Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 5297d71e8c85f9b04dc31c690e91c6a2f7eddb90 / . / docs / LanguageExtensions.html
blob: cdd197956fe0363d0a043a454983903492f4df61 [file] [log] [blame]
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<!-- Material used from: HTML 4.01 specs: http://www.w3.org/TR/html401/ -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]4<html>
5<head>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]6 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]7 <title>Clang Language Extensions</title>
Sean Hunt64f857b2011-06-23 01:22:53 +0000[diff] [blame]8 <link type="text/css" rel="stylesheet" href="../menu.css">
9 <link type="text/css" rel="stylesheet" href="../content.css">
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]10 <style type="text/css">
11 td {
12 vertical-align: top;
13 }
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]14 th { background-color: #ffddaa; }
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]15 </style>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]16</head>
17<body>
18
19<!--#include virtual="../menu.html.incl"-->
20
21<div id="content">
22
23<h1>Clang Language Extensions</h1>
24
25<ul>
26<li><a href="#intro">Introduction</a></li>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]27<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]28<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]29<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]30<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]31<li><a href="#deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> attributes</a></li>
32<li><a href="#attributes-on-enumerators">Attributes on enumerators</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]33<li><a href="#checking_language_features">Checks for Standard Language Features</a>
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]34 <ul>
35 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
36 <li><a href="#cxx_rtti">C++ RTTI</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]37 </ul></li>
38<li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]39 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]40 <li><a href="#cxx0x">C++11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]41 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]42 <li><a href="#cxx_access_control_sfinae">C++11 SFINAE includes
43 access control</a></li>
44 <li><a href="#cxx_alias_templates">C++11 alias templates</a></li>
45 <li><a href="#cxx_alignas">C++11 alignment specifiers</a></li>
46 <li><a href="#cxx_attributes">C++11 attributes</a></li>
47 <li><a href="#cxx_constexpr">C++11 generalized constant expressions</a></li>
48 <li><a href="#cxx_decltype">C++11 <tt>decltype()</tt></a></li>
49 <li><a href="#cxx_default_function_template_args">C++11 default template arguments in function templates</a></li>
Douglas Gregorf695a692011-11-01 01:19:34 +0000[diff] [blame]50 <li><a href="#cxx_defaulted_functions">C++11 defaulted functions</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]51 <li><a href="#cxx_delegating_constructor">C++11 delegating constructors</a></li>
52 <li><a href="#cxx_deleted_functions">C++11 deleted functions</a></li>
53 <li><a href="#cxx_explicit_conversions">C++11 explicit conversion functions</a></li>
54 <li><a href="#cxx_generalized_initializers">C++11 generalized initializers</a></li>
55 <li><a href="#cxx_implicit_moves">C++11 implicit move constructors/assignment operators</a></li>
56 <li><a href="#cxx_inheriting_constructors">C++11 inheriting constructors</a></li>
57 <li><a href="#cxx_inline_namespaces">C++11 inline namespaces</a></li>
58 <li><a href="#cxx_lambdas">C++11 lambdas</a></li>
59 <li><a href="#cxx_noexcept">C++11 noexcept specification</a></li>
60 <li><a href="#cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</a></li>
61 <li><a href="#cxx_nullptr">C++11 nullptr</a></li>
62 <li><a href="#cxx_override_control">C++11 override control</a></li>
63 <li><a href="#cxx_range_for">C++11 range-based for loop</a></li>
64 <li><a href="#cxx_raw_string_literals">C++11 raw string literals</a></li>
65 <li><a href="#cxx_rvalue_references">C++11 rvalue references</a></li>
66 <li><a href="#cxx_reference_qualified_functions">C++11 reference-qualified functions</a></li>
67 <li><a href="#cxx_static_assert">C++11 <tt>static_assert()</tt></a></li>
68 <li><a href="#cxx_auto_type">C++11 type inference</a></li>
69 <li><a href="#cxx_strong_enums">C++11 strongly-typed enumerations</a></li>
70 <li><a href="#cxx_trailing_return">C++11 trailing return type</a></li>
71 <li><a href="#cxx_unicode_literals">C++11 Unicode string literals</a></li>
72 <li><a href="#cxx_unrestricted_unions">C++11 unrestricted unions</a></li>
73 <li><a href="#cxx_user_literals">C++11 user-defined literals</a></li>
74 <li><a href="#cxx_variadic_templates">C++11 variadic templates</a></li>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]75 </ul></li>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]76 <li><a href="#c11">C11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]77 <ul>
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]78 <li><a href="#c_alignas">C11 alignment specifiers</a></li>
79 <li><a href="#c_generic_selections">C11 generic selections</a></li>
80 <li><a href="#c_static_assert">C11 <tt>_Static_assert()</tt></a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]81 </ul></li>
82 </ul> </li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]83<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]84<li><a href="#blocks">Blocks</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]85<li><a href="#objc_features">Objective-C Features</a>
86 <ul>
87 <li><a href="#objc_instancetype">Related result types</a></li>
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]88 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]89 <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]90 </ul>
91</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]92<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]93<li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]94<li><a href="#builtins">Builtin Functions</a>
95 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]96 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]97 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]98 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]99 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]100</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]101<li><a href="#targetspecific">Target-Specific Extensions</a>
102 <ul>
103 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
104 </ul>
105</li>
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]106<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]107<li><a href="#dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</a>
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]108 <ul>
109 <li><a href="#address_sanitizer">AddressSanitizer</a></li>
110 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]111</li>
112<li><a href="#threadsafety">Thread Safety Annotation Checking</a>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]113 <ul>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]114 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
115 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
116 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]117 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
118 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]119 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
120 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
121 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
122 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
123 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
124 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
125 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
126 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
127 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
128 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
129 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
130 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
131 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]132 </ul>
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]133</li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]134</ul>
135
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]136<!-- ======================================================================= -->
137<h2 id="intro">Introduction</h2>
138<!-- ======================================================================= -->
139
140<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]141addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]142range of GCC extensions. Please see the <a
143href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
144more information on these extensions.</p>
145
146<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]147<h2 id="feature_check">Feature Checking Macros</h2>
148<!-- ======================================================================= -->
149
150<p>Language extensions can be very useful, but only if you know you can depend
Chris Lattnerc70e1932011-03-21 16:25:11 +0000[diff] [blame]151on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]152function-like macros. This allows you to directly test for a feature in your
153code without having to resort to something like autoconf or fragile "compiler
154version checks".</p>
155
156<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]157<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]158<!-- ======================================================================= -->
159
160<p>This function-like macro takes a single identifier argument that is the name
161of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
162not. It can be used like this:</p>
163
164<blockquote>
165<pre>
166#ifndef __has_builtin // Optional of course.
167 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
168#endif
169
170...
171#if __has_builtin(__builtin_trap)
172 __builtin_trap();
173#else
174 abort();
175#endif
176...
177</pre>
178</blockquote>
179
180
181<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]182<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]183<!-- ======================================================================= -->
184
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]185<p>These function-like macros take a single identifier argument that is the
186name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
187is both supported by Clang and standardized in the current language standard
188or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
189<code>__has_extension</code> evaluates to 1 if the feature is supported by
190Clang in the current language (either as a language extension or a standard
191language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]192
193<blockquote>
194<pre>
195#ifndef __has_feature // Optional of course.
196 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
197#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]198#ifndef __has_extension
199 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
200#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]201
202...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]203#if __has_feature(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]204// This code will only be compiled with the -std=c++11 and -std=gnu++11
205// options, because rvalue references are only standardized in C++11.
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]206#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]207
208#if __has_extension(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]209// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]210// and -std=gnu++98 options, because rvalue references are supported as a
211// language extension in C++98.
212#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]213</pre>
214</blockquote>
215
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]216<p id="has_feature_back_compat">For backwards compatibility reasons,
217<code>__has_feature</code> can also be used to test for support for
218non-standardized features, i.e. features not prefixed <code>c_</code>,
219<code>cxx_</code> or <code>objc_</code>.</p>
220
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]221<p id="has_feature_for_non_language_features">
222Another use of <code>__has_feature</code> is to check for compiler features
223not related to the language standard, such as e.g.
224<a href="AddressSanitizer.html">AddressSanitizer</a>.
225
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]226<p>If the <code>-pedantic-errors</code> option is given,
227<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
228
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]229<p>The feature tag is described along with the language feature below.</p>
230
Richard Smith5297d712012-02-25 10:41:10 +0000[diff] [blame^]231<p>The feature name or extension name can also be specified with a preceding and
232following <code>__</code> (double underscore) to avoid interference from a macro
233with the same name. For instance, <code>__always_inline__</code> can be used
234instead of <code>always_inline</code>.</p>
235
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]236<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]237<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]238<!-- ======================================================================= -->
239
240<p>This function-like macro takes a single identifier argument that is the name
241of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
242can be used like this:</p>
243
244<blockquote>
245<pre>
246#ifndef __has_attribute // Optional of course.
247 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
248#endif
249
250...
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]251#if __has_attribute(always_inline)
252#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]253#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]254#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]255#endif
256...
257</pre>
258</blockquote>
259
260<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]261<h2 id="has_include">Include File Checking Macros</h2>
262<!-- ======================================================================= -->
263
264<p>Not all developments systems have the same include files.
265The <a href="#__has_include">__has_include</a> and
266<a href="#__has_include_next">__has_include_next</a> macros allow you to
267check for the existence of an include file before doing
268a possibly failing #include directive.</p>
269
270<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]271<h3><a name="__has_include">__has_include</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]272<!-- ======================================================================= -->
273
274<p>This function-like macro takes a single file name string argument that
275is the name of an include file. It evaluates to 1 if the file can
276be found using the include paths, or 0 otherwise:</p>
277
278<blockquote>
279<pre>
280// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]281#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]282# include "myinclude.h"
283#endif
284
285// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]286#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]287# include "myinclude.h"
288#endif
289</pre>
290</blockquote>
291
292<p>To test for this feature, use #if defined(__has_include).</p>
293
294<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]295<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]296<!-- ======================================================================= -->
297
298<p>This function-like macro takes a single file name string argument that
299is the name of an include file. It is like __has_include except that it
300looks for the second instance of the given file found in the include
301paths. It evaluates to 1 if the second instance of the file can
302be found using the include paths, or 0 otherwise:</p>
303
304<blockquote>
305<pre>
306// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]307#if __has_include_next("myinclude.h") &amp;&amp; __has_include_next(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]308# include_next "myinclude.h"
309#endif
310
311// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]312#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]313# include_next "myinclude.h"
314#endif
315</pre>
316</blockquote>
317
318<p>Note that __has_include_next, like the GNU extension
319#include_next directive, is intended for use in headers only,
320and will issue a warning if used in the top-level compilation
321file. A warning will also be issued if an absolute path
322is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]323
Ted Kremenekd7681502011-10-12 19:46:30 +0000[diff] [blame]324
325<!-- ======================================================================= -->
326<h3><a name="__has_warning">__has_warning</a></h3>
327<!-- ======================================================================= -->
328
329<p>This function-like macro takes a string literal that represents a command
330 line option for a warning and returns true if that is a valid warning
331 option.</p>
332
333<blockquote>
334<pre>
335#if __has_warning("-Wformat")
336...
337#endif
338</pre>
339</blockquote>
340
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]341<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]342<h2 id="builtinmacros">Builtin Macros</h2>
343<!-- ======================================================================= -->
344
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]345<dl>
346 <dt><code>__BASE_FILE__</code></dt>
347 <dd>Defined to a string that contains the name of the main input
348 file passed to Clang.</dd>
349
350 <dt><code>__COUNTER__</code></dt>
351 <dd>Defined to an integer value that starts at zero and is
352 incremented each time the <code>__COUNTER__</code> macro is
353 expanded.</dd>
354
355 <dt><code>__INCLUDE_LEVEL__</code></dt>
356 <dd>Defined to an integral value that is the include depth of the
357 file currently being translated. For the main file, this value is
358 zero.</dd>
359
360 <dt><code>__TIMESTAMP__</code></dt>
361 <dd>Defined to the date and time of the last modification of the
362 current source file.</dd>
363
364 <dt><code>__clang__</code></dt>
365 <dd>Defined when compiling with Clang</dd>
366
367 <dt><code>__clang_major__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]368 <dd>Defined to the major marketing version number of Clang (e.g., the
369 2 in 2.0.1). Note that marketing version numbers should not be used to
370 check for language features, as different vendors use different numbering
371 schemes. Instead, use the <a href="#feature_check">feature checking
372 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]373
374 <dt><code>__clang_minor__</code></dt>
375 <dd>Defined to the minor version number of Clang (e.g., the 0 in
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]376 2.0.1). Note that marketing version numbers should not be used to
377 check for language features, as different vendors use different numbering
378 schemes. Instead, use the <a href="#feature_check">feature checking
379 macros</a>.</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]380
381 <dt><code>__clang_patchlevel__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]382 <dd>Defined to the marketing patch level of Clang (e.g., the 1 in 2.0.1).</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]383
384 <dt><code>__clang_version__</code></dt>
Chris Lattnerd4b66b92011-12-15 19:06:36 +0000[diff] [blame]385 <dd>Defined to a string that captures the Clang marketing version, including
386 the Subversion tag or revision number, e.g., "1.5 (trunk 102332)".</dd>
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]387</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]388
389<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]390<h2 id="vectors">Vectors and Extended Vectors</h2>
391<!-- ======================================================================= -->
392
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]393<p>Supports the GCC, OpenCL, AltiVec and NEON vector extensions.</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]394
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]395<p>OpenCL vector types are created using <tt>ext_vector_type</tt> attribute. It
396support for <tt>V.xyzw</tt> syntax and other tidbits as seen in OpenCL. An
397example is:</p>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]398
399<blockquote>
400<pre>
401typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
402typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
403
404float4 foo(float2 a, float2 b) {
405 float4 c;
406 c.xz = a;
407 c.yw = b;
408 return c;
409}
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]410</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]411</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]412
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]413<p>Query for this feature with
414<tt>__has_extension(attribute_ext_vector_type)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]415
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]416<p>Giving <tt>-faltivec</tt> option to clang enables support for AltiVec vector
417syntax and functions. For example:</p>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]418
419<blockquote>
420<pre>
421vector float foo(vector int a) {
422 vector int b;
423 b = vec_add(a, a) + a;
424 return (vector float)b;
425}
426</pre>
427</blockquote>
428
429<p>NEON vector types are created using <tt>neon_vector_type</tt> and
430<tt>neon_polyvector_type</tt> attributes. For example:</p>
431
432<blockquote>
433<pre>
434typedef <b>__attribute__((neon_vector_type(8)))</b> int8_t int8x8_t;
435typedef <b>__attribute__((neon_polyvector_type(16)))</b> poly8_t poly8x16_t;
436
437int8x8_t foo(int8x8_t a) {
438 int8x8_t v;
439 v = a;
440 return v;
441}
442</pre>
443</blockquote>
444
445<!-- ======================================================================= -->
446<h3><a name="vector_literals">Vector Literals</a></h3>
447<!-- ======================================================================= -->
448
449<p>Vector literals can be used to create vectors from a set of scalars, or
450vectors. Either parentheses or braces form can be used. In the parentheses form
451the number of literal values specified must be one, i.e. referring to a scalar
452value, or must match the size of the vector type being created. If a single
453scalar literal value is specified, the scalar literal value will be replicated
454to all the components of the vector type. In the brackets form any number of
455literals can be specified. For example:</p>
456
457<blockquote>
458<pre>
459typedef int v4si __attribute__((__vector_size__(16)));
460typedef float float4 __attribute__((ext_vector_type(4)));
461typedef float float2 __attribute__((ext_vector_type(2)));
462
463v4si vsi = (v4si){1, 2, 3, 4};
464float4 vf = (float4)(1.0f, 2.0f, 3.0f, 4.0f);
465vector int vi1 = (vector int)(1); // vi1 will be (1, 1, 1, 1).
466vector int vi2 = (vector int){1}; // vi2 will be (1, 0, 0, 0).
467vector int vi3 = (vector int)(1, 2); // error
468vector int vi4 = (vector int){1, 2}; // vi4 will be (1, 2, 0, 0).
469vector int vi5 = (vector int)(1, 2, 3, 4);
470float4 vf = (float4)((float2)(1.0f, 2.0f), (float2)(3.0f, 4.0f));
471</pre>
472</blockquote>
473
474<!-- ======================================================================= -->
475<h3><a name="vector_operations">Vector Operations</a></h3>
476<!-- ======================================================================= -->
477
478<p>The table below shows the support for each operation by vector extension.
479A dash indicates that an operation is not accepted according to a corresponding
480specification.</p>
481
482<table width="500" border="1" cellspacing="0">
483 <tr>
Benjamin Kramer3419d7c2012-01-15 16:42:14 +0000[diff] [blame]484 <th>Operator</th>
485 <th>OpenCL</th>
486 <th>AltiVec</th>
487 <th>GCC</th>
488 <th>NEON</th>
Anton Yartsevda90c772012-01-15 16:22:24 +0000[diff] [blame]489 </tr>
490 <tr>
491 <td>[]</td>
492 <td align="center">yes</td>
493 <td align="center">yes</td>
494 <td align="center">yes</td>
495 <td align="center">-</td>
496 </tr>
497 <tr>
498 <td>unary operators +, -</td>
499 <td align="center">yes</td>
500 <td align="center">yes</td>
501 <td align="center">yes</td>
502 <td align="center">-</td>
503 </tr>
504 <tr>
505 <td>++, --</td>
506 <td align="center">yes</td>
507 <td align="center">yes</td>
508 <td align="center">-</td>
509 <td align="center">-</td>
510 </tr>
511 <tr>
512 <td>+, -, *, /, %</td>
513 <td align="center">yes</td>
514 <td align="center">yes</td>
515 <td align="center">yes</td>
516 <td align="center">-</td>
517 </tr>
518 <tr>
519 <td>bitwise operators &, |, ^, ~</td>
520 <td align="center">yes</td>
521 <td align="center">yes</td>
522 <td align="center">yes</td>
523 <td align="center">-</td>
524 </tr>
525 <tr>
526 <td>&gt&gt, &lt&lt</td>
527 <td align="center">yes</td>
528 <td align="center">yes</td>
529 <td align="center">yes</td>
530 <td align="center">-</td>
531 </tr>
532 <tr>
533 <td>!, &&,||</td>
534 <td align="center">no</td>
535 <td align="center">-</td>
536 <td align="center">-</td>
537 <td align="center">-</td>
538 </tr>
539 <tr>
540 <td>==,!=, >, <, >=, <=</td>
541 <td align="center">yes</td>
542 <td align="center">yes</td>
543 <td align="center">-</td>
544 <td align="center">-</td>
545 </tr>
546 <tr>
547 <td>=</td>
548 <td align="center">yes</td>
549 <td align="center">yes</td>
550 <td align="center">yes</td>
551 <td align="center">yes</td>
552 </tr>
553 <tr>
554 <td>:?</td>
555 <td align="center">yes</td>
556 <td align="center">-</td>
557 <td align="center">-</td>
558 <td align="center">-</td>
559 </tr>
560 <tr>
561 <td>sizeof</td>
562 <td align="center">yes</td>
563 <td align="center">yes</td>
564 <td align="center">yes</td>
565 <td align="center">yes</td>
566 </tr>
567</table>
568
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]569<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
570
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]571<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]572<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]573<!-- ======================================================================= -->
574
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]575<p>An optional string message can be added to the <tt>deprecated</tt>
576and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]577
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]578<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]579<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]580</blockquote>
581
582<p>If the deprecated or unavailable declaration is used, the message
583will be incorporated into the appropriate diagnostic:</p>
584
585<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]586<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!! [-Wdeprecated-declarations]
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]587 explode();
588 ^</pre>
589</blockquote>
590
591<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]592with <tt>__has_extension(attribute_deprecated_with_message)</tt>
593and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]594
595<!-- ======================================================================= -->
596<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
597<!-- ======================================================================= -->
598
599<p>Clang allows attributes to be written on individual enumerators.
600This allows enumerators to be deprecated, made unavailable, etc. The
601attribute must appear after the enumerator name and before any
602initializer, like so:</p>
603
604<blockquote>
605<pre>enum OperationMode {
606 OM_Invalid,
607 OM_Normal,
608 OM_Terrified __attribute__((deprecated)),
609 OM_AbortOnError __attribute__((deprecated)) = 4
610};</pre>
611</blockquote>
612
613<p>Attributes on the <tt>enum</tt> declaration do not apply to
614individual enumerators.</p>
615
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]616<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]617
618<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]619<h2 id="checking_language_features">Checks for Standard Language Features</h2>
620<!-- ======================================================================= -->
621
622<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
623enabled. Those features are listed here.</p>
624
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]625<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]626
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]627<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Sean Hunt647ba1b2011-06-23 00:42:53 +0000[diff] [blame]628example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]629
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]630<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]631
Ted Kremenek0eb95602009-12-03 02:06:43 +0000[diff] [blame]632<p>Use <tt>__has_feature(cxx_rtti)</tt> to determine if C++ RTTI has been enabled. For example,
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]633compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]634
635<!-- ======================================================================= -->
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]636<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
637<!-- ======================================================================= -->
638
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]639<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
640to query if certain upcoming standard language features are enabled. Those
641features are listed here. Features that are not yet implemented will be
642noted.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]643
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]644<h3 id="cxx0x">C++11</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]645
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]646<p>The features listed below are slated for inclusion in the upcoming
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]647C++11 standard. As a result, all these features are enabled
648with the <tt>-std=c++11</tt> option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]649
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]650<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]651
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]652<p>Use <tt>__has_feature(cxx_access_control_sfinae)</tt> or <tt>__has_extension(cxx_access_control_sfinae)</tt> to determine whether access-control errors (e.g., calling a private constructor) are considered to be template argument deduction errors (aka SFINAE errors), per <a href="http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#1170">C++ DR1170</a>.</p>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]653
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]654<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]655
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]656<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
657<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]658C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]659
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]660<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]661
662<p>Use <tt>__has_feature(cxx_alignas)</tt> or
663<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
664specifiers using <tt>alignas</tt> is enabled.</p>
665
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]666<h4 id="cxx_attributes">C++11 attributes</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]667
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]668<p>Use <tt>__has_feature(cxx_attributes)</tt> or
669<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]670parsing with C++11's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]671
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]672<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]673
674<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
675for generalized constant expressions (e.g., <tt>constexpr</tt>) is
Richard Smithb5216aa2012-02-14 22:56:17 +0000[diff] [blame]676enabled.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]677
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]678<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]679
680<p>Use <tt>__has_feature(cxx_decltype)</tt> or
681<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
682<tt>decltype()</tt> specifier is enabled.</p>
683
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]684<h4 id="cxx_default_function_template_args">C++11 default template arguments in function templates</h4>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]685
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]686<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
687<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
688if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]689
Douglas Gregorf695a692011-11-01 01:19:34 +0000[diff] [blame]690<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
691
692<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
693<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
694defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
695
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]696<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Sean Huntd9624992011-06-23 06:11:37 +0000[diff] [blame]697
698<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
699support for delegating constructors is enabled.</p>
700
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]701<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]702
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]703<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
704<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]705deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]706
Benjamin Kramer665a8dc2012-01-15 15:26:07 +0000[diff] [blame]707<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]708<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
709
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]710<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]711
712<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
713support for generalized initializers (using braced lists and
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]714<tt>std::initializer_list</tt>) is enabled. Clang does not currently implement
715this feature.</p>
716
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]717<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]718
Sebastian Redl72a81d22011-10-10 18:10:00 +0000[diff] [blame]719<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
720implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]721
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]722<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]723
724<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>
725
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]726<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]727
728<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
729<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
730inline namespaces is enabled.</p>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000[diff] [blame]731
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]732<h4 id="cxx_lambdas">C++11 lambdas</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]733
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]734<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
735<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregor46e021e2012-02-23 05:44:09 +0000[diff] [blame]736is enabled. </p>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]737
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]738<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]739
740<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
741<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
742exception specifications is enabled.</p>
743
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]744<h4 id="cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]745
746<p>Use <tt>__has_feature(cxx_nonstatic_member_init)</tt> to determine whether in-class initialization of non-static data members is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]747
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]748<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]749
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]750<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
751<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000[diff] [blame]752<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]753
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]754<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]755
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]756<p>Use <tt>__has_feature(cxx_override_control)</tt> or
757<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]758the override control keywords is enabled.</p>
759
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]760<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]761<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
762<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
763if support for reference-qualified functions (e.g., member functions with
764<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
765is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]766
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]767<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]768
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]769<p>Use <tt>__has_feature(cxx_range_for)</tt> or
770<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
771range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]772
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]773<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]774<p>Use <tt>__has_feature(cxx_raw_string_literals)</tt> to determine if support for raw string literals (e.g., <tt>R"foo\bar"</tt>) is enabled.</p>
775
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]776<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]777
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]778<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
779<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]780rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]781
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]782<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]783
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]784<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
785<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]786compile-time assertions using <tt>static_assert</tt> is enabled.</p>
787
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]788<h4 id="cxx_auto_type">C++11 type inference</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]789
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]790<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]791<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]792supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
793will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]794
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]795<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]796
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]797<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
798<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
799strongly typed, scoped enumerations is enabled.</p>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]800
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]801<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]802
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]803<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
804<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
805alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]806
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]807<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]808<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
809support for Unicode string literals is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]810
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]811<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]812
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]813<p>Use <tt>__has_feature(cxx_unrestricted_unions)</tt> to determine if support for unrestricted unions is enabled. Clang does not currently support this feature.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]814
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]815<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]816
817<p>Use <tt>__has_feature(cxx_user_literals)</tt> to determine if support for user-defined literals is enabled. Clang does not currently support this feature.</p>
818
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]819<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000[diff] [blame]820
821<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
822<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
823for variadic templates is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]824
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]825<h3 id="c11">C11</h3>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]826
827<p>The features listed below are slated for inclusion in the upcoming
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]828C11 standard. As a result, all these features are enabled
829with the <tt>-std=c11</tt> option when compiling C code.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]830
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]831<h4 id="c_alignas">C11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000[diff] [blame]832
833<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
834to determine if support for alignment specifiers using <tt>_Alignas</tt>
835is enabled.</p>
836
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]837<h4 id="c_generic_selections">C11 generic selections</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]838
839<p>Use <tt>__has_feature(c_generic_selections)</tt> or
840<tt>__has_extension(c_generic_selections)</tt> to determine if support for
841generic selections is enabled.</p>
842
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]843<p>As an extension, the C11 generic selection expression is available in all
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]844languages supported by Clang. The syntax is the same as that given in the
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]845C11 standard.</p>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]846
847<p>In C, type compatibility is decided according to the rules given in the
848appropriate standard, but in C++, which lacks the type compatibility rules
849used in C, types are considered compatible only if they are equivalent.</p>
850
Benjamin Kramerffbe9b92011-12-23 17:00:35 +0000[diff] [blame]851<h4 id="c_static_assert">C11 <tt>_Static_assert()</tt></h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]852
853<p>Use <tt>__has_feature(c_static_assert)</tt> or
854<tt>__has_extension(c_static_assert)</tt> to determine if support for
855compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
856
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]857<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]858<h2 id="checking_type_traits">Checks for Type Traits</h2>
859<!-- ======================================================================= -->
860
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]861<p>Clang supports the <a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html">GNU C++ type traits</a> and a subset of the <a href="http://msdn.microsoft.com/en-us/library/ms177194(v=VS.100).aspx">Microsoft Visual C++ Type traits</a>. For each supported type trait <code>__X</code>, <code>__has_extension(X)</code> indicates the presence of the type trait. For example:
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]862<blockquote>
863<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]864#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]865template&lt;typename From, typename To&gt;
866struct is_convertible_to {
867 static const bool value = __is_convertible_to(From, To);
868};
869#else
870// Emulate type trait
871#endif
872</pre>
873</blockquote>
874
875<p>The following type traits are supported by Clang:</p>
876<ul>
877 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
878 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
879 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
880 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
881 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
882 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
883 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
884 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
885 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
886 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
887 <li><code>__is_class</code> (GNU, Microsoft)</li>
888 <li><code>__is_convertible_to</code> (Microsoft)</li>
889 <li><code>__is_empty</code> (GNU, Microsoft)</li>
890 <li><code>__is_enum</code> (GNU, Microsoft)</li>
891 <li><code>__is_pod</code> (GNU, Microsoft)</li>
892 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
893 <li><code>__is_union</code> (GNU, Microsoft)</li>
894 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
Douglas Gregor5e9392b2011-12-03 18:14:24 +0000[diff] [blame]895 <li><code>__is_final</code>: Determines whether the given type is declared with a <code>final</code> class-virt-specifier.</li>
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]896 <li><code>__underlying_type(type)</code>: Retrieves the underlying type for a given <code>enum</code> type. This trait is required to implement the C++11 standard library.</li>
Douglas Gregor4ca8ac22012-02-24 07:38:34 +0000[diff] [blame]897 <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>
898 <li><code>__is_trivially_constructible(type, argtypes...)</code>: Determines whether a value of type <tt>type</tt> can be direct-initialized with arguments of types <tt>argtypes...</tt> such that no non-trivial functions are called as part of that initialization. This trait is required to implement the C++11 standard library.</li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]899</ul>
900
901<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]902<h2 id="blocks">Blocks</h2>
903<!-- ======================================================================= -->
904
Chris Lattnera7dbdf52009-03-09 07:03:22 +0000[diff] [blame]905<p>The syntax and high level language feature description is in <a
906href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
907details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +0000[diff] [blame]908href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]909
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]910
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]911<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]912
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]913<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]914<h2 id="objc_features">Objective-C Features</h2>
915<!-- ======================================================================= -->
916
917<h3 id="objc_instancetype">Related result types</h3>
918
919<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>
920
921<blockquote>
922<pre>
923@interface NSObject
924+ (id)alloc;
925- (id)init;
926@end
927
928@interface NSArray : NSObject
929@end
930</pre>
931</blockquote>
932
933<p>and this common initialization pattern</p>
934
935<blockquote>
936<pre>
937NSArray *array = [[NSArray alloc] init];
938</pre>
939</blockquote>
940
941<p>the type of the expression <code>[NSArray alloc]</code> is
942<code>NSArray*</code> because <code>alloc</code> implicitly has a
943related result type. Similarly, the type of the expression
944<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
945<code>init</code> has a related result type and its receiver is known
946to 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>
947
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]948<p>A method with a related result type can be declared by using the
949type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
950is a contextual keyword that is only permitted in the result type of
951an Objective-C method, e.g.</p>
952
953<pre>
954@interface A
955+ (<b>instancetype</b>)constructAnA;
956@end
957</pre>
958
959<p>The related result type can also be inferred for some methods.
960To determine whether a method has an inferred related result type, the first
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]961word in the camel-case selector (e.g., "init" in "initWithObjects") is
Douglas Gregor8a0ace62011-11-03 18:33:01 +0000[diff] [blame]962considered, and the method will have a related result type if its return
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]963type is compatible with the type of its class and if</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]964
965<ul>
966
967 <li>the first word is "alloc" or "new", and the method is a class
968 method, or</li>
969
970 <li>the first word is "autorelease", "init", "retain", or "self",
971 and the method is an instance method.</li>
972
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]973</ul>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]974
975<p>If a method with a related result type is overridden by a subclass
976method, the subclass method must also return a type that is compatible
977with the subclass type. For example:</p>
978
979<blockquote>
980<pre>
981@interface NSString : NSObject
982- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
983@end
984</pre>
985</blockquote>
986
987<p>Related result types only affect the type of a message send or
988property access via the given method. In all other respects, a method
Douglas Gregore97179c2011-09-08 01:46:34 +0000[diff] [blame]989with a related result type is treated the same way as method that
990returns <tt>id</tt>.</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]991
Douglas Gregoraebb6532011-09-08 17:19:31 +0000[diff] [blame]992<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
993the <tt>instancetype</tt> contextual keyword is available.</p>
994
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]995<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]996<h2 id="objc_arc">Automatic reference counting </h2>
997<!-- ======================================================================= -->
998
999<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>
1000
1001<!-- ======================================================================= -->
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1002<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
1003<!-- ======================================================================= -->
1004
David Blaikie5090e9f2011-10-18 05:49:30 +0000[diff] [blame]1005<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor5471bc82011-09-08 17:18:35 +0000[diff] [blame]1006underlying type within Objective-C. For example, one can write an
1007enumeration type as:</p>
1008
1009<pre>
1010typedef enum : unsigned char { Red, Green, Blue } Color;
1011</pre>
1012
1013<p>This specifies that the underlying type, which is used to store the
1014enumeration value, is <tt>unsigned char</tt>.</p>
1015
1016<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
1017support for fixed underlying types is available in Objective-C.</p>
1018
1019<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1020<h2 id="overloading-in-c">Function Overloading in C</h2>
1021<!-- ======================================================================= -->
1022
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1023<p>Clang provides support for C++ function overloading in C. Function
1024overloading in C is introduced using the <tt>overloadable</tt> attribute. For
1025example, one might provide several overloaded versions of a <tt>tgsin</tt>
1026function that invokes the appropriate standard function computing the sine of a
1027value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
1028precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1029
1030<blockquote>
1031<pre>
1032#include &lt;math.h&gt;
1033float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
1034double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
1035long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
1036</pre>
1037</blockquote>
1038
1039<p>Given these declarations, one can call <tt>tgsin</tt> with a
1040<tt>float</tt> value to receive a <tt>float</tt> result, with a
1041<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
1042overloading in C follows the rules of C++ function overloading to pick
1043the best overload given the call arguments, with a few C-specific
1044semantics:</p>
1045<ul>
1046 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
1047 double</tt> is ranked as a floating-point promotion (per C99) rather
1048 than as a floating-point conversion (as in C++).</li>
1049
1050 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
1051 <tt>U*</tt> is considered a pointer conversion (with conversion
1052 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
1053
1054 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
1055 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
1056 conversion is given "conversion" rank.</li>
1057</ul>
1058
1059<p>The declaration of <tt>overloadable</tt> functions is restricted to
1060function declarations and definitions. Most importantly, if any
1061function with a given name is given the <tt>overloadable</tt>
1062attribute, then all function declarations and definitions with that
1063name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]1064attribute. This rule even applies to redeclarations of functions whose original
1065declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1066
1067<blockquote>
1068<pre>
1069int f(int) __attribute__((overloadable));
1070float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
1071
1072int g(int) __attribute__((overloadable));
1073int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
1074</pre>
1075</blockquote>
1076
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1077<p>Functions marked <tt>overloadable</tt> must have
1078prototypes. Therefore, the following code is ill-formed:</p>
1079
1080<blockquote>
1081<pre>
1082int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
1083</pre>
1084</blockquote>
1085
1086<p>However, <tt>overloadable</tt> functions are allowed to use a
1087ellipsis 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>
1088
1089<blockquote>
1090<pre>
Chris Lattner02246802009-02-18 22:27:46 +0000[diff] [blame]1091void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]1092</pre>
1093</blockquote>
1094
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1095<p>Functions declared with the <tt>overloadable</tt> attribute have
1096their names mangled according to the same rules as C++ function
1097names. For example, the three <tt>tgsin</tt> functions in our
1098motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +0000[diff] [blame]1099<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1100caveats to this use of name mangling:</p>
1101
1102<ul>
1103
1104 <li>Future versions of Clang may change the name mangling of
1105 functions overloaded in C, so you should not depend on an specific
1106 mangling. To be completely safe, we strongly urge the use of
1107 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
1108
1109 <li>The <tt>overloadable</tt> attribute has almost no meaning when
1110 used in C++, because names will already be mangled and functions are
1111 already overloadable. However, when an <tt>overloadable</tt>
1112 function occurs within an <tt>extern "C"</tt> linkage specification,
1113 it's name <i>will</i> be mangled in the same way as it would in
1114 C.</li>
1115</ul>
1116
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]1117<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1118
Eli Friedman0c706c22011-09-19 23:17:44 +0000[diff] [blame]1119<!-- ======================================================================= -->
1120<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
1121<!-- ======================================================================= -->
1122
1123<p>clang supports an extension which allows the following in C:</p>
1124
1125<blockquote>
1126<pre>
1127#include &lt;math.h&gt;
1128#include &lt;complex.h&gt;
1129complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
1130</pre>
1131</blockquote>
1132
1133<p>This construct is useful because there is no way to separately
1134initialize the real and imaginary parts of a complex variable in
1135standard C, given that clang does not support <code>_Imaginary</code>.
1136(clang also supports the <code>__real__</code> and <code>__imag__</code>
1137extensions from gcc, which help in some cases, but are not usable in
1138static initializers.)
1139
1140<p>Note that this extension does not allow eliding the braces; the
1141meaning of the following two lines is different:</p>
1142
1143<blockquote>
1144<pre>
1145complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
1146complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
1147</pre>
1148</blockquote>
1149
1150<p>This extension also works in C++ mode, as far as that goes, but does not
1151 apply to the C++ <code>std::complex</code>. (In C++11, list
1152 initialization allows the same syntax to be used with
1153 <code>std::complex</code> with the same meaning.)
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1154
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]1155<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1156<h2 id="builtins">Builtin Functions</h2>
1157<!-- ======================================================================= -->
1158
1159<p>Clang supports a number of builtin library functions with the same syntax as
1160GCC, including things like <tt>__builtin_nan</tt>,
1161<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
1162<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
1163addition to the GCC builtins, Clang supports a number of builtins that GCC does
1164not, which are listed here.</p>
1165
1166<p>Please note that Clang does not and will not support all of the GCC builtins
1167for vector operations. Instead of using builtins, you should use the functions
1168defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
1169define portable wrappers for these. Many of the Clang versions of these
1170functions are implemented directly in terms of <a href="#vectors">extended
1171vector support</a> instead of builtins, in order to reduce the number of
1172builtins that we need to implement.</p>
1173
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1174<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1175<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1176<!-- ======================================================================= -->
1177
Chris Lattneraad826b2009-09-16 18:56:12 +0000[diff] [blame]1178<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1179permutation/shuffle/swizzle operations. This builtin is also very important for
1180the implementation of various target-specific header files like
1181<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1182</p>
1183
1184<p><b>Syntax:</b></p>
1185
1186<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1187__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1188</pre>
1189
1190<p><b>Examples:</b></p>
1191
1192<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1193 // Identity operation - return 4-element vector V1.
1194 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1195
1196 // "Splat" element 0 of V1 into a 4-element result.
1197 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1198
1199 // Reverse 4-element vector V1.
1200 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1201
1202 // Concatenate every other element of 4-element vectors V1 and V2.
1203 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1204
1205 // Concatenate every other element of 8-element vectors V1 and V2.
1206 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1207</pre>
1208
1209<p><b>Description:</b></p>
1210
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1211<p>The first two arguments to __builtin_shufflevector are vectors that have the
1212same element type. The remaining arguments are a list of integers that specify
1213the elements indices of the first two vectors that should be extracted and
1214returned in a new vector. These element indices are numbered sequentially
1215starting with the first vector, continuing into the second vector. Thus, if
1216vec1 is a 4-element vector, index 5 would refer to the second element of vec2.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1217</p>
1218
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]1219<p>The result of __builtin_shufflevector is a vector
1220with the same element type as vec1/vec2 but that has an element count equal to
1221the number of indices specified.
1222</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1223
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1224<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1225
1226<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1227<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1228<!-- ======================================================================= -->
1229
1230<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1231the program cannot be reached, even if the compiler might otherwise think it
1232can. This is useful to improve optimization and eliminates certain warnings.
1233For example, without the <tt>__builtin_unreachable</tt> in the example below,
1234the compiler assumes that the inline asm can fall through and prints a "function
1235declared 'noreturn' should not return" warning.
1236</p>
1237
1238<p><b>Syntax:</b></p>
1239
1240<pre>
1241__builtin_unreachable()
1242</pre>
1243
1244<p><b>Example of Use:</b></p>
1245
1246<pre>
1247void myabort(void) __attribute__((noreturn));
1248void myabort(void) {
1249 asm("int3");
1250 __builtin_unreachable();
1251}
1252</pre>
1253
1254<p><b>Description:</b></p>
1255
1256<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1257it has undefined behavior, it is a statement that it is never reached and the
1258optimizer can take advantage of this to produce better code. This builtin takes
1259no arguments and produces a void result.
1260</p>
1261
1262<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1263
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1264<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1265<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1266<!-- ======================================================================= -->
1267
1268<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1269memory.
1270</p>
1271
1272<p><b>Syntax:</b></p>
1273
1274<pre>
1275<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1276</pre>
1277
1278<p><b>Example of Use:</b></p>
1279
1280<pre>
Sean Hunt7e98b472011-06-23 01:21:01 +0000[diff] [blame]1281int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]1282</pre>
1283
1284<p><b>Description:</b></p>
1285
1286<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1287intrinsics to allow code to atomically swap the current value with the new
1288value. More importantly, it helps developers write more efficient and correct
1289code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1290relying on the platform specific implementation details of
1291__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1292</p>
1293
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]1294
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1295<!-- ======================================================================= -->
1296<h2 id="targetspecific">Target-Specific Extensions</h2>
1297<!-- ======================================================================= -->
1298
1299<p>Clang supports some language features conditionally on some targets.</p>
1300
1301<!-- ======================================================================= -->
1302<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1303<!-- ======================================================================= -->
1304
1305<p>The X86 backend has these language extensions:</p>
1306
1307<!-- ======================================================================= -->
1308<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1309<!-- ======================================================================= -->
1310
1311<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +0000[diff] [blame]1312relative to the X86 GS segment register, and address space #257 causes it to be
1313relative to the X86 FS segment. Note that this is a very very low-level
1314feature that should only be used if you know what you're doing (for example in
1315an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]1316
1317<p>Here is an example:</p>
1318
1319<pre>
1320#define GS_RELATIVE __attribute__((address_space(256)))
1321int foo(int GS_RELATIVE *P) {
1322 return *P;
1323}
1324</pre>
1325
1326<p>Which compiles to (on X86-32):</p>
1327
1328<pre>
1329_foo:
1330 movl 4(%esp), %eax
1331 movl %gs:(%eax), %eax
1332 ret
1333</pre>
1334
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1335<!-- ======================================================================= -->
1336<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1337<!-- ======================================================================= -->
1338
1339<p>Clang supports additional attributes that are useful for documenting program
1340invariants and rules for static analysis tools. The extensions documented here
1341are used by the <a
1342href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1343engine</a> that is part of Clang's Analysis library.</p>
1344
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1345<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1346
1347<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +0000[diff] [blame]1348attribute. This attribute, which is typically affixed to a function prototype,
1349indicates that a call to a given function never returns. Function prototypes for
1350common functions like <tt>exit</tt> are typically annotated with this attribute,
1351as well as a variety of common assertion handlers. Users can educate the static
1352analyzer about their own custom assertion handles (thus cutting down on false
1353positives due to false paths) by marking their own &quot;panic&quot; functions
1354with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1355
1356<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +0000[diff] [blame]1357there are special functions that for all intents and purposes should be
1358considered panic functions (i.e., they are only called when an internal program
1359error occurs) but may actually return so that the program can fail gracefully.
1360The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1361as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]1362pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1363<tt>noreturn</tt>).</p>
1364
1365<p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]1366same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1367placed at the end of function prototypes:</p>
1368
1369<pre>
1370 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1371</pre>
1372
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1373<p>Query for this feature with
1374<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1375
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1376<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1377
1378<p>Many methods in Objective-C have conventional meanings determined
1379by their selectors. For the purposes of static analysis, it is
1380sometimes useful to be able to mark a method as having a particular
1381conventional meaning despite not having the right selector, or as not
1382having the conventional meaning that its selector would suggest.
1383For these use cases, we provide an attribute to specifically describe
1384the <q>method family</q> that a method belongs to.</p>
1385
1386<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1387where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1388<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1389can only be placed at the end of a method declaration:</p>
1390
1391<pre>
1392 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1393</pre>
1394
1395<p>Users who do not wish to change the conventional meaning of a
1396method, and who merely want to document its non-standard retain and
1397release semantics, should use the
1398<a href="#attr_retain_release">retaining behavior attributes</a>
1399described below.</p>
1400
1401<p>Query for this feature with
1402<tt>__has_attribute(objc_method_family)</tt>.</p>
1403
1404<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1405
1406<p>In Objective-C, functions and methods are generally assumed to take
1407and return objects with +0 retain counts, with some exceptions for
1408special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1409there are exceptions, and so Clang provides attributes to allow these
1410exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1411ignore non-leaks). Some exceptions may be better described using
1412the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1413attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1414
1415<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1416<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1417and <tt>cf_returns_not_retained</tt> attributes can be placed on
1418methods and functions that return Objective-C or CoreFoundation
1419objects. They are commonly placed at the end of a function prototype
1420or method declaration:</p>
1421
1422<pre>
1423 id foo() <b>__attribute__((ns_returns_retained))</b>;
1424
1425 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1426</pre>
1427
1428<p>The <tt>*_returns_retained</tt> attributes specify that the
1429returned object has a +1 retain count.
1430The <tt>*_returns_not_retained</tt> attributes specify that the return
1431object has a +0 retain count, even if the normal convention for its
1432selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1433returned object is +0, but is guaranteed to live at least as long as the
1434next flush of an autorelease pool.</p>
1435
1436<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1437attributes can be placed on an parameter declaration; they specify
1438that the argument is expected to have a +1 retain count, which will be
1439balanced in some way by the function or method.
1440The <tt>ns_consumes_self</tt> attribute can only be placed on an
1441Objective-C method; it specifies that the method expects
1442its <tt>self</tt> parameter to have a +1 retain count, which it will
1443balance in some way.</p>
1444
1445<pre>
1446 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1447
1448 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1449 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1450</pre>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1451
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1452<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1453<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1454
Kostya Serebryanyce98c9b2011-11-28 20:51:02 +0000[diff] [blame]1455<!-- ======================================================================= -->
1456<h2 id="dynamicanalyzerspecific">Dynamic Analysis-Specific Extensions</h2>
1457<!-- ======================================================================= -->
1458<h3 id="address_sanitizer">AddressSanitizer</h3>
1459<p> Use <code>__has_feature(address_sanitizer)</code>
1460to check if the code is being built with <a
1461 href="AddressSanitizer.html">AddressSanitizer</a>.
1462</p>
Kostya Serebryany71efba02012-01-24 19:25:38 +0000[diff] [blame]1463<p>Use <tt>__attribute__((no_address_safety_analysis))</tt> on a function
1464declaration to specify that address safety instrumentation (e.g.
1465AddressSanitizer) should not be applied to that function.
1466</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1467
1468<!-- ======================================================================= -->
Caitlin Sadowski73cbbc82011-07-28 18:38:36 +0000[diff] [blame]1469<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1470<!-- ======================================================================= -->
1471
1472<p>Clang supports additional attributes for checking basic locking policies in
1473multithreaded programs.
1474Clang currently parses the following list of attributes, although
1475<b>the implementation for these annotations is currently in development.</b>
1476For more details, see the
1477<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1478</p>
1479
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1480<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1481
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1482<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1483declaration to specify that the thread safety analysis should not be run on that
1484function. This attribute provides an escape hatch (e.g. for situations when it
1485is difficult to annotate the locking policy). </p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1486
1487<h4 id="ts_lockable">lockable</h4>
1488
1489<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1490that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1491used to check consistency.</p>
1492
1493<h4 id="ts_scopedlockable">scoped_lockable</h4>
1494
1495<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1496specify that it has a "scoped" lockable type. Objects of this type will acquire
1497the lock upon construction and release it upon going out of scope.
1498 This annotation is primarily used to check
1499consistency.</p>
1500
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1501<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1502
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1503<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1504specify that the variable must be accessed while holding some lock.</p>
1505
1506<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1507
1508<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1509specify that the pointer must be dereferenced while holding some lock.</p>
1510
1511<h4 id="ts_guardedby">guarded_by(l)</h4>
1512
1513<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1514specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1515
1516<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1517
1518<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1519specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1520
1521<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1522
1523<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1524of a lockable variable to specify that the lock must be acquired before all
1525attribute arguments. Arguments must be lockable type, and there must be at
1526least one argument.</p>
1527
1528<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1529
1530<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1531of a lockable variable to specify that the lock must be acquired after all
1532attribute arguments. Arguments must be lockable type, and there must be at
1533least one argument.</p>
1534
1535<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1536
1537<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1538declaration to specify that the function acquires all listed locks
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1539exclusively. This attribute takes zero or more arguments: either of lockable
1540type or integers indexing into function parameters of lockable type. If no
1541arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1542enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1543
1544<h4 id="ts_slf">shared_lock_function(...)</h4>
1545
1546<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1547declaration to specify that the function acquires all listed locks, although
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1548 the locks may be shared (e.g. read locks). This attribute takes zero or more
1549arguments: either of lockable type or integers indexing into function
1550parameters of lockable type. If no arguments are given, the acquired lock is
1551implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1552
1553<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1554
1555<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1556declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1557all listed locks exclusively. This attribute takes one or more arguments. The
1558first argument is an integer or boolean value specifying the return value of a
1559successful lock acquisition. The remaining arugments are either of lockable type
1560or integers indexing into function parameters of lockable type. If only one
1561argument is given, the acquired lock is implicitly <tt>this</tt> of the
1562enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1563
1564<h4 id="ts_stf">shared_trylock_function(...)</h4>
1565
1566<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1567declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1568all listed locks, although the locks may be shared (e.g. read locks). This
1569attribute takes one or more arguments. The first argument is an integer or
1570boolean value specifying the return value of a successful lock acquisition. The
1571remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1572function parameters of lockable type. If only one argument is given, the
1573acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1574
1575<h4 id="ts_uf">unlock_function(...)</h4>
1576
1577<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1578declaration to specify that the function release all listed locks. This
1579attribute takes zero or more arguments: either of lockable type or integers
1580indexing into function parameters of lockable type. If no arguments are given,
1581the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1582
1583<h4 id="ts_lr">lock_returned(l)</h4>
1584
1585<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1586declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1587must be of lockable type). This annotation is used to aid in resolving lock
1588expressions.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1589
1590<h4 id="ts_le">locks_excluded(...)</h4>
1591
1592<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowskib51e0312011-08-09 17:59:31 +0000[diff] [blame]1593to specify that the function must not be called with the listed locks. Arguments
1594must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000[diff] [blame]1595
1596<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1597
1598<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1599declaration to specify that the function must be called while holding the listed
1600exclusive locks. Arguments must be lockable type, and there must be at
1601least one argument.</p>
1602
1603<h4 id="ts_slr">shared_locks_required(...)</h4>
1604
1605<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1606declaration to specify that the function must be called while holding the listed
1607shared locks. Arguments must be lockable type, and there must be at
1608least one argument.</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000[diff] [blame]1609
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1610</div>
1611</body>
1612</html>