blob: 92db81f2bbafa0dfa0f1637353397888cebc36d1 [file] [log] [blame]
Sean Hunt7e98b472011-06-23 01:21:01 +00001<!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 +00004<html>
5<head>
Sean Hunt7e98b472011-06-23 01:21:01 +00006 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Eli Friedman0c706c22011-09-19 23:17:44 +00007 <title>Clang Language Extensions</title>
Sean Hunt64f857b2011-06-23 01:22:53 +00008 <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 +000010 <style type="text/css">
11 td {
12 vertical-align: top;
13 }
Benjamin Kramer3419d7c2012-01-15 16:42:14 +000014 th { background-color: #ffddaa; }
Sean Hunt7e98b472011-06-23 01:21:01 +000015 </style>
Chris Lattner5ce933f2009-02-09 08:46:11 +000016</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 +000027<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompson92bd8c72009-11-02 22:28:12 +000028<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner81edc9f2009-04-13 02:45:46 +000029<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000030<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCall48209082010-11-08 19:48:17 +000031<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>
Daniel Dunbar85ff9692012-04-05 17:10:06 +000033<li><a href="#user_specified_system_framework">'User-Specified' System Frameworks</a></li>
Douglas Gregor93a70672012-03-11 04:53:21 +000034<li><a href="#availability">Availability attribute</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +000035<li><a href="#checking_language_features">Checks for Standard Language Features</a>
Ted Kremenek22c34102009-12-03 02:05:57 +000036 <ul>
Richard Smithfafbf062012-04-11 17:55:32 +000037 <li><a href="#cxx98">C++98</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +000038 <ul>
Richard Smithfafbf062012-04-11 17:55:32 +000039 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
40 <li><a href="#cxx_rtti">C++ RTTI</a></li>
41 </ul></li>
42 <li><a href="#cxx11">C++11</a>
43 <ul>
44 <li><a href="#cxx_access_control_sfinae">C++11 SFINAE includes access control</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +000045 <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 +000051 <li><a href="#cxx_defaulted_functions">C++11 defaulted functions</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +000052 <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>
Douglas Gregor7b156dd2012-04-04 00:48:39 +000060 <li><a href="#cxx_local_type_template_args">C++11 local and unnamed types as template arguments</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +000061 <li><a href="#cxx_noexcept">C++11 noexcept specification</a></li>
62 <li><a href="#cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</a></li>
63 <li><a href="#cxx_nullptr">C++11 nullptr</a></li>
64 <li><a href="#cxx_override_control">C++11 override control</a></li>
65 <li><a href="#cxx_range_for">C++11 range-based for loop</a></li>
66 <li><a href="#cxx_raw_string_literals">C++11 raw string literals</a></li>
67 <li><a href="#cxx_rvalue_references">C++11 rvalue references</a></li>
68 <li><a href="#cxx_reference_qualified_functions">C++11 reference-qualified functions</a></li>
69 <li><a href="#cxx_static_assert">C++11 <tt>static_assert()</tt></a></li>
70 <li><a href="#cxx_auto_type">C++11 type inference</a></li>
Richard Smithfafbf062012-04-11 17:55:32 +000071 <li><a href="#cxx_strong_enums">C++11 strongly-typed enumerations</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +000072 <li><a href="#cxx_trailing_return">C++11 trailing return type</a></li>
73 <li><a href="#cxx_unicode_literals">C++11 Unicode string literals</a></li>
74 <li><a href="#cxx_unrestricted_unions">C++11 unrestricted unions</a></li>
75 <li><a href="#cxx_user_literals">C++11 user-defined literals</a></li>
76 <li><a href="#cxx_variadic_templates">C++11 variadic templates</a></li>
Richard Smithfafbf062012-04-11 17:55:32 +000077 </ul></li>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +000078 <li><a href="#c11">C11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +000079 <ul>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +000080 <li><a href="#c_alignas">C11 alignment specifiers</a></li>
Richard Smithfafbf062012-04-11 17:55:32 +000081 <li><a href="#c_atomic">C11 atomic operations</a></li>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +000082 <li><a href="#c_generic_selections">C11 generic selections</a></li>
83 <li><a href="#c_static_assert">C11 <tt>_Static_assert()</tt></a></li>
Richard Smithfafbf062012-04-11 17:55:32 +000084 </ul></li>
85</ul></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +000086<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000087<li><a href="#blocks">Blocks</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +000088<li><a href="#objc_features">Objective-C Features</a>
89 <ul>
90 <li><a href="#objc_instancetype">Related result types</a></li>
John McCallf85e1932011-06-15 23:02:42 +000091 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor5471bc82011-09-08 17:18:35 +000092 <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li>
Douglas Gregor8a4e1822012-03-09 23:24:48 +000093 <li><a href="#objc_lambdas">Interoperability with C++11 lambdas</a></li>
Patrick Beardeb382ec2012-04-19 00:25:12 +000094 <li><a href="#objc_object_literals_subscripting">Object Literals and Subscripting</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +000095 </ul>
96</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +000097<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman0c706c22011-09-19 23:17:44 +000098<li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000099<li><a href="#builtins">Builtin Functions</a>
100 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000101 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +0000102 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000103 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000104 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000105</li>
Richard Smithe0d3b4c2012-05-03 18:27:39 +0000106<li><a href="#non-standard-attributes">Non-standard C++11 Attributes</a>
107<ul>
108 <li><a href="#clang__fallthrough">The <tt>clang::fallthrough</tt> attribute</a></li>
109</ul>
110</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000111<li><a href="#targetspecific">Target-Specific Extensions</a>
112 <ul>
113 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
114 </ul>
115</li>
John McCall87494012011-03-18 03:51:49 +0000116<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000117<li><a href="#dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</a>
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000118 <ul>
119 <li><a href="#address_sanitizer">AddressSanitizer</a></li>
120 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000121</li>
122<li><a href="#threadsafety">Thread Safety Annotation Checking</a>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000123 <ul>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000124 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
125 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
126 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000127 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
128 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000129 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
130 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
131 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
132 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
133 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
134 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
135 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
136 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
137 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
138 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
139 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
140 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
141 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000142 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000143</li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000144</ul>
145
Chris Lattner5ce933f2009-02-09 08:46:11 +0000146<!-- ======================================================================= -->
147<h2 id="intro">Introduction</h2>
148<!-- ======================================================================= -->
149
150<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000151addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000152range of GCC extensions. Please see the <a
153href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
154more information on these extensions.</p>
155
156<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000157<h2 id="feature_check">Feature Checking Macros</h2>
158<!-- ======================================================================= -->
159
160<p>Language extensions can be very useful, but only if you know you can depend
Chris Lattnerc70e1932011-03-21 16:25:11 +0000161on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000162function-like macros. This allows you to directly test for a feature in your
163code without having to resort to something like autoconf or fragile "compiler
164version checks".</p>
165
166<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000167<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000168<!-- ======================================================================= -->
169
170<p>This function-like macro takes a single identifier argument that is the name
171of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
172not. It can be used like this:</p>
173
174<blockquote>
175<pre>
176#ifndef __has_builtin // Optional of course.
177 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
178#endif
179
180...
181#if __has_builtin(__builtin_trap)
182 __builtin_trap();
183#else
184 abort();
185#endif
186...
187</pre>
188</blockquote>
189
190
191<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000192<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000193<!-- ======================================================================= -->
194
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000195<p>These function-like macros take a single identifier argument that is the
196name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
197is both supported by Clang and standardized in the current language standard
198or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
199<code>__has_extension</code> evaluates to 1 if the feature is supported by
200Clang in the current language (either as a language extension or a standard
201language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000202
203<blockquote>
204<pre>
205#ifndef __has_feature // Optional of course.
206 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
207#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000208#ifndef __has_extension
209 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
210#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000211
212...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000213#if __has_feature(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000214// This code will only be compiled with the -std=c++11 and -std=gnu++11
215// options, because rvalue references are only standardized in C++11.
Chris Lattner148772a2009-06-13 07:13:28 +0000216#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000217
218#if __has_extension(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000219// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000220// and -std=gnu++98 options, because rvalue references are supported as a
221// language extension in C++98.
222#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000223</pre>
224</blockquote>
225
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000226<p id="has_feature_back_compat">For backwards compatibility reasons,
227<code>__has_feature</code> can also be used to test for support for
228non-standardized features, i.e. features not prefixed <code>c_</code>,
229<code>cxx_</code> or <code>objc_</code>.</p>
230
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000231<p id="has_feature_for_non_language_features">
232Another use of <code>__has_feature</code> is to check for compiler features
233not related to the language standard, such as e.g.
234<a href="AddressSanitizer.html">AddressSanitizer</a>.
235
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000236<p>If the <code>-pedantic-errors</code> option is given,
237<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
238
Chris Lattner148772a2009-06-13 07:13:28 +0000239<p>The feature tag is described along with the language feature below.</p>
240
Richard Smith5297d712012-02-25 10:41:10 +0000241<p>The feature name or extension name can also be specified with a preceding and
242following <code>__</code> (double underscore) to avoid interference from a macro
Richard Smith1d9f4c12012-03-01 02:12:07 +0000243with the same name. For instance, <code>__cxx_rvalue_references__</code> can be
244used instead of <code>cxx_rvalue_references</code>.</p>
Richard Smith5297d712012-02-25 10:41:10 +0000245
John Thompson92bd8c72009-11-02 22:28:12 +0000246<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000247<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlssoncae50952010-10-20 02:31:43 +0000248<!-- ======================================================================= -->
249
250<p>This function-like macro takes a single identifier argument that is the name
251of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
252can be used like this:</p>
253
254<blockquote>
255<pre>
256#ifndef __has_attribute // Optional of course.
257 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
258#endif
259
260...
Anders Carlsson961003d2011-01-24 03:54:51 +0000261#if __has_attribute(always_inline)
262#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000263#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000264#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000265#endif
266...
267</pre>
268</blockquote>
269
Jean-Daniel Dupas8a5e7fd2012-03-01 14:53:16 +0000270<p>The attribute name can also be specified with a preceding and
271following <code>__</code> (double underscore) to avoid interference from a macro
272with the same name. For instance, <code>__always_inline__</code> can be used
273instead of <code>always_inline</code>.</p>
274
Anders Carlssoncae50952010-10-20 02:31:43 +0000275<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000276<h2 id="has_include">Include File Checking Macros</h2>
277<!-- ======================================================================= -->
278
279<p>Not all developments systems have the same include files.
280The <a href="#__has_include">__has_include</a> and
281<a href="#__has_include_next">__has_include_next</a> macros allow you to
282check for the existence of an include file before doing
283a possibly failing #include directive.</p>
284
285<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000286<h3><a name="__has_include">__has_include</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000287<!-- ======================================================================= -->
288
289<p>This function-like macro takes a single file name string argument that
290is the name of an include file. It evaluates to 1 if the file can
291be found using the include paths, or 0 otherwise:</p>
292
293<blockquote>
294<pre>
295// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000296#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000297# include "myinclude.h"
298#endif
299
300// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000301#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000302# include "myinclude.h"
303#endif
304</pre>
305</blockquote>
306
307<p>To test for this feature, use #if defined(__has_include).</p>
308
309<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000310<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000311<!-- ======================================================================= -->
312
313<p>This function-like macro takes a single file name string argument that
314is the name of an include file. It is like __has_include except that it
315looks for the second instance of the given file found in the include
316paths. It evaluates to 1 if the second instance of the file can
317be found using the include paths, or 0 otherwise:</p>
318
319<blockquote>
320<pre>
321// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000322#if __has_include_next("myinclude.h") &amp;&amp; __has_include_next(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000323# include_next "myinclude.h"
324#endif
325
326// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000327#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000328# include_next "myinclude.h"
329#endif
330</pre>
331</blockquote>
332
333<p>Note that __has_include_next, like the GNU extension
334#include_next directive, is intended for use in headers only,
335and will issue a warning if used in the top-level compilation
336file. A warning will also be issued if an absolute path
337is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000338
Ted Kremenekd7681502011-10-12 19:46:30 +0000339
340<!-- ======================================================================= -->
341<h3><a name="__has_warning">__has_warning</a></h3>
342<!-- ======================================================================= -->
343
344<p>This function-like macro takes a string literal that represents a command
345 line option for a warning and returns true if that is a valid warning
346 option.</p>
347
348<blockquote>
349<pre>
350#if __has_warning("-Wformat")
351...
352#endif
353</pre>
354</blockquote>
355
Chris Lattner148772a2009-06-13 07:13:28 +0000356<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000357<h2 id="builtinmacros">Builtin Macros</h2>
358<!-- ======================================================================= -->
359
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000360<dl>
361 <dt><code>__BASE_FILE__</code></dt>
362 <dd>Defined to a string that contains the name of the main input
363 file passed to Clang.</dd>
364
365 <dt><code>__COUNTER__</code></dt>
366 <dd>Defined to an integer value that starts at zero and is
367 incremented each time the <code>__COUNTER__</code> macro is
368 expanded.</dd>
369
370 <dt><code>__INCLUDE_LEVEL__</code></dt>
371 <dd>Defined to an integral value that is the include depth of the
372 file currently being translated. For the main file, this value is
373 zero.</dd>
374
375 <dt><code>__TIMESTAMP__</code></dt>
376 <dd>Defined to the date and time of the last modification of the
377 current source file.</dd>
378
379 <dt><code>__clang__</code></dt>
380 <dd>Defined when compiling with Clang</dd>
381
382 <dt><code>__clang_major__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000383 <dd>Defined to the major marketing version number of Clang (e.g., the
384 2 in 2.0.1). Note that marketing version numbers should not be used to
385 check for language features, as different vendors use different numbering
386 schemes. Instead, use the <a href="#feature_check">feature checking
387 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000388
389 <dt><code>__clang_minor__</code></dt>
390 <dd>Defined to the minor version number of Clang (e.g., the 0 in
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000391 2.0.1). Note that marketing version numbers should not be used to
392 check for language features, as different vendors use different numbering
393 schemes. Instead, use the <a href="#feature_check">feature checking
394 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000395
396 <dt><code>__clang_patchlevel__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000397 <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 +0000398
399 <dt><code>__clang_version__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000400 <dd>Defined to a string that captures the Clang marketing version, including
401 the Subversion tag or revision number, e.g., "1.5 (trunk 102332)".</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000402</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000403
404<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000405<h2 id="vectors">Vectors and Extended Vectors</h2>
406<!-- ======================================================================= -->
407
Anton Yartsevda90c772012-01-15 16:22:24 +0000408<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000409
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000410<p>OpenCL vector types are created using <tt>ext_vector_type</tt> attribute. It
411support for <tt>V.xyzw</tt> syntax and other tidbits as seen in OpenCL. An
412example is:</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000413
414<blockquote>
415<pre>
416typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
417typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
418
419float4 foo(float2 a, float2 b) {
420 float4 c;
421 c.xz = a;
422 c.yw = b;
423 return c;
424}
John McCall48209082010-11-08 19:48:17 +0000425</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000426</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000427
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000428<p>Query for this feature with
429<tt>__has_extension(attribute_ext_vector_type)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000430
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000431<p>Giving <tt>-faltivec</tt> option to clang enables support for AltiVec vector
432syntax and functions. For example:</p>
Anton Yartsevda90c772012-01-15 16:22:24 +0000433
434<blockquote>
435<pre>
436vector float foo(vector int a) {
437 vector int b;
438 b = vec_add(a, a) + a;
439 return (vector float)b;
440}
441</pre>
442</blockquote>
443
444<p>NEON vector types are created using <tt>neon_vector_type</tt> and
445<tt>neon_polyvector_type</tt> attributes. For example:</p>
446
447<blockquote>
448<pre>
449typedef <b>__attribute__((neon_vector_type(8)))</b> int8_t int8x8_t;
450typedef <b>__attribute__((neon_polyvector_type(16)))</b> poly8_t poly8x16_t;
451
452int8x8_t foo(int8x8_t a) {
453 int8x8_t v;
454 v = a;
455 return v;
456}
457</pre>
458</blockquote>
459
460<!-- ======================================================================= -->
461<h3><a name="vector_literals">Vector Literals</a></h3>
462<!-- ======================================================================= -->
463
464<p>Vector literals can be used to create vectors from a set of scalars, or
465vectors. Either parentheses or braces form can be used. In the parentheses form
466the number of literal values specified must be one, i.e. referring to a scalar
467value, or must match the size of the vector type being created. If a single
468scalar literal value is specified, the scalar literal value will be replicated
469to all the components of the vector type. In the brackets form any number of
470literals can be specified. For example:</p>
471
472<blockquote>
473<pre>
474typedef int v4si __attribute__((__vector_size__(16)));
475typedef float float4 __attribute__((ext_vector_type(4)));
476typedef float float2 __attribute__((ext_vector_type(2)));
477
478v4si vsi = (v4si){1, 2, 3, 4};
479float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
480vector int vi1 = (vector int)(1); // vi1 will be (1, 1, 1, 1).
481vector int vi2 = (vector int){1}; // vi2 will be (1, 0, 0, 0).
482vector int vi3 = (vector int)(1, 2); // error
483vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
484vector int vi5 = (vector int)(1, 2, 3, 4);
485float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));
486</pre>
487</blockquote>
488
489<!-- ======================================================================= -->
490<h3><a name="vector_operations">Vector Operations</a></h3>
491<!-- ======================================================================= -->
492
493<p>The table below shows the support for each operation by vector extension.
494A dash indicates that an operation is not accepted according to a corresponding
495specification.</p>
496
497<table width="500" border="1" cellspacing="0">
498 <tr>
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000499 <th>Operator</th>
500 <th>OpenCL</th>
501 <th>AltiVec</th>
502 <th>GCC</th>
503 <th>NEON</th>
Anton Yartsevda90c772012-01-15 16:22:24 +0000504 </tr>
505 <tr>
506 <td>[]</td>
507 <td align="center">yes</td>
508 <td align="center">yes</td>
509 <td align="center">yes</td>
510 <td align="center">-</td>
511 </tr>
512 <tr>
513 <td>unary operators +, -</td>
514 <td align="center">yes</td>
515 <td align="center">yes</td>
516 <td align="center">yes</td>
517 <td align="center">-</td>
518 </tr>
519 <tr>
520 <td>++, --</td>
521 <td align="center">yes</td>
522 <td align="center">yes</td>
523 <td align="center">-</td>
524 <td align="center">-</td>
525 </tr>
526 <tr>
527 <td>+, -, *, /, %</td>
528 <td align="center">yes</td>
529 <td align="center">yes</td>
530 <td align="center">yes</td>
531 <td align="center">-</td>
532 </tr>
533 <tr>
534 <td>bitwise operators &, |, ^, ~</td>
535 <td align="center">yes</td>
536 <td align="center">yes</td>
537 <td align="center">yes</td>
538 <td align="center">-</td>
539 </tr>
540 <tr>
541 <td>&gt&gt, &lt&lt</td>
542 <td align="center">yes</td>
543 <td align="center">yes</td>
544 <td align="center">yes</td>
545 <td align="center">-</td>
546 </tr>
547 <tr>
548 <td>!, &&,||</td>
549 <td align="center">no</td>
550 <td align="center">-</td>
551 <td align="center">-</td>
552 <td align="center">-</td>
553 </tr>
554 <tr>
555 <td>==,!=, >, <, >=, <=</td>
556 <td align="center">yes</td>
557 <td align="center">yes</td>
558 <td align="center">-</td>
559 <td align="center">-</td>
560 </tr>
561 <tr>
562 <td>=</td>
563 <td align="center">yes</td>
564 <td align="center">yes</td>
565 <td align="center">yes</td>
566 <td align="center">yes</td>
567 </tr>
568 <tr>
569 <td>:?</td>
570 <td align="center">yes</td>
571 <td align="center">-</td>
572 <td align="center">-</td>
573 <td align="center">-</td>
574 </tr>
575 <tr>
576 <td>sizeof</td>
577 <td align="center">yes</td>
578 <td align="center">yes</td>
579 <td align="center">yes</td>
580 <td align="center">yes</td>
581 </tr>
582</table>
583
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000584<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
585
Chris Lattner5ce933f2009-02-09 08:46:11 +0000586<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000587<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000588<!-- ======================================================================= -->
589
John McCall48209082010-11-08 19:48:17 +0000590<p>An optional string message can be added to the <tt>deprecated</tt>
591and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000592
John McCall48209082010-11-08 19:48:17 +0000593<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000594<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000595</blockquote>
596
597<p>If the deprecated or unavailable declaration is used, the message
598will be incorporated into the appropriate diagnostic:</p>
599
600<blockquote>
Benjamin Kramerb4556862012-03-19 19:12:30 +0000601<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!!
602 [-Wdeprecated-declarations]
John McCall48209082010-11-08 19:48:17 +0000603 explode();
604 ^</pre>
605</blockquote>
606
607<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000608with <tt>__has_extension(attribute_deprecated_with_message)</tt>
609and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000610
611<!-- ======================================================================= -->
612<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
613<!-- ======================================================================= -->
614
615<p>Clang allows attributes to be written on individual enumerators.
616This allows enumerators to be deprecated, made unavailable, etc. The
617attribute must appear after the enumerator name and before any
618initializer, like so:</p>
619
620<blockquote>
621<pre>enum OperationMode {
622 OM_Invalid,
623 OM_Normal,
624 OM_Terrified __attribute__((deprecated)),
625 OM_AbortOnError __attribute__((deprecated)) = 4
626};</pre>
627</blockquote>
628
629<p>Attributes on the <tt>enum</tt> declaration do not apply to
630individual enumerators.</p>
631
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000632<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000633
634<!-- ======================================================================= -->
Daniel Dunbar85ff9692012-04-05 17:10:06 +0000635<h2 id="user_specified_system_framework">'User-Specified' System Frameworks</h2>
636<!-- ======================================================================= -->
637
638<p>Clang provides a mechanism by which frameworks can be built in such a way
639that they will always be treated as being 'system frameworks', even if they are
640not present in a system framework directory. This can be useful to system
641framework developers who want to be able to test building other applications
642with development builds of their framework, including the manner in which the
643compiler changes warning behavior for system headers.</p>
644
645<p>Framework developers can opt-in to this mechanism by creating a
646'.system_framework' file at the top-level of their framework. That is, the
647framework should have contents like:</p>
648
649<pre>
650 .../TestFramework.framework
651 .../TestFramework.framework/.system_framework
652 .../TestFramework.framework/Headers
653 .../TestFramework.framework/Headers/TestFramework.h
654 ...
655</pre>
656
657<p>Clang will treat the presence of this file as an indicator that the framework
658should be treated as a system framework, regardless of how it was found in the
659framework search path. For consistency, we recommend that such files never be
660included in installed versions of the framework.</p>
661
662<!-- ======================================================================= -->
Dmitri Gribenkoccc4edf2012-05-27 14:08:44 +0000663<h2 id="availability">Availability attribute</h2>
Douglas Gregor93a70672012-03-11 04:53:21 +0000664<!-- ======================================================================= -->
665
666<p>Clang introduces the <code>availability</code> attribute, which can
667be placed on declarations to describe the lifecycle of that
668declaration relative to operating system versions. Consider the function declaration for a hypothetical function <code>f</code>:</p>
669
670<pre>
671void f(void) __attribute__((availability(macosx,introduced=10.4,deprecated=10.6,obsoleted=10.7)));
672</pre>
673
674<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>
675
676<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>
677
678<dl>
679 <dt>introduced=<i>version</i></dt>
680 <dd>The first version in which this declaration was introduced.</dd>
681
682 <dt>deprecated=<i>version</i></dt>
683 <dd>The first version in which this declaration was deprecated, meaning that users should migrate away from this API.</dd>
684
685 <dt>obsoleted=<i>version</i></dt>
686 <dd>The first version in which this declaration was obsoleted, meaning that it was removed completely and can no longer be used.</dd>
687
688 <dt>unavailable</dt>
689 <dd>This declaration is never available on this platform.</dd>
690
691 <dt>message=<i>string-literal</i></dt>
692 <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>
693</dl>
694
695<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>
696
697<dl>
698 <dt>ios</dt>
699 <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>
700
701 <dt>macosx</dt>
702 <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>
703</dl>
704
Douglas Gregor594f8412012-03-11 17:21:03 +0000705<p>A declaration can be used even when deploying back to a platform
706version prior to when the declaration was introduced. When this
707happens, the declaration is <a
708 href="https://developer.apple.com/library/mac/#documentation/MacOSX/Conceptual/BPFrameworks/Concepts/WeakLinking.html">weakly
709linked</a>, as if the <code>weak_import</code> attribute were added to the declaration. A weakly-linked declaration may or may not be present a run-time, and a program can determine whether the declaration is present by checking whether the address of that declaration is non-NULL.</p>
710
Douglas Gregor93a70672012-03-11 04:53:21 +0000711<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000712<h2 id="checking_language_features">Checks for Standard Language Features</h2>
713<!-- ======================================================================= -->
714
Richard Smithfafbf062012-04-11 17:55:32 +0000715<p>The <tt>__has_feature</tt> macro can be used to query if certain standard
716language features are enabled. The <tt>__has_extension</tt> macro can be used
717to query if language features are available as an extension when compiling for
718a standard which does not provide them. The features which can be tested are
719listed here.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000720
Richard Smithfafbf062012-04-11 17:55:32 +0000721<h3 id="cxx98">C++98</h3>
722
723<p>The features listed below are part of the C++98 standard. These features are
724enabled by default when compiling C++ code.</p>
725
726<h4 id="cxx_exceptions">C++ exceptions</h4>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000727
Ted Kremenek22c34102009-12-03 02:05:57 +0000728<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Richard Smithfafbf062012-04-11 17:55:32 +0000729example, compiling code with <tt>-fno-exceptions</tt> disables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000730
Richard Smithfafbf062012-04-11 17:55:32 +0000731<h4 id="cxx_rtti">C++ RTTI</h4>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000732
Ted Kremenek0eb95602009-12-03 02:06:43 +0000733<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 +0000734compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000735
Richard Smithfafbf062012-04-11 17:55:32 +0000736<h3 id="cxx11">C++11</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000737
Richard Smithfafbf062012-04-11 17:55:32 +0000738<p>The features listed below are part of the C++11 standard. As a result, all
739these features are enabled with the <tt>-std=c++11</tt> or <tt>-std=gnu++11</tt>
740option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000741
David Blaikie5090e9f2011-10-18 05:49:30 +0000742<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000743
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000744<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 +0000745
David Blaikie5090e9f2011-10-18 05:49:30 +0000746<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000747
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000748<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
749<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie5090e9f2011-10-18 05:49:30 +0000750C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000751
David Blaikie5090e9f2011-10-18 05:49:30 +0000752<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000753
754<p>Use <tt>__has_feature(cxx_alignas)</tt> or
755<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
756specifiers using <tt>alignas</tt> is enabled.</p>
757
David Blaikie5090e9f2011-10-18 05:49:30 +0000758<h4 id="cxx_attributes">C++11 attributes</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000759
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000760<p>Use <tt>__has_feature(cxx_attributes)</tt> or
761<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie5090e9f2011-10-18 05:49:30 +0000762parsing with C++11's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000763
David Blaikie5090e9f2011-10-18 05:49:30 +0000764<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000765
766<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
767for generalized constant expressions (e.g., <tt>constexpr</tt>) is
Richard Smithb5216aa2012-02-14 22:56:17 +0000768enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000769
David Blaikie5090e9f2011-10-18 05:49:30 +0000770<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000771
772<p>Use <tt>__has_feature(cxx_decltype)</tt> or
773<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
Douglas Gregor316551f2012-04-10 20:00:33 +0000774<tt>decltype()</tt> specifier is enabled. C++11's <tt>decltype</tt>
775does not require type-completeness of a function call expression.
776Use <tt>__has_feature(cxx_decltype_incomplete_return_types)</tt>
777or <tt>__has_extension(cxx_decltype_incomplete_return_types)</tt>
778to determine if support for this feature is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000779
David Blaikie5090e9f2011-10-18 05:49:30 +0000780<h4 id="cxx_default_function_template_args">C++11 default template arguments in function templates</h4>
Douglas Gregor07508002011-02-05 20:35:30 +0000781
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000782<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
783<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
784if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000785
Douglas Gregorf695a692011-11-01 01:19:34 +0000786<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
787
788<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
789<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
790defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
791
David Blaikie5090e9f2011-10-18 05:49:30 +0000792<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Sean Huntd9624992011-06-23 06:11:37 +0000793
794<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
795support for delegating constructors is enabled.</p>
796
David Blaikie5090e9f2011-10-18 05:49:30 +0000797<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000798
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000799<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
800<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000801deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000802
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000803<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000804<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
805
David Blaikie5090e9f2011-10-18 05:49:30 +0000806<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000807
808<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
809support for generalized initializers (using braced lists and
Richard Smith88189552012-02-26 07:09:21 +0000810<tt>std::initializer_list</tt>) is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000811
David Blaikie5090e9f2011-10-18 05:49:30 +0000812<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000813
Sebastian Redl72a81d22011-10-10 18:10:00 +0000814<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
815implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000816
David Blaikie5090e9f2011-10-18 05:49:30 +0000817<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000818
819<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>
820
David Blaikie5090e9f2011-10-18 05:49:30 +0000821<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000822
823<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
824<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
825inline namespaces is enabled.</p>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000826
David Blaikie5090e9f2011-10-18 05:49:30 +0000827<h4 id="cxx_lambdas">C++11 lambdas</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000828
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000829<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
830<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregor46e021e2012-02-23 05:44:09 +0000831is enabled. </p>
Douglas Gregorece38942011-08-29 17:28:38 +0000832
Douglas Gregor7b156dd2012-04-04 00:48:39 +0000833<h4 id="cxx_local_type_template_args">C++11 local and unnamed types as template arguments</h4>
834
835<p>Use <tt>__has_feature(cxx_local_type_template_args)</tt> or
836<tt>__has_extension(cxx_local_type_template_args)</tt> to determine if
837support for local and unnamed types as template arguments is enabled.</p>
838
David Blaikie5090e9f2011-10-18 05:49:30 +0000839<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000840
841<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
842<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
843exception specifications is enabled.</p>
844
David Blaikie5090e9f2011-10-18 05:49:30 +0000845<h4 id="cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000846
847<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 +0000848
David Blaikie5090e9f2011-10-18 05:49:30 +0000849<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000850
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000851<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
852<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000853<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000854
David Blaikie5090e9f2011-10-18 05:49:30 +0000855<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000856
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000857<p>Use <tt>__has_feature(cxx_override_control)</tt> or
858<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000859the override control keywords is enabled.</p>
860
David Blaikie5090e9f2011-10-18 05:49:30 +0000861<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000862<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
863<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
864if support for reference-qualified functions (e.g., member functions with
865<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
866is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000867
David Blaikie5090e9f2011-10-18 05:49:30 +0000868<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smitha391a462011-04-15 15:14:40 +0000869
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000870<p>Use <tt>__has_feature(cxx_range_for)</tt> or
871<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
872range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000873
David Blaikie5090e9f2011-10-18 05:49:30 +0000874<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Richard Smith80134582012-03-07 08:57:31 +0000875<p>Use <tt>__has_feature(cxx_raw_string_literals)</tt> to determine if support
876for raw string literals (e.g., <tt>R"x(foo\bar)x"</tt>) is enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000877
David Blaikie5090e9f2011-10-18 05:49:30 +0000878<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000879
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000880<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
881<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000882rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000883
David Blaikie5090e9f2011-10-18 05:49:30 +0000884<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000885
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000886<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
887<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000888compile-time assertions using <tt>static_assert</tt> is enabled.</p>
889
David Blaikie5090e9f2011-10-18 05:49:30 +0000890<h4 id="cxx_auto_type">C++11 type inference</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000891
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000892<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie5090e9f2011-10-18 05:49:30 +0000893<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000894supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
895will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000896
David Blaikie5090e9f2011-10-18 05:49:30 +0000897<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000898
Douglas Gregorece38942011-08-29 17:28:38 +0000899<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
900<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
901strongly typed, scoped enumerations is enabled.</p>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000902
David Blaikie5090e9f2011-10-18 05:49:30 +0000903<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000904
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000905<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
906<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
907alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000908
David Blaikie5090e9f2011-10-18 05:49:30 +0000909<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000910<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
911support for Unicode string literals is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000912
David Blaikie5090e9f2011-10-18 05:49:30 +0000913<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000914
Richard Smithec92bc72012-03-03 23:51:05 +0000915<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 +0000916
David Blaikie5090e9f2011-10-18 05:49:30 +0000917<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000918
Richard Smith9c1dda72012-03-09 08:41:27 +0000919<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 +0000920
David Blaikie5090e9f2011-10-18 05:49:30 +0000921<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000922
923<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
924<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
925for variadic templates is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000926
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000927<h3 id="c11">C11</h3>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000928
Richard Smithfafbf062012-04-11 17:55:32 +0000929<p>The features listed below are part of the C11 standard. As a result, all
930these features are enabled with the <tt>-std=c11</tt> or <tt>-std=gnu11</tt>
931option when compiling C code. Additionally, because these features are all
932backward-compatible, they are available as extensions in all language modes.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000933
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000934<h4 id="c_alignas">C11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000935
936<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
937to determine if support for alignment specifiers using <tt>_Alignas</tt>
938is enabled.</p>
939
Richard Smithfafbf062012-04-11 17:55:32 +0000940<h4 id="c_atomic">C11 atomic operations</h4>
941
942<p>Use <tt>__has_feature(c_atomic)</tt> or <tt>__has_extension(c_atomic)</tt>
943to determine if support for atomic types using <tt>_Atomic</tt> is enabled.
944Clang also provides <a href="#__c11_atomic">a set of builtins</a> which can be
Richard Smithc495e602012-04-19 17:46:52 +0000945used to implement the <tt>&lt;stdatomic.h&gt;</tt> operations on
946<tt>_Atomic</tt> types.</p>
Richard Smithfafbf062012-04-11 17:55:32 +0000947
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000948<h4 id="c_generic_selections">C11 generic selections</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000949
950<p>Use <tt>__has_feature(c_generic_selections)</tt> or
951<tt>__has_extension(c_generic_selections)</tt> to determine if support for
952generic selections is enabled.</p>
953
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000954<p>As an extension, the C11 generic selection expression is available in all
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000955languages supported by Clang. The syntax is the same as that given in the
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000956C11 standard.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000957
958<p>In C, type compatibility is decided according to the rules given in the
959appropriate standard, but in C++, which lacks the type compatibility rules
960used in C, types are considered compatible only if they are equivalent.</p>
961
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000962<h4 id="c_static_assert">C11 <tt>_Static_assert()</tt></h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000963
964<p>Use <tt>__has_feature(c_static_assert)</tt> or
965<tt>__has_extension(c_static_assert)</tt> to determine if support for
966compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
967
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000968<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000969<h2 id="checking_type_traits">Checks for Type Traits</h2>
970<!-- ======================================================================= -->
971
Sean Hunt7e98b472011-06-23 01:21:01 +0000972<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 +0000973<blockquote>
974<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000975#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000976template&lt;typename From, typename To&gt;
977struct is_convertible_to {
978 static const bool value = __is_convertible_to(From, To);
979};
980#else
981// Emulate type trait
982#endif
983</pre>
984</blockquote>
985
986<p>The following type traits are supported by Clang:</p>
987<ul>
988 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
989 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
990 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
991 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
992 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
993 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
994 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
995 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
996 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
997 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
998 <li><code>__is_class</code> (GNU, Microsoft)</li>
999 <li><code>__is_convertible_to</code> (Microsoft)</li>
1000 <li><code>__is_empty</code> (GNU, Microsoft)</li>
1001 <li><code>__is_enum</code> (GNU, Microsoft)</li>
1002 <li><code>__is_pod</code> (GNU, Microsoft)</li>
1003 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
1004 <li><code>__is_union</code> (GNU, Microsoft)</li>
1005 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
Douglas Gregor5e9392b2011-12-03 18:14:24 +00001006 <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 +00001007 <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 +00001008 <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>
1009 <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 +00001010</ul>
1011
1012<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +00001013<h2 id="blocks">Blocks</h2>
1014<!-- ======================================================================= -->
1015
Chris Lattnera7dbdf52009-03-09 07:03:22 +00001016<p>The syntax and high level language feature description is in <a
1017href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
1018details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +00001019href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +00001020
Chris Lattner148772a2009-06-13 07:13:28 +00001021
Peter Collingbournec1b5fa42011-05-13 20:54:45 +00001022<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +00001023
Chris Lattner5ce933f2009-02-09 08:46:11 +00001024<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +00001025<h2 id="objc_features">Objective-C Features</h2>
1026<!-- ======================================================================= -->
1027
1028<h3 id="objc_instancetype">Related result types</h3>
1029
1030<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>
1031
1032<blockquote>
1033<pre>
1034@interface NSObject
1035+ (id)alloc;
1036- (id)init;
1037@end
1038
1039@interface NSArray : NSObject
1040@end
1041</pre>
1042</blockquote>
1043
1044<p>and this common initialization pattern</p>
1045
1046<blockquote>
1047<pre>
1048NSArray *array = [[NSArray alloc] init];
1049</pre>
1050</blockquote>
1051
1052<p>the type of the expression <code>[NSArray alloc]</code> is
1053<code>NSArray*</code> because <code>alloc</code> implicitly has a
1054related result type. Similarly, the type of the expression
1055<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
1056<code>init</code> has a related result type and its receiver is known
1057to 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>
1058
Douglas Gregore97179c2011-09-08 01:46:34 +00001059<p>A method with a related result type can be declared by using the
1060type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
1061is a contextual keyword that is only permitted in the result type of
1062an Objective-C method, e.g.</p>
1063
1064<pre>
1065@interface A
1066+ (<b>instancetype</b>)constructAnA;
1067@end
1068</pre>
1069
1070<p>The related result type can also be inferred for some methods.
1071To determine whether a method has an inferred related result type, the first
Douglas Gregor926df6c2011-06-11 01:09:30 +00001072word in the camel-case selector (e.g., "init" in "initWithObjects") is
Douglas Gregor8a0ace62011-11-03 18:33:01 +00001073considered, and the method will have a related result type if its return
Sean Hunt7e98b472011-06-23 01:21:01 +00001074type is compatible with the type of its class and if</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +00001075
1076<ul>
1077
1078 <li>the first word is "alloc" or "new", and the method is a class
1079 method, or</li>
1080
1081 <li>the first word is "autorelease", "init", "retain", or "self",
1082 and the method is an instance method.</li>
1083
Sean Hunt7e98b472011-06-23 01:21:01 +00001084</ul>
Douglas Gregor926df6c2011-06-11 01:09:30 +00001085
1086<p>If a method with a related result type is overridden by a subclass
1087method, the subclass method must also return a type that is compatible
1088with the subclass type. For example:</p>
1089
1090<blockquote>
1091<pre>
1092@interface NSString : NSObject
1093- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
1094@end
1095</pre>
1096</blockquote>
1097
1098<p>Related result types only affect the type of a message send or
1099property access via the given method. In all other respects, a method
Douglas Gregore97179c2011-09-08 01:46:34 +00001100with a related result type is treated the same way as method that
1101returns <tt>id</tt>.</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +00001102
Douglas Gregoraebb6532011-09-08 17:19:31 +00001103<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
1104the <tt>instancetype</tt> contextual keyword is available.</p>
1105
Douglas Gregor926df6c2011-06-11 01:09:30 +00001106<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +00001107<h2 id="objc_arc">Automatic reference counting </h2>
1108<!-- ======================================================================= -->
1109
1110<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>
1111
1112<!-- ======================================================================= -->
Douglas Gregor5471bc82011-09-08 17:18:35 +00001113<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
1114<!-- ======================================================================= -->
1115
David Blaikie5090e9f2011-10-18 05:49:30 +00001116<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor5471bc82011-09-08 17:18:35 +00001117underlying type within Objective-C. For example, one can write an
1118enumeration type as:</p>
1119
1120<pre>
1121typedef enum : unsigned char { Red, Green, Blue } Color;
1122</pre>
1123
1124<p>This specifies that the underlying type, which is used to store the
1125enumeration value, is <tt>unsigned char</tt>.</p>
1126
1127<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
1128support for fixed underlying types is available in Objective-C.</p>
1129
1130<!-- ======================================================================= -->
Douglas Gregor8a4e1822012-03-09 23:24:48 +00001131<h2 id="objc_lambdas">Interoperability with C++11 lambdas</h2>
1132<!-- ======================================================================= -->
1133
1134<p>Clang provides interoperability between C++11 lambdas and
1135blocks-based APIs, by permitting a lambda to be implicitly converted
1136to a block pointer with the corresponding signature. For example,
1137consider an API such as <code>NSArray</code>'s array-sorting
1138method:</p>
1139
1140<pre> - (NSArray *)sortedArrayUsingComparator:(NSComparator)cmptr; </pre>
1141
1142<p><code>NSComparator</code> is simply a typedef for the block pointer
1143<code>NSComparisonResult (^)(id, id)</code>, and parameters of this
1144type are generally provided with block literals as arguments. However,
1145one can also use a C++11 lambda so long as it provides the same
1146signature (in this case, accepting two parameters of type
1147<code>id</code> and returning an <code>NSComparisonResult</code>):</p>
1148
1149<pre>
1150 NSArray *array = @[@"string 1", @"string 21", @"string 12", @"String 11",
1151 @"String 02"];
1152 const NSStringCompareOptions comparisonOptions
1153 = NSCaseInsensitiveSearch | NSNumericSearch |
1154 NSWidthInsensitiveSearch | NSForcedOrderingSearch;
1155 NSLocale *currentLocale = [NSLocale currentLocale];
1156 NSArray *sorted
1157 = [array sortedArrayUsingComparator:<b>[=](id s1, id s2) -&gt; NSComparisonResult {
1158 NSRange string1Range = NSMakeRange(0, [s1 length]);
1159 return [s1 compare:s2 options:comparisonOptions
1160 range:string1Range locale:currentLocale];
1161 }</b>];
1162 NSLog(@"sorted: %@", sorted);
1163</pre>
1164
1165<p>This code relies on an implicit conversion from the type of the
1166lambda expression (an unnamed, local class type called the <i>closure
1167type</i>) to the corresponding block pointer type. The conversion
1168itself is expressed by a conversion operator in that closure type
1169that produces a block pointer with the same signature as the lambda
1170itself, e.g.,</p>
1171
1172<pre>
1173 operator NSComparisonResult (^)(id, id)() const;
1174</pre>
1175
1176<p>This conversion function returns a new block that simply forwards
1177the two parameters to the lambda object (which it captures by copy),
1178then returns the result. The returned block is first copied (with
1179<tt>Block_copy</tt>) and then autoreleased. As an optimization, if a
1180lambda expression is immediately converted to a block pointer (as in
1181the first example, above), then the block is not copied and
1182autoreleased: rather, it is given the same lifetime as a block literal
1183written at that point in the program, which avoids the overhead of
1184copying a block to the heap in the common case.</p>
1185
Douglas Gregorbccda482012-03-10 22:20:11 +00001186<p>The conversion from a lambda to a block pointer is only available
1187in Objective-C++, and not in C++ with blocks, due to its use of
1188Objective-C memory management (autorelease).</p>
1189
Douglas Gregor8a4e1822012-03-09 23:24:48 +00001190<!-- ======================================================================= -->
Patrick Beardeb382ec2012-04-19 00:25:12 +00001191<h2 id="objc_object_literals_subscripting">Object Literals and Subscripting</h2>
Patrick Beard62f12342012-03-20 21:51:03 +00001192<!-- ======================================================================= -->
1193
Patrick Beardeb382ec2012-04-19 00:25:12 +00001194<p>Clang provides support for <a href="ObjectiveCLiterals.html">Object Literals
1195and Subscripting</a> in Objective-C, which simplifies common Objective-C
1196programming patterns, makes programs more concise, and improves the safety of
1197container creation. There are several feature macros associated with object
1198literals and subscripting: <code>__has_feature(objc_array_literals)</code>
1199tests the availability of array literals;
1200<code>__has_feature(objc_dictionary_literals)</code> tests the availability of
1201dictionary literals; <code>__has_feature(objc_subscripting)</code> tests the
1202availability of object subscripting.</p>
Patrick Beard62f12342012-03-20 21:51:03 +00001203
1204<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +00001205<h2 id="overloading-in-c">Function Overloading in C</h2>
1206<!-- ======================================================================= -->
1207
Chris Lattnerf161d412009-02-13 21:51:45 +00001208<p>Clang provides support for C++ function overloading in C. Function
1209overloading in C is introduced using the <tt>overloadable</tt> attribute. For
1210example, one might provide several overloaded versions of a <tt>tgsin</tt>
1211function that invokes the appropriate standard function computing the sine of a
1212value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
1213precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +00001214
1215<blockquote>
1216<pre>
1217#include &lt;math.h&gt;
1218float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
1219double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
1220long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
1221</pre>
1222</blockquote>
1223
1224<p>Given these declarations, one can call <tt>tgsin</tt> with a
1225<tt>float</tt> value to receive a <tt>float</tt> result, with a
1226<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
1227overloading in C follows the rules of C++ function overloading to pick
1228the best overload given the call arguments, with a few C-specific
1229semantics:</p>
1230<ul>
1231 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
1232 double</tt> is ranked as a floating-point promotion (per C99) rather
1233 than as a floating-point conversion (as in C++).</li>
1234
1235 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
1236 <tt>U*</tt> is considered a pointer conversion (with conversion
1237 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
1238
1239 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
1240 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
1241 conversion is given "conversion" rank.</li>
1242</ul>
1243
1244<p>The declaration of <tt>overloadable</tt> functions is restricted to
1245function declarations and definitions. Most importantly, if any
1246function with a given name is given the <tt>overloadable</tt>
1247attribute, then all function declarations and definitions with that
1248name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +00001249attribute. This rule even applies to redeclarations of functions whose original
1250declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +00001251
1252<blockquote>
1253<pre>
1254int f(int) __attribute__((overloadable));
1255float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
1256
1257int g(int) __attribute__((overloadable));
1258int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
1259</pre>
1260</blockquote>
1261
Douglas Gregor965acbb2009-02-18 07:07:28 +00001262<p>Functions marked <tt>overloadable</tt> must have
1263prototypes. Therefore, the following code is ill-formed:</p>
1264
1265<blockquote>
1266<pre>
1267int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
1268</pre>
1269</blockquote>
1270
1271<p>However, <tt>overloadable</tt> functions are allowed to use a
1272ellipsis 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>
1273
1274<blockquote>
1275<pre>
Chris Lattner02246802009-02-18 22:27:46 +00001276void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +00001277</pre>
1278</blockquote>
1279
Douglas Gregorcb54d432009-02-13 00:57:04 +00001280<p>Functions declared with the <tt>overloadable</tt> attribute have
1281their names mangled according to the same rules as C++ function
1282names. For example, the three <tt>tgsin</tt> functions in our
1283motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +00001284<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +00001285caveats to this use of name mangling:</p>
1286
1287<ul>
1288
1289 <li>Future versions of Clang may change the name mangling of
1290 functions overloaded in C, so you should not depend on an specific
1291 mangling. To be completely safe, we strongly urge the use of
1292 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
1293
1294 <li>The <tt>overloadable</tt> attribute has almost no meaning when
1295 used in C++, because names will already be mangled and functions are
1296 already overloadable. However, when an <tt>overloadable</tt>
1297 function occurs within an <tt>extern "C"</tt> linkage specification,
1298 it's name <i>will</i> be mangled in the same way as it would in
1299 C.</li>
1300</ul>
1301
Peter Collingbournec1b5fa42011-05-13 20:54:45 +00001302<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +00001303
Eli Friedman0c706c22011-09-19 23:17:44 +00001304<!-- ======================================================================= -->
1305<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
1306<!-- ======================================================================= -->
1307
1308<p>clang supports an extension which allows the following in C:</p>
1309
1310<blockquote>
1311<pre>
1312#include &lt;math.h&gt;
1313#include &lt;complex.h&gt;
1314complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
1315</pre>
1316</blockquote>
1317
1318<p>This construct is useful because there is no way to separately
1319initialize the real and imaginary parts of a complex variable in
1320standard C, given that clang does not support <code>_Imaginary</code>.
1321(clang also supports the <code>__real__</code> and <code>__imag__</code>
1322extensions from gcc, which help in some cases, but are not usable in
1323static initializers.)
1324
1325<p>Note that this extension does not allow eliding the braces; the
1326meaning of the following two lines is different:</p>
1327
1328<blockquote>
1329<pre>
1330complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
1331complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
1332</pre>
1333</blockquote>
1334
1335<p>This extension also works in C++ mode, as far as that goes, but does not
1336 apply to the C++ <code>std::complex</code>. (In C++11, list
1337 initialization allows the same syntax to be used with
1338 <code>std::complex</code> with the same meaning.)
Chris Lattner148772a2009-06-13 07:13:28 +00001339
Douglas Gregorcb54d432009-02-13 00:57:04 +00001340<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +00001341<h2 id="builtins">Builtin Functions</h2>
1342<!-- ======================================================================= -->
1343
1344<p>Clang supports a number of builtin library functions with the same syntax as
1345GCC, including things like <tt>__builtin_nan</tt>,
1346<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
1347<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
1348addition to the GCC builtins, Clang supports a number of builtins that GCC does
1349not, which are listed here.</p>
1350
1351<p>Please note that Clang does not and will not support all of the GCC builtins
1352for vector operations. Instead of using builtins, you should use the functions
1353defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
1354define portable wrappers for these. Many of the Clang versions of these
1355functions are implemented directly in terms of <a href="#vectors">extended
1356vector support</a> instead of builtins, in order to reduce the number of
1357builtins that we need to implement.</p>
1358
Chris Lattner5ce933f2009-02-09 08:46:11 +00001359<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +00001360<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +00001361<!-- ======================================================================= -->
1362
Chris Lattneraad826b2009-09-16 18:56:12 +00001363<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +00001364permutation/shuffle/swizzle operations. This builtin is also very important for
1365the implementation of various target-specific header files like
1366<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +00001367</p>
1368
1369<p><b>Syntax:</b></p>
1370
1371<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +00001372__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +00001373</pre>
1374
1375<p><b>Examples:</b></p>
1376
1377<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +00001378 // Identity operation - return 4-element vector V1.
1379 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1380
1381 // "Splat" element 0 of V1 into a 4-element result.
1382 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1383
1384 // Reverse 4-element vector V1.
1385 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1386
1387 // Concatenate every other element of 4-element vectors V1 and V2.
1388 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1389
1390 // Concatenate every other element of 8-element vectors V1 and V2.
1391 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +00001392</pre>
1393
1394<p><b>Description:</b></p>
1395
Chris Lattner6f72da52009-02-13 20:00:20 +00001396<p>The first two arguments to __builtin_shufflevector are vectors that have the
1397same element type. The remaining arguments are a list of integers that specify
1398the elements indices of the first two vectors that should be extracted and
1399returned in a new vector. These element indices are numbered sequentially
1400starting with the first vector, continuing into the second vector. Thus, if
1401vec1 is a 4-element vector, index 5 would refer to the second element of vec2.
Chris Lattner5ce933f2009-02-09 08:46:11 +00001402</p>
1403
Chris Lattner6f72da52009-02-13 20:00:20 +00001404<p>The result of __builtin_shufflevector is a vector
1405with the same element type as vec1/vec2 but that has an element count equal to
1406the number of indices specified.
1407</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +00001408
Chris Lattner21190d52009-09-21 03:09:59 +00001409<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1410
1411<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +00001412<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattner21190d52009-09-21 03:09:59 +00001413<!-- ======================================================================= -->
1414
1415<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1416the program cannot be reached, even if the compiler might otherwise think it
1417can. This is useful to improve optimization and eliminates certain warnings.
1418For example, without the <tt>__builtin_unreachable</tt> in the example below,
1419the compiler assumes that the inline asm can fall through and prints a "function
1420declared 'noreturn' should not return" warning.
1421</p>
1422
1423<p><b>Syntax:</b></p>
1424
1425<pre>
1426__builtin_unreachable()
1427</pre>
1428
1429<p><b>Example of Use:</b></p>
1430
1431<pre>
1432void myabort(void) __attribute__((noreturn));
1433void myabort(void) {
1434 asm("int3");
1435 __builtin_unreachable();
1436}
1437</pre>
1438
1439<p><b>Description:</b></p>
1440
1441<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1442it has undefined behavior, it is a statement that it is never reached and the
1443optimizer can take advantage of this to produce better code. This builtin takes
1444no arguments and produces a void result.
1445</p>
1446
1447<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1448
Chris Lattner23aa9c82011-04-09 03:57:26 +00001449<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +00001450<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner23aa9c82011-04-09 03:57:26 +00001451<!-- ======================================================================= -->
1452
1453<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1454memory.
1455</p>
1456
1457<p><b>Syntax:</b></p>
1458
1459<pre>
1460<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1461</pre>
1462
1463<p><b>Example of Use:</b></p>
1464
1465<pre>
Sean Hunt7e98b472011-06-23 01:21:01 +00001466int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner23aa9c82011-04-09 03:57:26 +00001467</pre>
1468
1469<p><b>Description:</b></p>
1470
1471<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1472intrinsics to allow code to atomically swap the current value with the new
1473value. More importantly, it helps developers write more efficient and correct
1474code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1475relying on the platform specific implementation details of
1476__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1477</p>
1478
Richard Smithfafbf062012-04-11 17:55:32 +00001479<!-- ======================================================================= -->
1480<h3><a name="__c11_atomic">__c11_atomic builtins</a></h3>
1481<!-- ======================================================================= -->
1482
1483<p>Clang provides a set of builtins which are intended to be used to implement
1484C11's <tt>&lt;stdatomic.h&gt;</tt> header. These builtins provide the semantics
1485of the <tt>_explicit</tt> form of the corresponding C11 operation, and are named
1486with a <tt>__c11_</tt> prefix. The supported operations are:</p>
1487
1488<ul>
1489 <li><tt>__c11_atomic_init</tt></li>
1490 <li><tt>__c11_atomic_thread_fence</tt></li>
1491 <li><tt>__c11_atomic_signal_fence</tt></li>
1492 <li><tt>__c11_atomic_is_lock_free</tt></li>
1493 <li><tt>__c11_atomic_store</tt></li>
1494 <li><tt>__c11_atomic_load</tt></li>
1495 <li><tt>__c11_atomic_exchange</tt></li>
1496 <li><tt>__c11_atomic_compare_exchange_strong</tt></li>
1497 <li><tt>__c11_atomic_compare_exchange_weak</tt></li>
1498 <li><tt>__c11_atomic_fetch_add</tt></li>
1499 <li><tt>__c11_atomic_fetch_sub</tt></li>
1500 <li><tt>__c11_atomic_fetch_and</tt></li>
1501 <li><tt>__c11_atomic_fetch_or</tt></li>
1502 <li><tt>__c11_atomic_fetch_xor</tt></li>
1503</ul>
1504
Richard Smithe0d3b4c2012-05-03 18:27:39 +00001505<!-- ======================================================================= -->
1506<h2 id="non-standard-attributes">Non-standard C++11 Attributes</h2>
1507<!-- ======================================================================= -->
1508
Richard Smith207653c2012-05-03 20:05:46 +00001509<p>Clang supports one non-standard C++11 attribute. It resides in the
1510<tt>clang</tt> attribute namespace.</p>
Richard Smithe0d3b4c2012-05-03 18:27:39 +00001511
1512<!-- ======================================================================= -->
1513<h3 id="clang__fallthrough">The <tt>clang::fallthrough</tt> attribute</h3>
1514<!-- ======================================================================= -->
1515
Richard Smith207653c2012-05-03 20:05:46 +00001516<p>The <tt>clang::fallthrough</tt> attribute is used along with the
1517<tt>-Wimplicit-fallthrough</tt> argument to annotate intentional fall-through
1518between switch labels. It can only be applied to a null statement placed at a
Richard Smithe0d3b4c2012-05-03 18:27:39 +00001519point of execution between any statement and the next switch label. It is common
1520to mark these places with a specific comment, but this attribute is meant to
1521replace comments with a more strict annotation, which can be checked by the
1522compiler. This attribute doesn't change semantics of the code and can be used
Richard Smith207653c2012-05-03 20:05:46 +00001523wherever an intended fall-through occurs. It is designed to mimic
1524control-flow statements like <tt>break;</tt>, so it can be placed in most places
1525where <tt>break;</tt> can, but only if there are no statements on the execution
1526path between it and the next switch label.</p>
Richard Smithe0d3b4c2012-05-03 18:27:39 +00001527<p>Here is an example:</p>
1528<pre>
1529// compile with -Wimplicit-fallthrough
1530switch (n) {
1531case 33:
1532 f();
1533case 44: // warning: unannotated fall-through
1534 g();
1535 <b>[[clang::fallthrough]];</b>
1536case 55: // no warning
1537 if (x) {
1538 h();
1539 break;
1540 }
1541 else {
1542 i();
1543 <b>[[clang::fallthrough]];</b>
1544 }
1545case 66: // no warning
1546 p();
Richard Smithfff4a442012-05-03 20:10:41 +00001547 <b>[[clang::fallthrough]];</b> // warning: fallthrough annotation does not directly precede case label
Richard Smithe0d3b4c2012-05-03 18:27:39 +00001548 q();
1549case 77: // warning: unannotated fall-through
1550 r();
1551}
1552</pre>
Chris Lattner21190d52009-09-21 03:09:59 +00001553
Chris Lattner1177f912009-04-09 19:58:15 +00001554<!-- ======================================================================= -->
1555<h2 id="targetspecific">Target-Specific Extensions</h2>
1556<!-- ======================================================================= -->
1557
1558<p>Clang supports some language features conditionally on some targets.</p>
1559
1560<!-- ======================================================================= -->
1561<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1562<!-- ======================================================================= -->
1563
1564<p>The X86 backend has these language extensions:</p>
1565
1566<!-- ======================================================================= -->
1567<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1568<!-- ======================================================================= -->
1569
1570<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +00001571relative to the X86 GS segment register, and address space #257 causes it to be
1572relative to the X86 FS segment. Note that this is a very very low-level
1573feature that should only be used if you know what you're doing (for example in
1574an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +00001575
1576<p>Here is an example:</p>
1577
1578<pre>
1579#define GS_RELATIVE __attribute__((address_space(256)))
1580int foo(int GS_RELATIVE *P) {
1581 return *P;
1582}
1583</pre>
1584
1585<p>Which compiles to (on X86-32):</p>
1586
1587<pre>
1588_foo:
1589 movl 4(%esp), %eax
1590 movl %gs:(%eax), %eax
1591 ret
1592</pre>
1593
Ted Kremeneked869312009-04-10 05:03:33 +00001594<!-- ======================================================================= -->
1595<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1596<!-- ======================================================================= -->
1597
1598<p>Clang supports additional attributes that are useful for documenting program
1599invariants and rules for static analysis tools. The extensions documented here
1600are used by the <a
1601href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1602engine</a> that is part of Clang's Analysis library.</p>
1603
John McCall87494012011-03-18 03:51:49 +00001604<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +00001605
1606<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +00001607attribute. This attribute, which is typically affixed to a function prototype,
1608indicates that a call to a given function never returns. Function prototypes for
1609common functions like <tt>exit</tt> are typically annotated with this attribute,
1610as well as a variety of common assertion handlers. Users can educate the static
1611analyzer about their own custom assertion handles (thus cutting down on false
1612positives due to false paths) by marking their own &quot;panic&quot; functions
1613with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +00001614
1615<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +00001616there are special functions that for all intents and purposes should be
1617considered panic functions (i.e., they are only called when an internal program
1618error occurs) but may actually return so that the program can fail gracefully.
1619The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1620as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +00001621pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +00001622<tt>noreturn</tt>).</p>
1623
1624<p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the
Chris Lattner28935892009-04-10 05:54:56 +00001625same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +00001626placed at the end of function prototypes:</p>
1627
1628<pre>
1629 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +00001630</pre>
1631
John McCall87494012011-03-18 03:51:49 +00001632<p>Query for this feature with
1633<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +00001634
John McCall87494012011-03-18 03:51:49 +00001635<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1636
1637<p>Many methods in Objective-C have conventional meanings determined
1638by their selectors. For the purposes of static analysis, it is
1639sometimes useful to be able to mark a method as having a particular
1640conventional meaning despite not having the right selector, or as not
1641having the conventional meaning that its selector would suggest.
1642For these use cases, we provide an attribute to specifically describe
1643the <q>method family</q> that a method belongs to.</p>
1644
1645<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1646where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1647<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1648can only be placed at the end of a method declaration:</p>
1649
1650<pre>
1651 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1652</pre>
1653
1654<p>Users who do not wish to change the conventional meaning of a
1655method, and who merely want to document its non-standard retain and
1656release semantics, should use the
1657<a href="#attr_retain_release">retaining behavior attributes</a>
1658described below.</p>
1659
1660<p>Query for this feature with
1661<tt>__has_attribute(objc_method_family)</tt>.</p>
1662
1663<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +00001664
1665<p>In Objective-C, functions and methods are generally assumed to take
1666and return objects with +0 retain counts, with some exceptions for
1667special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1668there are exceptions, and so Clang provides attributes to allow these
1669exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +00001670ignore non-leaks). Some exceptions may be better described using
1671the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1672attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +00001673
1674<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1675<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1676and <tt>cf_returns_not_retained</tt> attributes can be placed on
1677methods and functions that return Objective-C or CoreFoundation
1678objects. They are commonly placed at the end of a function prototype
1679or method declaration:</p>
1680
1681<pre>
1682 id foo() <b>__attribute__((ns_returns_retained))</b>;
1683
1684 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1685</pre>
1686
1687<p>The <tt>*_returns_retained</tt> attributes specify that the
1688returned object has a +1 retain count.
1689The <tt>*_returns_not_retained</tt> attributes specify that the return
1690object has a +0 retain count, even if the normal convention for its
1691selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1692returned object is +0, but is guaranteed to live at least as long as the
1693next flush of an autorelease pool.</p>
1694
1695<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1696attributes can be placed on an parameter declaration; they specify
1697that the argument is expected to have a +1 retain count, which will be
1698balanced in some way by the function or method.
1699The <tt>ns_consumes_self</tt> attribute can only be placed on an
1700Objective-C method; it specifies that the method expects
1701its <tt>self</tt> parameter to have a +1 retain count, which it will
1702balance in some way.</p>
1703
1704<pre>
1705 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1706
1707 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1708 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1709</pre>
Ted Kremeneked869312009-04-10 05:03:33 +00001710
John McCall87494012011-03-18 03:51:49 +00001711<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1712<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1713
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +00001714<!-- ======================================================================= -->
1715<h2 id="dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</h2>
1716<!-- ======================================================================= -->
1717<h3 id="address_sanitizer">AddressSanitizer</h3>
1718<p> Use <code>__has_feature(address_sanitizer)</code>
1719to check if the code is being built with <a
1720 href="AddressSanitizer.html">AddressSanitizer</a>.
1721</p>
Kostya Serebryany71efba02012-01-24 19:25:38 +00001722<p>Use <tt>__attribute__((no_address_safety_analysis))</tt> on a function
1723declaration to specify that address safety instrumentation (e.g.
1724AddressSanitizer) should not be applied to that function.
1725</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001726
1727<!-- ======================================================================= -->
Caitlin Sadowski73cbbc82011-07-28 18:38:36 +00001728<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001729<!-- ======================================================================= -->
1730
1731<p>Clang supports additional attributes for checking basic locking policies in
1732multithreaded programs.
1733Clang currently parses the following list of attributes, although
1734<b>the implementation for these annotations is currently in development.</b>
1735For more details, see the
1736<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1737</p>
1738
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001739<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001740
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001741<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1742declaration to specify that the thread safety analysis should not be run on that
1743function. This attribute provides an escape hatch (e.g. for situations when it
1744is difficult to annotate the locking policy). </p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001745
1746<h4 id="ts_lockable">lockable</h4>
1747
1748<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1749that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1750used to check consistency.</p>
1751
1752<h4 id="ts_scopedlockable">scoped_lockable</h4>
1753
1754<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1755specify that it has a "scoped" lockable type. Objects of this type will acquire
1756the lock upon construction and release it upon going out of scope.
1757 This annotation is primarily used to check
1758consistency.</p>
1759
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001760<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001761
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001762<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1763specify that the variable must be accessed while holding some lock.</p>
1764
1765<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1766
1767<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1768specify that the pointer must be dereferenced while holding some lock.</p>
1769
1770<h4 id="ts_guardedby">guarded_by(l)</h4>
1771
1772<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001773specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001774
1775<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1776
1777<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001778specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001779
1780<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1781
1782<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1783of a lockable variable to specify that the lock must be acquired before all
1784attribute arguments. Arguments must be lockable type, and there must be at
1785least one argument.</p>
1786
1787<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1788
1789<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1790of a lockable variable to specify that the lock must be acquired after all
1791attribute arguments. Arguments must be lockable type, and there must be at
1792least one argument.</p>
1793
1794<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1795
1796<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1797declaration to specify that the function acquires all listed locks
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001798exclusively. This attribute takes zero or more arguments: either of lockable
1799type or integers indexing into function parameters of lockable type. If no
1800arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1801enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001802
1803<h4 id="ts_slf">shared_lock_function(...)</h4>
1804
1805<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1806declaration to specify that the function acquires all listed locks, although
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001807 the locks may be shared (e.g. read locks). This attribute takes zero or more
1808arguments: either of lockable type or integers indexing into function
1809parameters of lockable type. If no arguments are given, the acquired lock is
1810implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001811
1812<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1813
1814<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1815declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001816all listed locks exclusively. This attribute takes one or more arguments. The
1817first argument is an integer or boolean value specifying the return value of a
1818successful lock acquisition. The remaining arugments are either of lockable type
1819or integers indexing into function parameters of lockable type. If only one
1820argument is given, the acquired lock is implicitly <tt>this</tt> of the
1821enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001822
1823<h4 id="ts_stf">shared_trylock_function(...)</h4>
1824
1825<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1826declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001827all listed locks, although the locks may be shared (e.g. read locks). This
1828attribute takes one or more arguments. The first argument is an integer or
1829boolean value specifying the return value of a successful lock acquisition. The
1830remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001831function parameters of lockable type. If only one argument is given, the
1832acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1833
1834<h4 id="ts_uf">unlock_function(...)</h4>
1835
1836<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001837declaration to specify that the function release all listed locks. This
1838attribute takes zero or more arguments: either of lockable type or integers
1839indexing into function parameters of lockable type. If no arguments are given,
1840the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001841
1842<h4 id="ts_lr">lock_returned(l)</h4>
1843
1844<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001845declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1846must be of lockable type). This annotation is used to aid in resolving lock
1847expressions.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001848
1849<h4 id="ts_le">locks_excluded(...)</h4>
1850
1851<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001852to specify that the function must not be called with the listed locks. Arguments
1853must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001854
1855<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1856
1857<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1858declaration to specify that the function must be called while holding the listed
1859exclusive locks. Arguments must be lockable type, and there must be at
1860least one argument.</p>
1861
1862<h4 id="ts_slr">shared_locks_required(...)</h4>
1863
1864<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1865declaration to specify that the function must be called while holding the listed
1866shared locks. Arguments must be lockable type, and there must be at
1867least one argument.</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001868
Chris Lattner5ce933f2009-02-09 08:46:11 +00001869</div>
1870</body>
1871</html>