Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 93a7067302d670ad771ba02ea3ff757dae43c853 / . / docs / LanguageExtensions.html
blob: bdb04ac1a5aad6c126635b4893c8d0cf61045dca [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>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]92 </ul>
93</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]94<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]95<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]96<li><a href="#builtins">Builtin Functions</a>
97 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]98 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]99 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]100 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]101 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]102</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]103<li><a href="#targetspecific">Target-Specific Extensions</a>
104 <ul>
105 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
106 </ul>
107</li>
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]108<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]109<li><a href="#dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</a>
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]110 <ul>
111 <li><a href="#address_sanitizer">AddressSanitizer</a></li>
112 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]113</li>
114<li><a href="#threadsafety">Thread Safety Annotation Checking</a>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]115 <ul>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]116 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
117 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
118 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]119 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
120 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]121 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
122 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
123 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
124 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
125 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
126 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
127 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
128 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
129 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
130 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
131 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
132 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
133 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]134 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]135</li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]136</ul>
137
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]138<!-- ======================================================================= -->
139<h2 id="intro">Introduction</h2>
140<!-- ======================================================================= -->
141
142<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]143addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]144range of GCC extensions. Please see the <a
145href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
146more information on these extensions.</p>
147
148<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]149<h2 id="feature_check">Feature Checking Macros</h2>
150<!-- ======================================================================= -->
151
152<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]153on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]154function-like macros. This allows you to directly test for a feature in your
155code without having to resort to something like autoconf or fragile "compiler
156version checks".</p>
157
158<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]159<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]160<!-- ======================================================================= -->
161
162<p>This function-like macro takes a single identifier argument that is the name
163of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
164not. It can be used like this:</p>
165
166<blockquote>
167<pre>
168#ifndef __has_builtin // Optional of course.
169 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
170#endif
171
172...
173#if __has_builtin(__builtin_trap)
174 __builtin_trap();
175#else
176 abort();
177#endif
178...
179</pre>
180</blockquote>
181
182
183<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]184<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]185<!-- ======================================================================= -->
186
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]187<p>These function-like macros take a single identifier argument that is the
188name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
189is both supported by Clang and standardized in the current language standard
190or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
191<code>__has_extension</code> evaluates to 1 if the feature is supported by
192Clang in the current language (either as a language extension or a standard
193language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]194
195<blockquote>
196<pre>
197#ifndef __has_feature // Optional of course.
198 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
199#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]200#ifndef __has_extension
201 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
202#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]203
204...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]205#if __has_feature(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]206// This code will only be compiled with the -std=c++11 and -std=gnu++11
207// options, because rvalue references are only standardized in C++11.
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]208#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]209
210#if __has_extension(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]211// 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]212// and -std=gnu++98 options, because rvalue references are supported as a
213// language extension in C++98.
214#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]215</pre>
216</blockquote>
217
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]218<p id="has_feature_back_compat">For backwards compatibility reasons,
219<code>__has_feature</code> can also be used to test for support for
220non-standardized features, i.e. features not prefixed <code>c_</code>,
221<code>cxx_</code> or <code>objc_</code>.</p>
222
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]223<p id="has_feature_for_non_language_features">
224Another use of <code>__has_feature</code> is to check for compiler features
225not related to the language standard, such as e.g.
226<a href="AddressSanitizer.html">AddressSanitizer</a>.
227
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]228<p>If the <code>-pedantic-errors</code> option is given,
229<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
230
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]231<p>The feature tag is described along with the language feature below.</p>
232
Richard Smith5297d712012-02-25 10:41:10 +0000[diff] [blame]233<p>The feature name or extension name can also be specified with a preceding and
234following <code>__</code> (double underscore) to avoid interference from a macro
Richard Smith1d9f4c12012-03-01 02:12:07 +0000[diff] [blame]235with the same name. For instance, <code>__cxx_rvalue_references__</code> can be
236used instead of <code>cxx_rvalue_references</code>.</p>
Richard Smith5297d712012-02-25 10:41:10 +0000[diff] [blame]237
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]238<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]239<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]240<!-- ======================================================================= -->
241
242<p>This function-like macro takes a single identifier argument that is the name
243of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
244can be used like this:</p>
245
246<blockquote>
247<pre>
248#ifndef __has_attribute // Optional of course.
249 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
250#endif
251
252...
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]253#if __has_attribute(always_inline)
254#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]255#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]256#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]257#endif
258...
259</pre>
260</blockquote>
261
Jean-Daniel Dupas8a5e7fd2012-03-01 14:53:16 +0000[diff] [blame]262<p>The attribute name can also be specified with a preceding and
263following <code>__</code> (double underscore) to avoid interference from a macro
264with the same name. For instance, <code>__always_inline__</code> can be used
265instead of <code>always_inline</code>.</p>
266
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]267<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]268<h2 id="has_include">Include File Checking Macros</h2>
269<!-- ======================================================================= -->
270
271<p>Not all developments systems have the same include files.
272The <a href="#__has_include">__has_include</a> and
273<a href="#__has_include_next">__has_include_next</a> macros allow you to
274check for the existence of an include file before doing
275a possibly failing #include directive.</p>
276
277<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]278<h3><a name="__has_include">__has_include</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]279<!-- ======================================================================= -->
280
281<p>This function-like macro takes a single file name string argument that
282is the name of an include file. It evaluates to 1 if the file can
283be found using the include paths, or 0 otherwise:</p>
284
285<blockquote>
286<pre>
287// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]288#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]289# include "myinclude.h"
290#endif
291
292// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]293#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]294# include "myinclude.h"
295#endif
296</pre>
297</blockquote>
298
299<p>To test for this feature, use #if defined(__has_include).</p>
300
301<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]302<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]303<!-- ======================================================================= -->
304
305<p>This function-like macro takes a single file name string argument that
306is the name of an include file. It is like __has_include except that it
307looks for the second instance of the given file found in the include
308paths. It evaluates to 1 if the second instance of the file can
309be found using the include paths, or 0 otherwise:</p>
310
311<blockquote>
312<pre>
313// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]314#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]315# include_next "myinclude.h"
316#endif
317
318// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]319#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]320# include_next "myinclude.h"
321#endif
322</pre>
323</blockquote>
324
325<p>Note that __has_include_next, like the GNU extension
326#include_next directive, is intended for use in headers only,
327and will issue a warning if used in the top-level compilation
328file. A warning will also be issued if an absolute path
329is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]330
Ted Kremenekd7681502011-10-12 19:46:30 +0000[diff] [blame]331
332<!-- ======================================================================= -->
333<h3><a name="__has_warning">__has_warning</a></h3>
334<!-- ======================================================================= -->
335
336<p>This function-like macro takes a string literal that represents a command
337 line option for a warning and returns true if that is a valid warning
338 option.</p>
339
340<blockquote>
341<pre>
342#if __has_warning("-Wformat")
343...
344#endif
345</pre>
346</blockquote>
347
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]348<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]349<h2 id="builtinmacros">Builtin Macros</h2>
350<!-- ======================================================================= -->
351
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]352<dl>
353 <dt><code>__BASE_FILE__</code></dt>
354 <dd>Defined to a string that contains the name of the main input
355 file passed to Clang.</dd>
356
357 <dt><code>__COUNTER__</code></dt>
358 <dd>Defined to an integer value that starts at zero and is
359 incremented each time the <code>__COUNTER__</code> macro is
360 expanded.</dd>
361
362 <dt><code>__INCLUDE_LEVEL__</code></dt>
363 <dd>Defined to an integral value that is the include depth of the
364 file currently being translated. For the main file, this value is
365 zero.</dd>
366
367 <dt><code>__TIMESTAMP__</code></dt>
368 <dd>Defined to the date and time of the last modification of the
369 current source file.</dd>
370
371 <dt><code>__clang__</code></dt>
372 <dd>Defined when compiling with Clang</dd>
373
374 <dt><code>__clang_major__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]375 <dd>Defined to the major marketing version number of Clang (e.g., the
376 2 in 2.0.1). Note that marketing version numbers should not be used to
377 check for language features, as different vendors use different numbering
378 schemes. Instead, use the <a href="#feature_check">feature checking
379 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]380
381 <dt><code>__clang_minor__</code></dt>
382 <dd>Defined to the minor version number of Clang (e.g., the 0 in
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]383 2.0.1). Note that marketing version numbers should not be used to
384 check for language features, as different vendors use different numbering
385 schemes. Instead, use the <a href="#feature_check">feature checking
386 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]387
388 <dt><code>__clang_patchlevel__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]389 <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]390
391 <dt><code>__clang_version__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]392 <dd>Defined to a string that captures the Clang marketing version, including
393 the Subversion tag or revision number, e.g., "1.5 (trunk 102332)".</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]394</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]395
396<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]397<h2 id="vectors">Vectors and Extended Vectors</h2>
398<!-- ======================================================================= -->
399
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]400<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]401
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]402<p>OpenCL vector types are created using <tt>ext_vector_type</tt> attribute. It
403support for <tt>V.xyzw</tt> syntax and other tidbits as seen in OpenCL. An
404example is:</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]405
406<blockquote>
407<pre>
408typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
409typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
410
411float4 foo(float2 a, float2 b) {
412 float4 c;
413 c.xz = a;
414 c.yw = b;
415 return c;
416}
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]417</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]418</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]419
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]420<p>Query for this feature with
421<tt>__has_extension(attribute_ext_vector_type)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]422
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]423<p>Giving <tt>-faltivec</tt> option to clang enables support for AltiVec vector
424syntax and functions. For example:</p>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]425
426<blockquote>
427<pre>
428vector float foo(vector int a) {
429 vector int b;
430 b = vec_add(a, a) + a;
431 return (vector float)b;
432}
433</pre>
434</blockquote>
435
436<p>NEON vector types are created using <tt>neon_vector_type</tt> and
437<tt>neon_polyvector_type</tt> attributes. For example:</p>
438
439<blockquote>
440<pre>
441typedef <b>__attribute__((neon_vector_type(8)))</b> int8_t int8x8_t;
442typedef <b>__attribute__((neon_polyvector_type(16)))</b> poly8_t poly8x16_t;
443
444int8x8_t foo(int8x8_t a) {
445 int8x8_t v;
446 v = a;
447 return v;
448}
449</pre>
450</blockquote>
451
452<!-- ======================================================================= -->
453<h3><a name="vector_literals">Vector Literals</a></h3>
454<!-- ======================================================================= -->
455
456<p>Vector literals can be used to create vectors from a set of scalars, or
457vectors. Either parentheses or braces form can be used. In the parentheses form
458the number of literal values specified must be one, i.e. referring to a scalar
459value, or must match the size of the vector type being created. If a single
460scalar literal value is specified, the scalar literal value will be replicated
461to all the components of the vector type. In the brackets form any number of
462literals can be specified. For example:</p>
463
464<blockquote>
465<pre>
466typedef int v4si __attribute__((__vector_size__(16)));
467typedef float float4 __attribute__((ext_vector_type(4)));
468typedef float float2 __attribute__((ext_vector_type(2)));
469
470v4si vsi = (v4si){1, 2, 3, 4};
471float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
472vector int vi1 = (vector int)(1); // vi1 will be (1, 1, 1, 1).
473vector int vi2 = (vector int){1}; // vi2 will be (1, 0, 0, 0).
474vector int vi3 = (vector int)(1, 2); // error
475vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
476vector int vi5 = (vector int)(1, 2, 3, 4);
477float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));
478</pre>
479</blockquote>
480
481<!-- ======================================================================= -->
482<h3><a name="vector_operations">Vector Operations</a></h3>
483<!-- ======================================================================= -->
484
485<p>The table below shows the support for each operation by vector extension.
486A dash indicates that an operation is not accepted according to a corresponding
487specification.</p>
488
489<table width="500" border="1" cellspacing="0">
490 <tr>
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]491 <th>Operator</th>
492 <th>OpenCL</th>
493 <th>AltiVec</th>
494 <th>GCC</th>
495 <th>NEON</th>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]496 </tr>
497 <tr>
498 <td>[]</td>
499 <td align="center">yes</td>
500 <td align="center">yes</td>
501 <td align="center">yes</td>
502 <td align="center">-</td>
503 </tr>
504 <tr>
505 <td>unary operators +, -</td>
506 <td align="center">yes</td>
507 <td align="center">yes</td>
508 <td align="center">yes</td>
509 <td align="center">-</td>
510 </tr>
511 <tr>
512 <td>++, --</td>
513 <td align="center">yes</td>
514 <td align="center">yes</td>
515 <td align="center">-</td>
516 <td align="center">-</td>
517 </tr>
518 <tr>
519 <td>+, -, *, /, %</td>
520 <td align="center">yes</td>
521 <td align="center">yes</td>
522 <td align="center">yes</td>
523 <td align="center">-</td>
524 </tr>
525 <tr>
526 <td>bitwise operators &, |, ^, ~</td>
527 <td align="center">yes</td>
528 <td align="center">yes</td>
529 <td align="center">yes</td>
530 <td align="center">-</td>
531 </tr>
532 <tr>
533 <td>&gt&gt, &lt&lt</td>
534 <td align="center">yes</td>
535 <td align="center">yes</td>
536 <td align="center">yes</td>
537 <td align="center">-</td>
538 </tr>
539 <tr>
540 <td>!, &&,||</td>
541 <td align="center">no</td>
542 <td align="center">-</td>
543 <td align="center">-</td>
544 <td align="center">-</td>
545 </tr>
546 <tr>
547 <td>==,!=, >, <, >=, <=</td>
548 <td align="center">yes</td>
549 <td align="center">yes</td>
550 <td align="center">-</td>
551 <td align="center">-</td>
552 </tr>
553 <tr>
554 <td>=</td>
555 <td align="center">yes</td>
556 <td align="center">yes</td>
557 <td align="center">yes</td>
558 <td align="center">yes</td>
559 </tr>
560 <tr>
561 <td>:?</td>
562 <td align="center">yes</td>
563 <td align="center">-</td>
564 <td align="center">-</td>
565 <td align="center">-</td>
566 </tr>
567 <tr>
568 <td>sizeof</td>
569 <td align="center">yes</td>
570 <td align="center">yes</td>
571 <td align="center">yes</td>
572 <td align="center">yes</td>
573 </tr>
574</table>
575
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]576<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
577
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]578<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]579<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]580<!-- ======================================================================= -->
581
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]582<p>An optional string message can be added to the <tt>deprecated</tt>
583and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]584
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]585<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]586<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]587</blockquote>
588
589<p>If the deprecated or unavailable declaration is used, the message
590will be incorporated into the appropriate diagnostic:</p>
591
592<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]593<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!! [-Wdeprecated-declarations]
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]594 explode();
595 ^</pre>
596</blockquote>
597
598<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]599with <tt>__has_extension(attribute_deprecated_with_message)</tt>
600and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]601
602<!-- ======================================================================= -->
603<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
604<!-- ======================================================================= -->
605
606<p>Clang allows attributes to be written on individual enumerators.
607This allows enumerators to be deprecated, made unavailable, etc. The
608attribute must appear after the enumerator name and before any
609initializer, like so:</p>
610
611<blockquote>
612<pre>enum OperationMode {
613 OM_Invalid,
614 OM_Normal,
615 OM_Terrified __attribute__((deprecated)),
616 OM_AbortOnError __attribute__((deprecated)) = 4
617};</pre>
618</blockquote>
619
620<p>Attributes on the <tt>enum</tt> declaration do not apply to
621individual enumerators.</p>
622
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]623<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]624
625<!-- ======================================================================= -->
Douglas Gregor93a70672012-03-11 04:53:21 +0000[diff] [blame^]626<h2 id="availability">Availability attribute</h2
627<!-- ======================================================================= -->
628
629<p>Clang introduces the <code>availability</code> attribute, which can
630be placed on declarations to describe the lifecycle of that
631declaration relative to operating system versions. Consider the function declaration for a hypothetical function <code>f</code>:</p>
632
633<pre>
634void f(void) __attribute__((availability(macosx,introduced=10.4,deprecated=10.6,obsoleted=10.7)));
635</pre>
636
637<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>
638
639<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>
640
641<dl>
642 <dt>introduced=<i>version</i></dt>
643 <dd>The first version in which this declaration was introduced.</dd>
644
645 <dt>deprecated=<i>version</i></dt>
646 <dd>The first version in which this declaration was deprecated, meaning that users should migrate away from this API.</dd>
647
648 <dt>obsoleted=<i>version</i></dt>
649 <dd>The first version in which this declaration was obsoleted, meaning that it was removed completely and can no longer be used.</dd>
650
651 <dt>unavailable</dt>
652 <dd>This declaration is never available on this platform.</dd>
653
654 <dt>message=<i>string-literal</i></dt>
655 <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>
656</dl>
657
658<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>
659
660<dl>
661 <dt>ios</dt>
662 <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>
663
664 <dt>macosx</dt>
665 <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>
666</dl>
667
668<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]669<h2 id="checking_language_features">Checks for Standard Language Features</h2>
670<!-- ======================================================================= -->
671
672<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
673enabled. Those features are listed here.</p>
674
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]675<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]676
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]677<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]678example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]679
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]680<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]681
Ted Kremenek0eb95602009-12-03 02:06:43 +0000[diff] [blame]682<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]683compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]684
685<!-- ======================================================================= -->
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]686<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
687<!-- ======================================================================= -->
688
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]689<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
690to query if certain upcoming standard language features are enabled. Those
691features are listed here. Features that are not yet implemented will be
692noted.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]693
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]694<h3 id="cxx0x">C++11</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]695
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]696<p>The features listed below are slated for inclusion in the upcoming
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]697C++11 standard. As a result, all these features are enabled
698with the <tt>-std=c++11</tt> option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]699
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]700<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]701
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]702<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]703
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]704<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]705
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]706<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
707<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]708C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]709
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]710<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]711
712<p>Use <tt>__has_feature(cxx_alignas)</tt> or
713<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
714specifiers using <tt>alignas</tt> is enabled.</p>
715
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]716<h4 id="cxx_attributes">C++11 attributes</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]717
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]718<p>Use <tt>__has_feature(cxx_attributes)</tt> or
719<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]720parsing with C++11's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]721
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]722<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]723
724<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
725for generalized constant expressions (e.g., <tt>constexpr</tt>) is
Richard Smithb5216aa2012-02-14 22:56:17 +0000[diff] [blame]726enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]727
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]728<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]729
730<p>Use <tt>__has_feature(cxx_decltype)</tt> or
731<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
732<tt>decltype()</tt> specifier is enabled.</p>
733
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]734<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]735
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]736<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
737<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
738if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]739
Douglas Gregorf695a692011-11-01 01:19:34 +0000[diff] [blame]740<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
741
742<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
743<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
744defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
745
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]746<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Sean Huntd9624992011-06-23 06:11:37 +0000[diff] [blame]747
748<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
749support for delegating constructors is enabled.</p>
750
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]751<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]752
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]753<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
754<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]755deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]756
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]757<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]758<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
759
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]760<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]761
762<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
763support for generalized initializers (using braced lists and
Richard Smith88189552012-02-26 07:09:21 +0000[diff] [blame]764<tt>std::initializer_list</tt>) is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]765
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]766<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]767
Sebastian Redl72a81d22011-10-10 18:10:00 +0000[diff] [blame]768<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
769implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]770
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]771<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]772
773<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>
774
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]775<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]776
777<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
778<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
779inline namespaces is enabled.</p>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]780
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]781<h4 id="cxx_lambdas">C++11 lambdas</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]782
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]783<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
784<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregor46e021e2012-02-23 05:44:09 +0000[diff] [blame]785is enabled. </p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]786
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]787<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]788
789<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
790<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
791exception specifications is enabled.</p>
792
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]793<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]794
795<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]796
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]797<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]798
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]799<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
800<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000[diff] [blame]801<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]802
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]803<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]804
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]805<p>Use <tt>__has_feature(cxx_override_control)</tt> or
806<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]807the override control keywords is enabled.</p>
808
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]809<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]810<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
811<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
812if support for reference-qualified functions (e.g., member functions with
813<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
814is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]815
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]816<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]817
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]818<p>Use <tt>__has_feature(cxx_range_for)</tt> or
819<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
820range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]821
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]822<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Richard Smith80134582012-03-07 08:57:31 +0000[diff] [blame]823<p>Use <tt>__has_feature(cxx_raw_string_literals)</tt> to determine if support
824for 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]825
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]826<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]827
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]828<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
829<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]830rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]831
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]832<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]833
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]834<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
835<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]836compile-time assertions using <tt>static_assert</tt> is enabled.</p>
837
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]838<h4 id="cxx_auto_type">C++11 type inference</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]839
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]840<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]841<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]842supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
843will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]844
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]845<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]846
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]847<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
848<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
849strongly typed, scoped enumerations is enabled.</p>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]850
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]851<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]852
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]853<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
854<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
855alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]856
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]857<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]858<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
859support for Unicode string literals is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]860
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]861<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]862
Richard Smithec92bc72012-03-03 23:51:05 +0000[diff] [blame]863<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]864
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]865<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]866
Richard Smith9c1dda72012-03-09 08:41:27 +0000[diff] [blame]867<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]868
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]869<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]870
871<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
872<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
873for variadic templates is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]874
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]875<h3 id="c11">C11</h3>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]876
877<p>The features listed below are slated for inclusion in the upcoming
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]878C11 standard. As a result, all these features are enabled
879with the <tt>-std=c11</tt> option when compiling C code.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]880
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]881<h4 id="c_alignas">C11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]882
883<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
884to determine if support for alignment specifiers using <tt>_Alignas</tt>
885is enabled.</p>
886
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]887<h4 id="c_generic_selections">C11 generic selections</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]888
889<p>Use <tt>__has_feature(c_generic_selections)</tt> or
890<tt>__has_extension(c_generic_selections)</tt> to determine if support for
891generic selections is enabled.</p>
892
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]893<p>As an extension, the C11 generic selection expression is available in all
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]894languages supported by Clang. The syntax is the same as that given in the
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]895C11 standard.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]896
897<p>In C, type compatibility is decided according to the rules given in the
898appropriate standard, but in C++, which lacks the type compatibility rules
899used in C, types are considered compatible only if they are equivalent.</p>
900
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]901<h4 id="c_static_assert">C11 <tt>_Static_assert()</tt></h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]902
903<p>Use <tt>__has_feature(c_static_assert)</tt> or
904<tt>__has_extension(c_static_assert)</tt> to determine if support for
905compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
906
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]907<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]908<h2 id="checking_type_traits">Checks for Type Traits</h2>
909<!-- ======================================================================= -->
910
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]911<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]912<blockquote>
913<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]914#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]915template&lt;typename From, typename To&gt;
916struct is_convertible_to {
917 static const bool value = __is_convertible_to(From, To);
918};
919#else
920// Emulate type trait
921#endif
922</pre>
923</blockquote>
924
925<p>The following type traits are supported by Clang:</p>
926<ul>
927 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
928 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
929 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
930 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
931 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
932 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
933 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
934 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
935 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
936 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
937 <li><code>__is_class</code> (GNU, Microsoft)</li>
938 <li><code>__is_convertible_to</code> (Microsoft)</li>
939 <li><code>__is_empty</code> (GNU, Microsoft)</li>
940 <li><code>__is_enum</code> (GNU, Microsoft)</li>
941 <li><code>__is_pod</code> (GNU, Microsoft)</li>
942 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
943 <li><code>__is_union</code> (GNU, Microsoft)</li>
944 <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]945 <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]946 <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]947 <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>
948 <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]949</ul>
950
951<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]952<h2 id="blocks">Blocks</h2>
953<!-- ======================================================================= -->
954
Chris Lattnera7dbdf52009-03-09 07:03:22 +0000[diff] [blame]955<p>The syntax and high level language feature description is in <a
956href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
957details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +0000[diff] [blame]958href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]959
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]960
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]961<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]962
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]963<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]964<h2 id="objc_features">Objective-C Features</h2>
965<!-- ======================================================================= -->
966
967<h3 id="objc_instancetype">Related result types</h3>
968
969<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>
970
971<blockquote>
972<pre>
973@interface NSObject
974+ (id)alloc;
975- (id)init;
976@end
977
978@interface NSArray : NSObject
979@end
980</pre>
981</blockquote>
982
983<p>and this common initialization pattern</p>
984
985<blockquote>
986<pre>
987NSArray *array = [[NSArray alloc] init];
988</pre>
989</blockquote>
990
991<p>the type of the expression <code>[NSArray alloc]</code> is
992<code>NSArray*</code> because <code>alloc</code> implicitly has a
993related result type. Similarly, the type of the expression
994<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
995<code>init</code> has a related result type and its receiver is known
996to 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>
997
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]998<p>A method with a related result type can be declared by using the
999type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
1000is a contextual keyword that is only permitted in the result type of
1001an Objective-C method, e.g.</p>
1002
1003<pre>
1004@interface A
1005+ (<b>instancetype</b>)constructAnA;
1006@end
1007</pre>
1008
1009<p>The related result type can also be inferred for some methods.
1010To determine whether a method has an inferred related result type, the first
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1011word in the camel-case selector (e.g., "init" in "initWithObjects") is
Douglas Gregor8a0ace62011-11-03 18:33:01 +0000[diff] [blame]1012considered, and the method will have a related result type if its return
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1013type is compatible with the type of its class and if</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1014
1015<ul>
1016
1017 <li>the first word is "alloc" or "new", and the method is a class
1018 method, or</li>
1019
1020 <li>the first word is "autorelease", "init", "retain", or "self",
1021 and the method is an instance method.</li>
1022
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1023</ul>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1024
1025<p>If a method with a related result type is overridden by a subclass
1026method, the subclass method must also return a type that is compatible
1027with the subclass type. For example:</p>
1028
1029<blockquote>
1030<pre>
1031@interface NSString : NSObject
1032- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
1033@end
1034</pre>
1035</blockquote>
1036
1037<p>Related result types only affect the type of a message send or
1038property access via the given method. In all other respects, a method
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]1039with a related result type is treated the same way as method that
1040returns <tt>id</tt>.</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1041
Douglas Gregoraebb6532011-09-08 17:19:31 +0000[diff] [blame]1042<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
1043the <tt>instancetype</tt> contextual keyword is available.</p>
1044
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]1045<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]1046<h2 id="objc_arc">Automatic reference counting </h2>
1047<!-- ======================================================================= -->
1048
1049<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>
1050
1051<!-- ======================================================================= -->
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1052<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
1053<!-- ======================================================================= -->
1054
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]1055<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1056underlying type within Objective-C. For example, one can write an
1057enumeration type as:</p>
1058
1059<pre>
1060typedef enum : unsigned char { Red, Green, Blue } Color;
1061</pre>
1062
1063<p>This specifies that the underlying type, which is used to store the
1064enumeration value, is <tt>unsigned char</tt>.</p>
1065
1066<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
1067support for fixed underlying types is available in Objective-C.</p>
1068
1069<!-- ======================================================================= -->
Douglas Gregor8a4e1822012-03-09 23:24:48 +0000[diff] [blame]1070<h2 id="objc_lambdas">Interoperability with C++11 lambdas</h2>
1071<!-- ======================================================================= -->
1072
1073<p>Clang provides interoperability between C++11 lambdas and
1074blocks-based APIs, by permitting a lambda to be implicitly converted
1075to a block pointer with the corresponding signature. For example,
1076consider an API such as <code>NSArray</code>'s array-sorting
1077method:</p>
1078
1079<pre> - (NSArray *)sortedArrayUsingComparator:(NSComparator)cmptr; </pre>
1080
1081<p><code>NSComparator</code> is simply a typedef for the block pointer
1082<code>NSComparisonResult (^)(id, id)</code>, and parameters of this
1083type are generally provided with block literals as arguments. However,
1084one can also use a C++11 lambda so long as it provides the same
1085signature (in this case, accepting two parameters of type
1086<code>id</code> and returning an <code>NSComparisonResult</code>):</p>
1087
1088<pre>
1089 NSArray *array = @[@"string 1", @"string 21", @"string 12", @"String 11",
1090 @"String 02"];
1091 const NSStringCompareOptions comparisonOptions
1092 = NSCaseInsensitiveSearch | NSNumericSearch |
1093 NSWidthInsensitiveSearch | NSForcedOrderingSearch;
1094 NSLocale *currentLocale = [NSLocale currentLocale];
1095 NSArray *sorted
1096 = [array sortedArrayUsingComparator:<b>[=](id s1, id s2) -&gt; NSComparisonResult {
1097 NSRange string1Range = NSMakeRange(0, [s1 length]);
1098 return [s1 compare:s2 options:comparisonOptions
1099 range:string1Range locale:currentLocale];
1100 }</b>];
1101 NSLog(@"sorted: %@", sorted);
1102</pre>
1103
1104<p>This code relies on an implicit conversion from the type of the
1105lambda expression (an unnamed, local class type called the <i>closure
1106type</i>) to the corresponding block pointer type. The conversion
1107itself is expressed by a conversion operator in that closure type
1108that produces a block pointer with the same signature as the lambda
1109itself, e.g.,</p>
1110
1111<pre>
1112 operator NSComparisonResult (^)(id, id)() const;
1113</pre>
1114
1115<p>This conversion function returns a new block that simply forwards
1116the two parameters to the lambda object (which it captures by copy),
1117then returns the result. The returned block is first copied (with
1118<tt>Block_copy</tt>) and then autoreleased. As an optimization, if a
1119lambda expression is immediately converted to a block pointer (as in
1120the first example, above), then the block is not copied and
1121autoreleased: rather, it is given the same lifetime as a block literal
1122written at that point in the program, which avoids the overhead of
1123copying a block to the heap in the common case.</p>
1124
Douglas Gregorbccda482012-03-10 22:20:11 +0000[diff] [blame]1125<p>The conversion from a lambda to a block pointer is only available
1126in Objective-C++, and not in C++ with blocks, due to its use of
1127Objective-C memory management (autorelease).</p>
1128
Douglas Gregor8a4e1822012-03-09 23:24:48 +0000[diff] [blame]1129<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1130<h2 id="overloading-in-c">Function Overloading in C</h2>
1131<!-- ======================================================================= -->
1132
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1133<p>Clang provides support for C++ function overloading in C. Function
1134overloading in C is introduced using the <tt>overloadable</tt> attribute. For
1135example, one might provide several overloaded versions of a <tt>tgsin</tt>
1136function that invokes the appropriate standard function computing the sine of a
1137value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
1138precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1139
1140<blockquote>
1141<pre>
1142#include &lt;math.h&gt;
1143float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
1144double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
1145long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
1146</pre>
1147</blockquote>
1148
1149<p>Given these declarations, one can call <tt>tgsin</tt> with a
1150<tt>float</tt> value to receive a <tt>float</tt> result, with a
1151<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
1152overloading in C follows the rules of C++ function overloading to pick
1153the best overload given the call arguments, with a few C-specific
1154semantics:</p>
1155<ul>
1156 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
1157 double</tt> is ranked as a floating-point promotion (per C99) rather
1158 than as a floating-point conversion (as in C++).</li>
1159
1160 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
1161 <tt>U*</tt> is considered a pointer conversion (with conversion
1162 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
1163
1164 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
1165 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
1166 conversion is given "conversion" rank.</li>
1167</ul>
1168
1169<p>The declaration of <tt>overloadable</tt> functions is restricted to
1170function declarations and definitions. Most importantly, if any
1171function with a given name is given the <tt>overloadable</tt>
1172attribute, then all function declarations and definitions with that
1173name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1174attribute. This rule even applies to redeclarations of functions whose original
1175declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1176
1177<blockquote>
1178<pre>
1179int f(int) __attribute__((overloadable));
1180float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
1181
1182int g(int) __attribute__((overloadable));
1183int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
1184</pre>
1185</blockquote>
1186
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1187<p>Functions marked <tt>overloadable</tt> must have
1188prototypes. Therefore, the following code is ill-formed:</p>
1189
1190<blockquote>
1191<pre>
1192int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
1193</pre>
1194</blockquote>
1195
1196<p>However, <tt>overloadable</tt> functions are allowed to use a
1197ellipsis 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>
1198
1199<blockquote>
1200<pre>
Chris Lattner02246802009-02-18 22:27:46 +0000[diff] [blame]1201void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1202</pre>
1203</blockquote>
1204
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1205<p>Functions declared with the <tt>overloadable</tt> attribute have
1206their names mangled according to the same rules as C++ function
1207names. For example, the three <tt>tgsin</tt> functions in our
1208motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +0000[diff] [blame]1209<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1210caveats to this use of name mangling:</p>
1211
1212<ul>
1213
1214 <li>Future versions of Clang may change the name mangling of
1215 functions overloaded in C, so you should not depend on an specific
1216 mangling. To be completely safe, we strongly urge the use of
1217 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
1218
1219 <li>The <tt>overloadable</tt> attribute has almost no meaning when
1220 used in C++, because names will already be mangled and functions are
1221 already overloadable. However, when an <tt>overloadable</tt>
1222 function occurs within an <tt>extern "C"</tt> linkage specification,
1223 it's name <i>will</i> be mangled in the same way as it would in
1224 C.</li>
1225</ul>
1226
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]1227<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1228
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]1229<!-- ======================================================================= -->
1230<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
1231<!-- ======================================================================= -->
1232
1233<p>clang supports an extension which allows the following in C:</p>
1234
1235<blockquote>
1236<pre>
1237#include &lt;math.h&gt;
1238#include &lt;complex.h&gt;
1239complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
1240</pre>
1241</blockquote>
1242
1243<p>This construct is useful because there is no way to separately
1244initialize the real and imaginary parts of a complex variable in
1245standard C, given that clang does not support <code>_Imaginary</code>.
1246(clang also supports the <code>__real__</code> and <code>__imag__</code>
1247extensions from gcc, which help in some cases, but are not usable in
1248static initializers.)
1249
1250<p>Note that this extension does not allow eliding the braces; the
1251meaning of the following two lines is different:</p>
1252
1253<blockquote>
1254<pre>
1255complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
1256complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
1257</pre>
1258</blockquote>
1259
1260<p>This extension also works in C++ mode, as far as that goes, but does not
1261 apply to the C++ <code>std::complex</code>. (In C++11, list
1262 initialization allows the same syntax to be used with
1263 <code>std::complex</code> with the same meaning.)
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1264
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1265<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1266<h2 id="builtins">Builtin Functions</h2>
1267<!-- ======================================================================= -->
1268
1269<p>Clang supports a number of builtin library functions with the same syntax as
1270GCC, including things like <tt>__builtin_nan</tt>,
1271<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
1272<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
1273addition to the GCC builtins, Clang supports a number of builtins that GCC does
1274not, which are listed here.</p>
1275
1276<p>Please note that Clang does not and will not support all of the GCC builtins
1277for vector operations. Instead of using builtins, you should use the functions
1278defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
1279define portable wrappers for these. Many of the Clang versions of these
1280functions are implemented directly in terms of <a href="#vectors">extended
1281vector support</a> instead of builtins, in order to reduce the number of
1282builtins that we need to implement.</p>
1283
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1284<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1285<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1286<!-- ======================================================================= -->
1287
Chris Lattneraad826b2009-09-16 18:56:12 +0000[diff] [blame]1288<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1289permutation/shuffle/swizzle operations. This builtin is also very important for
1290the implementation of various target-specific header files like
1291<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1292</p>
1293
1294<p><b>Syntax:</b></p>
1295
1296<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1297__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1298</pre>
1299
1300<p><b>Examples:</b></p>
1301
1302<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1303 // Identity operation - return 4-element vector V1.
1304 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1305
1306 // "Splat" element 0 of V1 into a 4-element result.
1307 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1308
1309 // Reverse 4-element vector V1.
1310 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1311
1312 // Concatenate every other element of 4-element vectors V1 and V2.
1313 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1314
1315 // Concatenate every other element of 8-element vectors V1 and V2.
1316 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1317</pre>
1318
1319<p><b>Description:</b></p>
1320
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1321<p>The first two arguments to __builtin_shufflevector are vectors that have the
1322same element type. The remaining arguments are a list of integers that specify
1323the elements indices of the first two vectors that should be extracted and
1324returned in a new vector. These element indices are numbered sequentially
1325starting with the first vector, continuing into the second vector. Thus, if
1326vec1 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]1327</p>
1328
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1329<p>The result of __builtin_shufflevector is a vector
1330with the same element type as vec1/vec2 but that has an element count equal to
1331the number of indices specified.
1332</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1333
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1334<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1335
1336<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1337<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1338<!-- ======================================================================= -->
1339
1340<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1341the program cannot be reached, even if the compiler might otherwise think it
1342can. This is useful to improve optimization and eliminates certain warnings.
1343For example, without the <tt>__builtin_unreachable</tt> in the example below,
1344the compiler assumes that the inline asm can fall through and prints a "function
1345declared 'noreturn' should not return" warning.
1346</p>
1347
1348<p><b>Syntax:</b></p>
1349
1350<pre>
1351__builtin_unreachable()
1352</pre>
1353
1354<p><b>Example of Use:</b></p>
1355
1356<pre>
1357void myabort(void) __attribute__((noreturn));
1358void myabort(void) {
1359 asm("int3");
1360 __builtin_unreachable();
1361}
1362</pre>
1363
1364<p><b>Description:</b></p>
1365
1366<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1367it has undefined behavior, it is a statement that it is never reached and the
1368optimizer can take advantage of this to produce better code. This builtin takes
1369no arguments and produces a void result.
1370</p>
1371
1372<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1373
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1374<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1375<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1376<!-- ======================================================================= -->
1377
1378<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1379memory.
1380</p>
1381
1382<p><b>Syntax:</b></p>
1383
1384<pre>
1385<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1386</pre>
1387
1388<p><b>Example of Use:</b></p>
1389
1390<pre>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1391int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1392</pre>
1393
1394<p><b>Description:</b></p>
1395
1396<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1397intrinsics to allow code to atomically swap the current value with the new
1398value. More importantly, it helps developers write more efficient and correct
1399code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1400relying on the platform specific implementation details of
1401__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1402</p>
1403
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1404
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1405<!-- ======================================================================= -->
1406<h2 id="targetspecific">Target-Specific Extensions</h2>
1407<!-- ======================================================================= -->
1408
1409<p>Clang supports some language features conditionally on some targets.</p>
1410
1411<!-- ======================================================================= -->
1412<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1413<!-- ======================================================================= -->
1414
1415<p>The X86 backend has these language extensions:</p>
1416
1417<!-- ======================================================================= -->
1418<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1419<!-- ======================================================================= -->
1420
1421<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +0000[diff] [blame]1422relative to the X86 GS segment register, and address space #257 causes it to be
1423relative to the X86 FS segment. Note that this is a very very low-level
1424feature that should only be used if you know what you're doing (for example in
1425an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1426
1427<p>Here is an example:</p>
1428
1429<pre>
1430#define GS_RELATIVE __attribute__((address_space(256)))
1431int foo(int GS_RELATIVE *P) {
1432 return *P;
1433}
1434</pre>
1435
1436<p>Which compiles to (on X86-32):</p>
1437
1438<pre>
1439_foo:
1440 movl 4(%esp), %eax
1441 movl %gs:(%eax), %eax
1442 ret
1443</pre>
1444
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1445<!-- ======================================================================= -->
1446<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1447<!-- ======================================================================= -->
1448
1449<p>Clang supports additional attributes that are useful for documenting program
1450invariants and rules for static analysis tools. The extensions documented here
1451are used by the <a
1452href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1453engine</a> that is part of Clang's Analysis library.</p>
1454
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1455<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1456
1457<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +0000[diff] [blame]1458attribute. This attribute, which is typically affixed to a function prototype,
1459indicates that a call to a given function never returns. Function prototypes for
1460common functions like <tt>exit</tt> are typically annotated with this attribute,
1461as well as a variety of common assertion handlers. Users can educate the static
1462analyzer about their own custom assertion handles (thus cutting down on false
1463positives due to false paths) by marking their own &quot;panic&quot; functions
1464with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1465
1466<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +0000[diff] [blame]1467there are special functions that for all intents and purposes should be
1468considered panic functions (i.e., they are only called when an internal program
1469error occurs) but may actually return so that the program can fail gracefully.
1470The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1471as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]1472pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1473<tt>noreturn</tt>).</p>
1474
1475<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]1476same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1477placed at the end of function prototypes:</p>
1478
1479<pre>
1480 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1481</pre>
1482
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1483<p>Query for this feature with
1484<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1485
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1486<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1487
1488<p>Many methods in Objective-C have conventional meanings determined
1489by their selectors. For the purposes of static analysis, it is
1490sometimes useful to be able to mark a method as having a particular
1491conventional meaning despite not having the right selector, or as not
1492having the conventional meaning that its selector would suggest.
1493For these use cases, we provide an attribute to specifically describe
1494the <q>method family</q> that a method belongs to.</p>
1495
1496<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1497where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1498<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1499can only be placed at the end of a method declaration:</p>
1500
1501<pre>
1502 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1503</pre>
1504
1505<p>Users who do not wish to change the conventional meaning of a
1506method, and who merely want to document its non-standard retain and
1507release semantics, should use the
1508<a href="#attr_retain_release">retaining behavior attributes</a>
1509described below.</p>
1510
1511<p>Query for this feature with
1512<tt>__has_attribute(objc_method_family)</tt>.</p>
1513
1514<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1515
1516<p>In Objective-C, functions and methods are generally assumed to take
1517and return objects with +0 retain counts, with some exceptions for
1518special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1519there are exceptions, and so Clang provides attributes to allow these
1520exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1521ignore non-leaks). Some exceptions may be better described using
1522the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1523attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1524
1525<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1526<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1527and <tt>cf_returns_not_retained</tt> attributes can be placed on
1528methods and functions that return Objective-C or CoreFoundation
1529objects. They are commonly placed at the end of a function prototype
1530or method declaration:</p>
1531
1532<pre>
1533 id foo() <b>__attribute__((ns_returns_retained))</b>;
1534
1535 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1536</pre>
1537
1538<p>The <tt>*_returns_retained</tt> attributes specify that the
1539returned object has a +1 retain count.
1540The <tt>*_returns_not_retained</tt> attributes specify that the return
1541object has a +0 retain count, even if the normal convention for its
1542selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1543returned object is +0, but is guaranteed to live at least as long as the
1544next flush of an autorelease pool.</p>
1545
1546<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1547attributes can be placed on an parameter declaration; they specify
1548that the argument is expected to have a +1 retain count, which will be
1549balanced in some way by the function or method.
1550The <tt>ns_consumes_self</tt> attribute can only be placed on an
1551Objective-C method; it specifies that the method expects
1552its <tt>self</tt> parameter to have a +1 retain count, which it will
1553balance in some way.</p>
1554
1555<pre>
1556 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1557
1558 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1559 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1560</pre>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1561
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1562<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1563<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1564
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]1565<!-- ======================================================================= -->
1566<h2 id="dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</h2>
1567<!-- ======================================================================= -->
1568<h3 id="address_sanitizer">AddressSanitizer</h3>
1569<p> Use <code>__has_feature(address_sanitizer)</code>
1570to check if the code is being built with <a
1571 href="AddressSanitizer.html">AddressSanitizer</a>.
1572</p>
Kostya Serebryany71efba02012-01-24 19:25:38 +0000[diff] [blame]1573<p>Use <tt>__attribute__((no_address_safety_analysis))</tt> on a function
1574declaration to specify that address safety instrumentation (e.g.
1575AddressSanitizer) should not be applied to that function.
1576</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1577
1578<!-- ======================================================================= -->
Caitlin Sadowski73cbbc82011-07-28 18:38:36 +0000[diff] [blame]1579<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1580<!-- ======================================================================= -->
1581
1582<p>Clang supports additional attributes for checking basic locking policies in
1583multithreaded programs.
1584Clang currently parses the following list of attributes, although
1585<b>the implementation for these annotations is currently in development.</b>
1586For more details, see the
1587<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1588</p>
1589
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1590<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1591
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1592<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1593declaration to specify that the thread safety analysis should not be run on that
1594function. This attribute provides an escape hatch (e.g. for situations when it
1595is difficult to annotate the locking policy). </p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1596
1597<h4 id="ts_lockable">lockable</h4>
1598
1599<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1600that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1601used to check consistency.</p>
1602
1603<h4 id="ts_scopedlockable">scoped_lockable</h4>
1604
1605<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1606specify that it has a "scoped" lockable type. Objects of this type will acquire
1607the lock upon construction and release it upon going out of scope.
1608 This annotation is primarily used to check
1609consistency.</p>
1610
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1611<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1612
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1613<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1614specify that the variable must be accessed while holding some lock.</p>
1615
1616<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1617
1618<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1619specify that the pointer must be dereferenced while holding some lock.</p>
1620
1621<h4 id="ts_guardedby">guarded_by(l)</h4>
1622
1623<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1624specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1625
1626<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1627
1628<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1629specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1630
1631<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1632
1633<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1634of a lockable variable to specify that the lock must be acquired before all
1635attribute arguments. Arguments must be lockable type, and there must be at
1636least one argument.</p>
1637
1638<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1639
1640<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1641of a lockable variable to specify that the lock must be acquired after all
1642attribute arguments. Arguments must be lockable type, and there must be at
1643least one argument.</p>
1644
1645<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1646
1647<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1648declaration to specify that the function acquires all listed locks
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1649exclusively. This attribute takes zero or more arguments: either of lockable
1650type or integers indexing into function parameters of lockable type. If no
1651arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1652enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1653
1654<h4 id="ts_slf">shared_lock_function(...)</h4>
1655
1656<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1657declaration to specify that the function acquires all listed locks, although
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1658 the locks may be shared (e.g. read locks). This attribute takes zero or more
1659arguments: either of lockable type or integers indexing into function
1660parameters of lockable type. If no arguments are given, the acquired lock is
1661implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1662
1663<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1664
1665<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1666declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1667all listed locks exclusively. This attribute takes one or more arguments. The
1668first argument is an integer or boolean value specifying the return value of a
1669successful lock acquisition. The remaining arugments are either of lockable type
1670or integers indexing into function parameters of lockable type. If only one
1671argument is given, the acquired lock is implicitly <tt>this</tt> of the
1672enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1673
1674<h4 id="ts_stf">shared_trylock_function(...)</h4>
1675
1676<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1677declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1678all listed locks, although the locks may be shared (e.g. read locks). This
1679attribute takes one or more arguments. The first argument is an integer or
1680boolean value specifying the return value of a successful lock acquisition. The
1681remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1682function parameters of lockable type. If only one argument is given, the
1683acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1684
1685<h4 id="ts_uf">unlock_function(...)</h4>
1686
1687<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1688declaration to specify that the function release all listed locks. This
1689attribute takes zero or more arguments: either of lockable type or integers
1690indexing into function parameters of lockable type. If no arguments are given,
1691the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1692
1693<h4 id="ts_lr">lock_returned(l)</h4>
1694
1695<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1696declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1697must be of lockable type). This annotation is used to aid in resolving lock
1698expressions.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1699
1700<h4 id="ts_le">locks_excluded(...)</h4>
1701
1702<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1703to specify that the function must not be called with the listed locks. Arguments
1704must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1705
1706<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1707
1708<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1709declaration to specify that the function must be called while holding the listed
1710exclusive locks. Arguments must be lockable type, and there must be at
1711least one argument.</p>
1712
1713<h4 id="ts_slr">shared_locks_required(...)</h4>
1714
1715<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1716declaration to specify that the function must be called while holding the listed
1717shared locks. Arguments must be lockable type, and there must be at
1718least one argument.</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1719
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1720</div>
1721</body>
1722</html>