned object has a +1 retain count.
1554The <tt>*_returns_not_retained</tt> attributes specify that the return
1555object has a +0 retain count, even if the normal convention for its
1556selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1557returned object is +0, but is guaranteed to live at least as long as the
1558next flush of an autorelease pool.</p>
1559
1560<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1561attributes can be placed on an parameter declaration; they specify
1562that the argument is expected to have a +1 retain count, which will be
1563balanced in some way by the function or method.
1564The <tt>ns_consumes_self</tt> attribute can only be placed on an
1565Objective-C method; it specifies that the method expects
1566its <tt>self</tt> parameter to have a +1 retain count, which it will
1567balance in some way.</p>
1568
1569<pre>
1570 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1571
1572 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1573 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1574</pre>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1575
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1576<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1577<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1578
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]1579<!-- ======================================================================= -->
1580<h2 id="dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</h2>
1581<!-- ======================================================================= -->
1582<h3 id="address_sanitizer">AddressSanitizer</h3>
1583<p> Use <code>__has_feature(address_sanitizer)</code>
1584to check if the code is being built with <a
1585 href="AddressSanitizer.html">AddressSanitizer</a>.
1586</p>
Kostya Serebryany71efba02012-01-24 19:25:38 +0000[diff] [blame]1587<p>Use <tt>__attribute__((no_address_safety_analysis))</tt> on a function
1588declaration to specify that address safety instrumentation (e.g.
1589AddressSanitizer) should not be applied to that function.
1590</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1591
1592<!-- ======================================================================= -->
Caitlin Sadowski73cbbc82011-07-28 18:38:36 +0000[diff] [blame]1593<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1594<!-- ======================================================================= -->
1595
1596<p>Clang supports additional attributes for checking basic locking policies in
1597multithreaded programs.
1598Clang currently parses the following list of attributes, although
1599<b>the implementation for these annotations is currently in development.</b>
1600For more details, see the
1601<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1602</p>
1603
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1604<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1605
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1606<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1607declaration to specify that the thread safety analysis should not be run on that
1608function. This attribute provides an escape hatch (e.g. for situations when it
1609is difficult to annotate the locking policy). </p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1610
1611<h4 id="ts_lockable">lockable</h4>
1612
1613<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1614that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1615used to check consistency.</p>
1616
1617<h4 id="ts_scopedlockable">scoped_lockable</h4>
1618
1619<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1620specify that it has a "scoped" lockable type. Objects of this type will acquire
1621the lock upon construction and release it upon going out of scope.
1622 This annotation is primarily used to check
1623consistency.</p>
1624
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1625<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1626
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1627<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1628specify that the variable must be accessed while holding some lock.</p>
1629
1630<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1631
1632<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1633specify that the pointer must be dereferenced while holding some lock.</p>
1634
1635<h4 id="ts_guardedby">guarded_by(l)</h4>
1636
1637<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1638specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1639
1640<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1641
1642<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1643specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1644
1645<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1646
1647<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1648of a lockable variable to specify that the lock must be acquired before all
1649attribute arguments. Arguments must be lockable type, and there must be at
1650least one argument.</p>
1651
1652<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1653
1654<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1655of a lockable variable to specify that the lock must be acquired after all
1656attribute arguments. Arguments must be lockable type, and there must be at
1657least one argument.</p>
1658
1659<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1660
1661<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1662declaration to specify that the function acquires all listed locks
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1663exclusively. This attribute takes zero or more arguments: either of lockable
1664type or integers indexing into function parameters of lockable type. If no
1665arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1666enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1667
1668<h4 id="ts_slf">shared_lock_function(...)</h4>
1669
1670<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1671declaration to specify that the function acquires all listed locks, although
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1672 the locks may be shared (e.g. read locks). This attribute takes zero or more
1673arguments: either of lockable type or integers indexing into function
1674parameters of lockable type. If no arguments are given, the acquired lock is
1675implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1676
1677<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1678
1679<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1680declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1681all listed locks exclusively. This attribute takes one or more arguments. The
1682first argument is an integer or boolean value specifying the return value of a
1683successful lock acquisition. The remaining arugments are either of lockable type
1684or integers indexing into function parameters of lockable type. If only one
1685argument is given, the acquired lock is implicitly <tt>this</tt> of the
1686enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1687
1688<h4 id="ts_stf">shared_trylock_function(...)</h4>
1689
1690<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1691declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1692all listed locks, although the locks may be shared (e.g. read locks). This
1693attribute takes one or more arguments. The first argument is an integer or
1694boolean value specifying the return value of a successful lock acquisition. The
1695remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1696function parameters of lockable type. If only one argument is given, the
1697acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1698
1699<h4 id="ts_uf">unlock_function(...)</h4>
1700
1701<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1702declaration to specify that the function release all listed locks. This
1703attribute takes zero or more arguments: either of lockable type or integers
1704indexing into function parameters of lockable type. If no arguments are given,
1705the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1706
1707<h4 id="ts_lr">lock_returned(l)</h4>
1708
1709<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1710declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1711must be of lockable type). This annotation is used to aid in resolving lock
1712expressions.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1713
1714<h4 id="ts_le">locks_excluded(...)</h4>
1715
1716<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1717to specify that the function must not be called with the listed locks. Arguments
1718must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1719
1720<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1721
1722<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1723declaration to specify that the function must be called while holding the listed
1724exclusive locks. Arguments must be lockable type, and there must be at
1725least one argument.</p>
1726
1727<h4 id="ts_slr">shared_locks_required(...)</h4>
1728
1729<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1730declaration to specify that the function must be called while holding the listed
1731shared locks. Arguments must be lockable type, and there must be at
1732least one argument.</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1733
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1734</div>
1735</body>
1736</html>