blob: 34e205613be25399b24ac0ebcfde23b60400587a [file] [log] [blame]
Sean Hunt7e98b472011-06-23 01:21:01 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<!-- Material used from: HTML 4.01 specs: http://www.w3.org/TR/html401/ -->
Chris Lattner5ce933f2009-02-09 08:46:11 +00004<html>
5<head>
Sean Hunt7e98b472011-06-23 01:21:01 +00006 <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
Eli Friedman0c706c22011-09-19 23:17:44 +00007 <title>Clang Language Extensions</title>
Sean Hunt64f857b2011-06-23 01:22:53 +00008 <link type="text/css" rel="stylesheet" href="../menu.css">
9 <link type="text/css" rel="stylesheet" href="../content.css">
Sean Hunt7e98b472011-06-23 01:21:01 +000010 <style type="text/css">
11 td {
12 vertical-align: top;
13 }
14 </style>
Chris Lattner5ce933f2009-02-09 08:46:11 +000015</head>
16<body>
17
18<!--#include virtual="../menu.html.incl"-->
19
20<div id="content">
21
22<h1>Clang Language Extensions</h1>
23
24<ul>
25<li><a href="#intro">Introduction</a></li>
Chris Lattner148772a2009-06-13 07:13:28 +000026<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompson92bd8c72009-11-02 22:28:12 +000027<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner81edc9f2009-04-13 02:45:46 +000028<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000029<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCall48209082010-11-08 19:48:17 +000030<li><a href="#deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> attributes</a></li>
31<li><a href="#attributes-on-enumerators">Attributes on enumerators</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +000032<li><a href="#checking_language_features">Checks for Standard Language Features</a>
Ted Kremenek22c34102009-12-03 02:05:57 +000033 <ul>
34 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
35 <li><a href="#cxx_rtti">C++ RTTI</a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +000036 </ul></li>
37<li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +000038 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +000039 <li><a href="#cxx0x">C++11</a>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +000040 <ul>
David Blaikie5090e9f2011-10-18 05:49:30 +000041 <li><a href="#cxx_access_control_sfinae">C++11 SFINAE includes
42 access control</a></li>
43 <li><a href="#cxx_alias_templates">C++11 alias templates</a></li>
44 <li><a href="#cxx_alignas">C++11 alignment specifiers</a></li>
45 <li><a href="#cxx_attributes">C++11 attributes</a></li>
46 <li><a href="#cxx_constexpr">C++11 generalized constant expressions</a></li>
47 <li><a href="#cxx_decltype">C++11 <tt>decltype()</tt></a></li>
48 <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 +000049 <li><a href="#cxx_defaulted_functions">C++11 defaulted functions</a></li>
David Blaikie5090e9f2011-10-18 05:49:30 +000050 <li><a href="#cxx_delegating_constructor">C++11 delegating constructors</a></li>
51 <li><a href="#cxx_deleted_functions">C++11 deleted functions</a></li>
52 <li><a href="#cxx_explicit_conversions">C++11 explicit conversion functions</a></li>
53 <li><a href="#cxx_generalized_initializers">C++11 generalized initializers</a></li>
54 <li><a href="#cxx_implicit_moves">C++11 implicit move constructors/assignment operators</a></li>
55 <li><a href="#cxx_inheriting_constructors">C++11 inheriting constructors</a></li>
56 <li><a href="#cxx_inline_namespaces">C++11 inline namespaces</a></li>
57 <li><a href="#cxx_lambdas">C++11 lambdas</a></li>
58 <li><a href="#cxx_noexcept">C++11 noexcept specification</a></li>
59 <li><a href="#cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</a></li>
60 <li><a href="#cxx_nullptr">C++11 nullptr</a></li>
61 <li><a href="#cxx_override_control">C++11 override control</a></li>
62 <li><a href="#cxx_range_for">C++11 range-based for loop</a></li>
63 <li><a href="#cxx_raw_string_literals">C++11 raw string literals</a></li>
64 <li><a href="#cxx_rvalue_references">C++11 rvalue references</a></li>
65 <li><a href="#cxx_reference_qualified_functions">C++11 reference-qualified functions</a></li>
66 <li><a href="#cxx_static_assert">C++11 <tt>static_assert()</tt></a></li>
67 <li><a href="#cxx_auto_type">C++11 type inference</a></li>
68 <li><a href="#cxx_strong_enums">C++11 strongly-typed enumerations</a></li>
69 <li><a href="#cxx_trailing_return">C++11 trailing return type</a></li>
70 <li><a href="#cxx_unicode_literals">C++11 Unicode string literals</a></li>
71 <li><a href="#cxx_unrestricted_unions">C++11 unrestricted unions</a></li>
72 <li><a href="#cxx_user_literals">C++11 user-defined literals</a></li>
73 <li><a href="#cxx_variadic_templates">C++11 variadic templates</a></li>
Douglas Gregorece38942011-08-29 17:28:38 +000074 </ul></li>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +000075 <li><a href="#c1x">C1X</a>
76 <ul>
Peter Collingbournefd5f6862011-10-14 23:44:46 +000077 <li><a href="#c_alignas">C1X alignment specifiers</a></li>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +000078 <li><a href="#c_generic_selections">C1X generic selections</a></li>
79 <li><a href="#c_static_assert">C1X <tt>_Static_assert()</tt></a></li>
Sean Hunt7e98b472011-06-23 01:21:01 +000080 </ul></li>
81 </ul> </li>
Douglas Gregorafdf1372011-02-03 21:57:35 +000082<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000083<li><a href="#blocks">Blocks</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +000084<li><a href="#objc_features">Objective-C Features</a>
85 <ul>
86 <li><a href="#objc_instancetype">Related result types</a></li>
John McCallf85e1932011-06-15 23:02:42 +000087 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor5471bc82011-09-08 17:18:35 +000088 <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +000089 </ul>
90</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +000091<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Eli Friedman0c706c22011-09-19 23:17:44 +000092<li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +000093<li><a href="#builtins">Builtin Functions</a>
94 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +000095 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +000096 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +000097 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +000098 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +000099</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000100<li><a href="#targetspecific">Target-Specific Extensions</a>
101 <ul>
102 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
103 </ul>
104</li>
John McCall87494012011-03-18 03:51:49 +0000105<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000106<li><a href="#threadsafety">Thread Safety Annotation Checking</a></li>
107 <ul>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000108 <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li>
109 <li><a href="#ts_lockable"><tt>lockable</tt></a></li>
110 <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000111 <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li>
112 <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +0000113 <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li>
114 <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li>
115 <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li>
116 <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li>
117 <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li>
118 <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li>
119 <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li>
120 <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li>
121 <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li>
122 <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li>
123 <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li>
124 <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li>
125 <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +0000126 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000127</ul>
128
Chris Lattner5ce933f2009-02-09 08:46:11 +0000129<!-- ======================================================================= -->
130<h2 id="intro">Introduction</h2>
131<!-- ======================================================================= -->
132
133<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000134addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000135range of GCC extensions. Please see the <a
136href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
137more information on these extensions.</p>
138
139<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000140<h2 id="feature_check">Feature Checking Macros</h2>
141<!-- ======================================================================= -->
142
143<p>Language extensions can be very useful, but only if you know you can depend
Chris Lattnerc70e1932011-03-21 16:25:11 +0000144on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000145function-like macros. This allows you to directly test for a feature in your
146code without having to resort to something like autoconf or fragile "compiler
147version checks".</p>
148
149<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000150<h3><a name="__has_builtin">__has_builtin</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000151<!-- ======================================================================= -->
152
153<p>This function-like macro takes a single identifier argument that is the name
154of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
155not. It can be used like this:</p>
156
157<blockquote>
158<pre>
159#ifndef __has_builtin // Optional of course.
160 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
161#endif
162
163...
164#if __has_builtin(__builtin_trap)
165 __builtin_trap();
166#else
167 abort();
168#endif
169...
170</pre>
171</blockquote>
172
173
174<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000175<h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000176<!-- ======================================================================= -->
177
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000178<p>These function-like macros take a single identifier argument that is the
179name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
180is both supported by Clang and standardized in the current language standard
181or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
182<code>__has_extension</code> evaluates to 1 if the feature is supported by
183Clang in the current language (either as a language extension or a standard
184language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000185
186<blockquote>
187<pre>
188#ifndef __has_feature // Optional of course.
189 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
190#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000191#ifndef __has_extension
192 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
193#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000194
195...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000196#if __has_feature(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000197// This code will only be compiled with the -std=c++11 and -std=gnu++11
198// options, because rvalue references are only standardized in C++11.
Chris Lattner148772a2009-06-13 07:13:28 +0000199#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000200
201#if __has_extension(cxx_rvalue_references)
David Blaikie5090e9f2011-10-18 05:49:30 +0000202// This code will be compiled with the -std=c++11, -std=gnu++11, -std=c++98
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000203// and -std=gnu++98 options, because rvalue references are supported as a
204// language extension in C++98.
205#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000206</pre>
207</blockquote>
208
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000209<p id="has_feature_back_compat">For backwards compatibility reasons,
210<code>__has_feature</code> can also be used to test for support for
211non-standardized features, i.e. features not prefixed <code>c_</code>,
212<code>cxx_</code> or <code>objc_</code>.</p>
213
214<p>If the <code>-pedantic-errors</code> option is given,
215<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
216
Chris Lattner148772a2009-06-13 07:13:28 +0000217<p>The feature tag is described along with the language feature below.</p>
218
John Thompson92bd8c72009-11-02 22:28:12 +0000219<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000220<h3><a name="__has_attribute">__has_attribute</a></h3>
Anders Carlssoncae50952010-10-20 02:31:43 +0000221<!-- ======================================================================= -->
222
223<p>This function-like macro takes a single identifier argument that is the name
224of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
225can be used like this:</p>
226
227<blockquote>
228<pre>
229#ifndef __has_attribute // Optional of course.
230 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
231#endif
232
233...
Anders Carlsson961003d2011-01-24 03:54:51 +0000234#if __has_attribute(always_inline)
235#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000236#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000237#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000238#endif
239...
240</pre>
241</blockquote>
242
243<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000244<h2 id="has_include">Include File Checking Macros</h2>
245<!-- ======================================================================= -->
246
247<p>Not all developments systems have the same include files.
248The <a href="#__has_include">__has_include</a> and
249<a href="#__has_include_next">__has_include_next</a> macros allow you to
250check for the existence of an include file before doing
251a possibly failing #include directive.</p>
252
253<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000254<h3><a name="__has_include">__has_include</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000255<!-- ======================================================================= -->
256
257<p>This function-like macro takes a single file name string argument that
258is the name of an include file. It evaluates to 1 if the file can
259be found using the include paths, or 0 otherwise:</p>
260
261<blockquote>
262<pre>
263// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000264#if __has_include("myinclude.h") &amp;&amp; __has_include(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000265# include "myinclude.h"
266#endif
267
268// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000269#if defined(__has_include) &amp;&amp; __has_include("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000270# include "myinclude.h"
271#endif
272</pre>
273</blockquote>
274
275<p>To test for this feature, use #if defined(__has_include).</p>
276
277<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000278<h3><a name="__has_include_next">__has_include_next</a></h3>
John Thompson92bd8c72009-11-02 22:28:12 +0000279<!-- ======================================================================= -->
280
281<p>This function-like macro takes a single file name string argument that
282is the name of an include file. It is like __has_include except that it
283looks for the second instance of the given file found in the include
284paths. It evaluates to 1 if the second instance of the file can
285be found using the include paths, or 0 otherwise:</p>
286
287<blockquote>
288<pre>
289// Note the two possible file name string formats.
Sean Hunt7e98b472011-06-23 01:21:01 +0000290#if __has_include_next("myinclude.h") &amp;&amp; __has_include_next(&lt;stdint.h&gt;)
John Thompson92bd8c72009-11-02 22:28:12 +0000291# include_next "myinclude.h"
292#endif
293
294// To avoid problem with non-clang compilers not having this macro.
Sean Hunt7e98b472011-06-23 01:21:01 +0000295#if defined(__has_include_next) &amp;&amp; __has_include_next("myinclude.h")
John Thompson92bd8c72009-11-02 22:28:12 +0000296# include_next "myinclude.h"
297#endif
298</pre>
299</blockquote>
300
301<p>Note that __has_include_next, like the GNU extension
302#include_next directive, is intended for use in headers only,
303and will issue a warning if used in the top-level compilation
304file. A warning will also be issued if an absolute path
305is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000306
Ted Kremenekd7681502011-10-12 19:46:30 +0000307
308<!-- ======================================================================= -->
309<h3><a name="__has_warning">__has_warning</a></h3>
310<!-- ======================================================================= -->
311
312<p>This function-like macro takes a string literal that represents a command
313 line option for a warning and returns true if that is a valid warning
314 option.</p>
315
316<blockquote>
317<pre>
318#if __has_warning("-Wformat")
319...
320#endif
321</pre>
322</blockquote>
323
Chris Lattner148772a2009-06-13 07:13:28 +0000324<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000325<h2 id="builtinmacros">Builtin Macros</h2>
326<!-- ======================================================================= -->
327
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000328<dl>
329 <dt><code>__BASE_FILE__</code></dt>
330 <dd>Defined to a string that contains the name of the main input
331 file passed to Clang.</dd>
332
333 <dt><code>__COUNTER__</code></dt>
334 <dd>Defined to an integer value that starts at zero and is
335 incremented each time the <code>__COUNTER__</code> macro is
336 expanded.</dd>
337
338 <dt><code>__INCLUDE_LEVEL__</code></dt>
339 <dd>Defined to an integral value that is the include depth of the
340 file currently being translated. For the main file, this value is
341 zero.</dd>
342
343 <dt><code>__TIMESTAMP__</code></dt>
344 <dd>Defined to the date and time of the last modification of the
345 current source file.</dd>
346
347 <dt><code>__clang__</code></dt>
348 <dd>Defined when compiling with Clang</dd>
349
350 <dt><code>__clang_major__</code></dt>
351 <dd>Defined to the major version number of Clang (e.g., the 2 in
352 2.0.1).</dd>
353
354 <dt><code>__clang_minor__</code></dt>
355 <dd>Defined to the minor version number of Clang (e.g., the 0 in
356 2.0.1).</dd>
357
358 <dt><code>__clang_patchlevel__</code></dt>
359 <dd>Defined to the patch level of Clang (e.g., the 1 in 2.0.1).</dd>
360
361 <dt><code>__clang_version__</code></dt>
362 <dd>Defined to a string that captures the Clang version, including
363 the Subversion tag or revision number, e.g., "1.5 (trunk
364 102332)".</dd>
365</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000366
367<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000368<h2 id="vectors">Vectors and Extended Vectors</h2>
369<!-- ======================================================================= -->
370
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000371<p>Supports the GCC vector extensions, plus some stuff like V[1].</p>
372
373<p>Also supports <tt>ext_vector</tt>, which additionally support for V.xyzw
374syntax and other tidbits as seen in OpenCL. An example is:</p>
375
376<blockquote>
377<pre>
378typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
379typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
380
381float4 foo(float2 a, float2 b) {
382 float4 c;
383 c.xz = a;
384 c.yw = b;
385 return c;
386}
John McCall48209082010-11-08 19:48:17 +0000387</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000388</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000389
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000390<p>Query for this feature with __has_extension(attribute_ext_vector_type).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000391
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000392<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
393
Chris Lattner5ce933f2009-02-09 08:46:11 +0000394<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000395<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000396<!-- ======================================================================= -->
397
John McCall48209082010-11-08 19:48:17 +0000398<p>An optional string message can be added to the <tt>deprecated</tt>
399and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000400
John McCall48209082010-11-08 19:48:17 +0000401<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000402<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000403</blockquote>
404
405<p>If the deprecated or unavailable declaration is used, the message
406will be incorporated into the appropriate diagnostic:</p>
407
408<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000409<pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!! [-Wdeprecated-declarations]
John McCall48209082010-11-08 19:48:17 +0000410 explode();
411 ^</pre>
412</blockquote>
413
414<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000415with <tt>__has_extension(attribute_deprecated_with_message)</tt>
416and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000417
418<!-- ======================================================================= -->
419<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
420<!-- ======================================================================= -->
421
422<p>Clang allows attributes to be written on individual enumerators.
423This allows enumerators to be deprecated, made unavailable, etc. The
424attribute must appear after the enumerator name and before any
425initializer, like so:</p>
426
427<blockquote>
428<pre>enum OperationMode {
429 OM_Invalid,
430 OM_Normal,
431 OM_Terrified __attribute__((deprecated)),
432 OM_AbortOnError __attribute__((deprecated)) = 4
433};</pre>
434</blockquote>
435
436<p>Attributes on the <tt>enum</tt> declaration do not apply to
437individual enumerators.</p>
438
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000439<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000440
441<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000442<h2 id="checking_language_features">Checks for Standard Language Features</h2>
443<!-- ======================================================================= -->
444
445<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
446enabled. Those features are listed here.</p>
447
Ted Kremenek22c34102009-12-03 02:05:57 +0000448<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000449
Ted Kremenek22c34102009-12-03 02:05:57 +0000450<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Sean Hunt647ba1b2011-06-23 00:42:53 +0000451example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000452
Ted Kremenek22c34102009-12-03 02:05:57 +0000453<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000454
Ted Kremenek0eb95602009-12-03 02:06:43 +0000455<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 +0000456compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000457
458<!-- ======================================================================= -->
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000459<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
460<!-- ======================================================================= -->
461
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000462<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
463to query if certain upcoming standard language features are enabled. Those
464features are listed here. Features that are not yet implemented will be
465noted.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000466
David Blaikie5090e9f2011-10-18 05:49:30 +0000467<h3 id="cxx0x">C++11</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000468
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000469<p>The features listed below are slated for inclusion in the upcoming
David Blaikie5090e9f2011-10-18 05:49:30 +0000470C++11 standard. As a result, all these features are enabled
471with the <tt>-std=c++11</tt> option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000472
David Blaikie5090e9f2011-10-18 05:49:30 +0000473<h4 id="cxx_access_control_sfinae">C++11 SFINAE includes access control</h4>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000474
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000475<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 +0000476
David Blaikie5090e9f2011-10-18 05:49:30 +0000477<h4 id="cxx_alias_templates">C++11 alias templates</h4>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000478
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000479<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
480<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
David Blaikie5090e9f2011-10-18 05:49:30 +0000481C++11's alias declarations and alias templates is enabled.</p>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000482
David Blaikie5090e9f2011-10-18 05:49:30 +0000483<h4 id="cxx_alignas">C++11 alignment specifiers</h4>
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000484
485<p>Use <tt>__has_feature(cxx_alignas)</tt> or
486<tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment
487specifiers using <tt>alignas</tt> is enabled.</p>
488
David Blaikie5090e9f2011-10-18 05:49:30 +0000489<h4 id="cxx_attributes">C++11 attributes</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000490
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000491<p>Use <tt>__has_feature(cxx_attributes)</tt> or
492<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
David Blaikie5090e9f2011-10-18 05:49:30 +0000493parsing with C++11's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000494
David Blaikie5090e9f2011-10-18 05:49:30 +0000495<h4 id="cxx_constexpr">C++11 generalized constant expressions</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000496
497<p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support
498for generalized constant expressions (e.g., <tt>constexpr</tt>) is
499enabled. Clang does not currently implement this feature.</p>
500
David Blaikie5090e9f2011-10-18 05:49:30 +0000501<h4 id="cxx_decltype">C++11 <tt>decltype()</tt></h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000502
503<p>Use <tt>__has_feature(cxx_decltype)</tt> or
504<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
505<tt>decltype()</tt> specifier is enabled.</p>
506
David Blaikie5090e9f2011-10-18 05:49:30 +0000507<h4 id="cxx_default_function_template_args">C++11 default template arguments in function templates</h4>
Douglas Gregor07508002011-02-05 20:35:30 +0000508
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000509<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
510<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
511if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000512
Douglas Gregorf695a692011-11-01 01:19:34 +0000513<h4 id="cxx_defaulted_functions">C++11 <tt>default</tt>ed functions</h4>
514
515<p>Use <tt>__has_feature(cxx_defaulted_functions)</tt> or
516<tt>__has_extension(cxx_defaulted_functions)</tt> to determine if support for
517defaulted function definitions (with <tt>= default</tt>) is enabled.</p>
518
David Blaikie5090e9f2011-10-18 05:49:30 +0000519<h4 id="cxx_delegating_constructors">C++11 delegating constructors</h4>
Sean Huntd9624992011-06-23 06:11:37 +0000520
521<p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if
522support for delegating constructors is enabled.</p>
523
David Blaikie5090e9f2011-10-18 05:49:30 +0000524<h4 id="cxx_deleted_functions">C++11 <tt>delete</tt>d functions</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000525
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000526<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
527<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000528deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000529
David Blaikie5090e9f2011-10-18 05:49:30 +0000530<h4 id="cxx_explicit_conversions">C++11 explicit conversion functions</h3>
Douglas Gregorece38942011-08-29 17:28:38 +0000531<p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p>
532
David Blaikie5090e9f2011-10-18 05:49:30 +0000533<h4 id="cxx_generalized_initializers">C++11 generalized initializers</h4>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000534
535<p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if
536support for generalized initializers (using braced lists and
Douglas Gregorece38942011-08-29 17:28:38 +0000537<tt>std::initializer_list</tt>) is enabled. Clang does not currently implement
538this feature.</p>
539
David Blaikie5090e9f2011-10-18 05:49:30 +0000540<h4 id="cxx_implicit_moves">C++11 implicit move constructors/assignment operators</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000541
Sebastian Redl72a81d22011-10-10 18:10:00 +0000542<p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will
543implicitly generate move constructors and move assignment operators where needed.</p>
Douglas Gregorece38942011-08-29 17:28:38 +0000544
David Blaikie5090e9f2011-10-18 05:49:30 +0000545<h4 id="cxx_inheriting_constructors">C++11 inheriting constructors</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000546
547<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>
548
David Blaikie5090e9f2011-10-18 05:49:30 +0000549<h4 id="cxx_inline_namespaces">C++11 inline namespaces</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000550
551<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
552<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
553inline namespaces is enabled.</p>
Sean Hunte1f6dea2011-08-07 00:34:32 +0000554
David Blaikie5090e9f2011-10-18 05:49:30 +0000555<h4 id="cxx_lambdas">C++11 lambdas</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000556
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000557<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
558<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
Douglas Gregorece38942011-08-29 17:28:38 +0000559is enabled. Clang does not currently implement this feature.</p>
560
David Blaikie5090e9f2011-10-18 05:49:30 +0000561<h4 id="cxx_noexcept">C++11 noexcept</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000562
563<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
564<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
565exception specifications is enabled.</p>
566
David Blaikie5090e9f2011-10-18 05:49:30 +0000567<h4 id="cxx_nonstatic_member_init">C++11 in-class non-static data member initialization</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000568
569<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 +0000570
David Blaikie5090e9f2011-10-18 05:49:30 +0000571<h4 id="cxx_nullptr">C++11 <tt>nullptr</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000572
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000573<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
574<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000575<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000576
David Blaikie5090e9f2011-10-18 05:49:30 +0000577<h4 id="cxx_override_control">C++11 <tt>override control</tt></h4>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000578
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000579<p>Use <tt>__has_feature(cxx_override_control)</tt> or
580<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000581the override control keywords is enabled.</p>
582
David Blaikie5090e9f2011-10-18 05:49:30 +0000583<h4 id="cxx_reference_qualified_functions">C++11 reference-qualified functions</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000584<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
585<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
586if support for reference-qualified functions (e.g., member functions with
587<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
588is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000589
David Blaikie5090e9f2011-10-18 05:49:30 +0000590<h4 id="cxx_range_for">C++11 range-based <tt>for</tt> loop</h4>
Richard Smitha391a462011-04-15 15:14:40 +0000591
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000592<p>Use <tt>__has_feature(cxx_range_for)</tt> or
593<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
594range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000595
David Blaikie5090e9f2011-10-18 05:49:30 +0000596<h4 id="cxx_raw_string_literals">C++11 raw string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000597<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>
598
David Blaikie5090e9f2011-10-18 05:49:30 +0000599<h4 id="cxx_rvalue_references">C++11 rvalue references</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000600
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000601<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
602<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000603rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000604
David Blaikie5090e9f2011-10-18 05:49:30 +0000605<h4 id="cxx_static_assert">C++11 <tt>static_assert()</tt></h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000606
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000607<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
608<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000609compile-time assertions using <tt>static_assert</tt> is enabled.</p>
610
David Blaikie5090e9f2011-10-18 05:49:30 +0000611<h4 id="cxx_auto_type">C++11 type inference</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000612
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000613<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
David Blaikie5090e9f2011-10-18 05:49:30 +0000614<tt>__has_extension(cxx_auto_type)</tt> to determine C++11 type inference is
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000615supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
616will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000617
David Blaikie5090e9f2011-10-18 05:49:30 +0000618<h4 id="cxx_strong_enums">C++11 strongly typed enumerations</h4>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000619
Douglas Gregorece38942011-08-29 17:28:38 +0000620<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
621<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
622strongly typed, scoped enumerations is enabled.</p>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000623
David Blaikie5090e9f2011-10-18 05:49:30 +0000624<h4 id="cxx_trailing_return">C++11 trailing return type</h4>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000625
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000626<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
627<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
628alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000629
David Blaikie5090e9f2011-10-18 05:49:30 +0000630<h4 id="cxx_unicode_literals">C++11 Unicode string literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000631<p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if
632support for Unicode string literals is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000633
David Blaikie5090e9f2011-10-18 05:49:30 +0000634<h4 id="cxx_unrestricted_unions">C++11 unrestricted unions</h4>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000635
Douglas Gregorece38942011-08-29 17:28:38 +0000636<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 +0000637
David Blaikie5090e9f2011-10-18 05:49:30 +0000638<h4 id="cxx_user_literals">C++11 user-defined literals</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000639
640<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>
641
David Blaikie5090e9f2011-10-18 05:49:30 +0000642<h4 id="cxx_variadic_templates">C++11 variadic templates</h4>
Douglas Gregorece38942011-08-29 17:28:38 +0000643
644<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
645<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
646for variadic templates is enabled.</p>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000647
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000648<h3 id="c1x">C1X</h3>
649
650<p>The features listed below are slated for inclusion in the upcoming
651C1X standard. As a result, all these features are enabled
652with the <tt>-std=c1x</tt> option when compiling C code.</p>
653
Peter Collingbournefd5f6862011-10-14 23:44:46 +0000654<h4 id="c_alignas">C1X alignment specifiers</h4>
655
656<p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt>
657to determine if support for alignment specifiers using <tt>_Alignas</tt>
658is enabled.</p>
659
Sean Hunt7e98b472011-06-23 01:21:01 +0000660<h4 id="c_generic_selections">C1X generic selections</h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000661
662<p>Use <tt>__has_feature(c_generic_selections)</tt> or
663<tt>__has_extension(c_generic_selections)</tt> to determine if support for
664generic selections is enabled.</p>
665
666<p>As an extension, the C1X generic selection expression is available in all
667languages supported by Clang. The syntax is the same as that given in the
668C1X draft standard.</p>
669
670<p>In C, type compatibility is decided according to the rules given in the
671appropriate standard, but in C++, which lacks the type compatibility rules
672used in C, types are considered compatible only if they are equivalent.</p>
673
Sean Hunt7e98b472011-06-23 01:21:01 +0000674<h4 id="c_static_assert">C1X <tt>_Static_assert()</tt></h4>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000675
676<p>Use <tt>__has_feature(c_static_assert)</tt> or
677<tt>__has_extension(c_static_assert)</tt> to determine if support for
678compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
679
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000680<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000681<h2 id="checking_type_traits">Checks for Type Traits</h2>
682<!-- ======================================================================= -->
683
Sean Hunt7e98b472011-06-23 01:21:01 +0000684<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 +0000685<blockquote>
686<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000687#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000688template&lt;typename From, typename To&gt;
689struct is_convertible_to {
690 static const bool value = __is_convertible_to(From, To);
691};
692#else
693// Emulate type trait
694#endif
695</pre>
696</blockquote>
697
698<p>The following type traits are supported by Clang:</p>
699<ul>
700 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
701 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
702 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
703 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
704 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
705 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
706 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
707 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
708 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
709 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
710 <li><code>__is_class</code> (GNU, Microsoft)</li>
711 <li><code>__is_convertible_to</code> (Microsoft)</li>
712 <li><code>__is_empty</code> (GNU, Microsoft)</li>
713 <li><code>__is_enum</code> (GNU, Microsoft)</li>
714 <li><code>__is_pod</code> (GNU, Microsoft)</li>
715 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
716 <li><code>__is_union</code> (GNU, Microsoft)</li>
717 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
David Blaikie5090e9f2011-10-18 05:49:30 +0000718 <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 Gregorafdf1372011-02-03 21:57:35 +0000719</ul>
720
721<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000722<h2 id="blocks">Blocks</h2>
723<!-- ======================================================================= -->
724
Chris Lattnera7dbdf52009-03-09 07:03:22 +0000725<p>The syntax and high level language feature description is in <a
726href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
727details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +0000728href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000729
Chris Lattner148772a2009-06-13 07:13:28 +0000730
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000731<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000732
Chris Lattner5ce933f2009-02-09 08:46:11 +0000733<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +0000734<h2 id="objc_features">Objective-C Features</h2>
735<!-- ======================================================================= -->
736
737<h3 id="objc_instancetype">Related result types</h3>
738
739<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>
740
741<blockquote>
742<pre>
743@interface NSObject
744+ (id)alloc;
745- (id)init;
746@end
747
748@interface NSArray : NSObject
749@end
750</pre>
751</blockquote>
752
753<p>and this common initialization pattern</p>
754
755<blockquote>
756<pre>
757NSArray *array = [[NSArray alloc] init];
758</pre>
759</blockquote>
760
761<p>the type of the expression <code>[NSArray alloc]</code> is
762<code>NSArray*</code> because <code>alloc</code> implicitly has a
763related result type. Similarly, the type of the expression
764<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
765<code>init</code> has a related result type and its receiver is known
766to 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>
767
Douglas Gregore97179c2011-09-08 01:46:34 +0000768<p>A method with a related result type can be declared by using the
769type <tt>instancetype</tt> as its result type. <tt>instancetype</tt>
770is a contextual keyword that is only permitted in the result type of
771an Objective-C method, e.g.</p>
772
773<pre>
774@interface A
775+ (<b>instancetype</b>)constructAnA;
776@end
777</pre>
778
779<p>The related result type can also be inferred for some methods.
780To determine whether a method has an inferred related result type, the first
Douglas Gregor926df6c2011-06-11 01:09:30 +0000781word in the camel-case selector (e.g., "init" in "initWithObjects") is
782considered, and the method will a related result type if its return
Sean Hunt7e98b472011-06-23 01:21:01 +0000783type is compatible with the type of its class and if</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000784
785<ul>
786
787 <li>the first word is "alloc" or "new", and the method is a class
788 method, or</li>
789
790 <li>the first word is "autorelease", "init", "retain", or "self",
791 and the method is an instance method.</li>
792
Sean Hunt7e98b472011-06-23 01:21:01 +0000793</ul>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000794
795<p>If a method with a related result type is overridden by a subclass
796method, the subclass method must also return a type that is compatible
797with the subclass type. For example:</p>
798
799<blockquote>
800<pre>
801@interface NSString : NSObject
802- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
803@end
804</pre>
805</blockquote>
806
807<p>Related result types only affect the type of a message send or
808property access via the given method. In all other respects, a method
Douglas Gregore97179c2011-09-08 01:46:34 +0000809with a related result type is treated the same way as method that
810returns <tt>id</tt>.</p>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000811
Douglas Gregoraebb6532011-09-08 17:19:31 +0000812<p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether
813the <tt>instancetype</tt> contextual keyword is available.</p>
814
Douglas Gregor926df6c2011-06-11 01:09:30 +0000815<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +0000816<h2 id="objc_arc">Automatic reference counting </h2>
817<!-- ======================================================================= -->
818
819<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>
820
821<!-- ======================================================================= -->
Douglas Gregor5471bc82011-09-08 17:18:35 +0000822<h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2>
823<!-- ======================================================================= -->
824
David Blaikie5090e9f2011-10-18 05:49:30 +0000825<p>Clang provides support for C++11 enumerations with a fixed
Douglas Gregor5471bc82011-09-08 17:18:35 +0000826underlying type within Objective-C. For example, one can write an
827enumeration type as:</p>
828
829<pre>
830typedef enum : unsigned char { Red, Green, Blue } Color;
831</pre>
832
833<p>This specifies that the underlying type, which is used to store the
834enumeration value, is <tt>unsigned char</tt>.</p>
835
836<p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether
837support for fixed underlying types is available in Objective-C.</p>
838
839<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +0000840<h2 id="overloading-in-c">Function Overloading in C</h2>
841<!-- ======================================================================= -->
842
Chris Lattnerf161d412009-02-13 21:51:45 +0000843<p>Clang provides support for C++ function overloading in C. Function
844overloading in C is introduced using the <tt>overloadable</tt> attribute. For
845example, one might provide several overloaded versions of a <tt>tgsin</tt>
846function that invokes the appropriate standard function computing the sine of a
847value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
848precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000849
850<blockquote>
851<pre>
852#include &lt;math.h&gt;
853float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
854double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
855long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
856</pre>
857</blockquote>
858
859<p>Given these declarations, one can call <tt>tgsin</tt> with a
860<tt>float</tt> value to receive a <tt>float</tt> result, with a
861<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
862overloading in C follows the rules of C++ function overloading to pick
863the best overload given the call arguments, with a few C-specific
864semantics:</p>
865<ul>
866 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
867 double</tt> is ranked as a floating-point promotion (per C99) rather
868 than as a floating-point conversion (as in C++).</li>
869
870 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
871 <tt>U*</tt> is considered a pointer conversion (with conversion
872 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
873
874 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
875 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
876 conversion is given "conversion" rank.</li>
877</ul>
878
879<p>The declaration of <tt>overloadable</tt> functions is restricted to
880function declarations and definitions. Most importantly, if any
881function with a given name is given the <tt>overloadable</tt>
882attribute, then all function declarations and definitions with that
883name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +0000884attribute. This rule even applies to redeclarations of functions whose original
885declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000886
887<blockquote>
888<pre>
889int f(int) __attribute__((overloadable));
890float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
891
892int g(int) __attribute__((overloadable));
893int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
894</pre>
895</blockquote>
896
Douglas Gregor965acbb2009-02-18 07:07:28 +0000897<p>Functions marked <tt>overloadable</tt> must have
898prototypes. Therefore, the following code is ill-formed:</p>
899
900<blockquote>
901<pre>
902int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
903</pre>
904</blockquote>
905
906<p>However, <tt>overloadable</tt> functions are allowed to use a
907ellipsis 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>
908
909<blockquote>
910<pre>
Chris Lattner02246802009-02-18 22:27:46 +0000911void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +0000912</pre>
913</blockquote>
914
Douglas Gregorcb54d432009-02-13 00:57:04 +0000915<p>Functions declared with the <tt>overloadable</tt> attribute have
916their names mangled according to the same rules as C++ function
917names. For example, the three <tt>tgsin</tt> functions in our
918motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +0000919<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +0000920caveats to this use of name mangling:</p>
921
922<ul>
923
924 <li>Future versions of Clang may change the name mangling of
925 functions overloaded in C, so you should not depend on an specific
926 mangling. To be completely safe, we strongly urge the use of
927 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
928
929 <li>The <tt>overloadable</tt> attribute has almost no meaning when
930 used in C++, because names will already be mangled and functions are
931 already overloadable. However, when an <tt>overloadable</tt>
932 function occurs within an <tt>extern "C"</tt> linkage specification,
933 it's name <i>will</i> be mangled in the same way as it would in
934 C.</li>
935</ul>
936
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000937<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000938
Eli Friedman0c706c22011-09-19 23:17:44 +0000939<!-- ======================================================================= -->
940<h2 id="complex-list-init">Initializer lists for complex numbers in C</h2>
941<!-- ======================================================================= -->
942
943<p>clang supports an extension which allows the following in C:</p>
944
945<blockquote>
946<pre>
947#include &lt;math.h&gt;
948#include &lt;complex.h&gt;
949complex float x = { 1.0f, INFINITY }; // Init to (1, Inf)
950</pre>
951</blockquote>
952
953<p>This construct is useful because there is no way to separately
954initialize the real and imaginary parts of a complex variable in
955standard C, given that clang does not support <code>_Imaginary</code>.
956(clang also supports the <code>__real__</code> and <code>__imag__</code>
957extensions from gcc, which help in some cases, but are not usable in
958static initializers.)
959
960<p>Note that this extension does not allow eliding the braces; the
961meaning of the following two lines is different:</p>
962
963<blockquote>
964<pre>
965complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1)
966complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0)
967</pre>
968</blockquote>
969
970<p>This extension also works in C++ mode, as far as that goes, but does not
971 apply to the C++ <code>std::complex</code>. (In C++11, list
972 initialization allows the same syntax to be used with
973 <code>std::complex</code> with the same meaning.)
Chris Lattner148772a2009-06-13 07:13:28 +0000974
Douglas Gregorcb54d432009-02-13 00:57:04 +0000975<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000976<h2 id="builtins">Builtin Functions</h2>
977<!-- ======================================================================= -->
978
979<p>Clang supports a number of builtin library functions with the same syntax as
980GCC, including things like <tt>__builtin_nan</tt>,
981<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
982<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
983addition to the GCC builtins, Clang supports a number of builtins that GCC does
984not, which are listed here.</p>
985
986<p>Please note that Clang does not and will not support all of the GCC builtins
987for vector operations. Instead of using builtins, you should use the functions
988defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
989define portable wrappers for these. Many of the Clang versions of these
990functions are implemented directly in terms of <a href="#vectors">extended
991vector support</a> instead of builtins, in order to reduce the number of
992builtins that we need to implement.</p>
993
Chris Lattner5ce933f2009-02-09 08:46:11 +0000994<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +0000995<h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000996<!-- ======================================================================= -->
997
Chris Lattneraad826b2009-09-16 18:56:12 +0000998<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +0000999permutation/shuffle/swizzle operations. This builtin is also very important for
1000the implementation of various target-specific header files like
1001<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +00001002</p>
1003
1004<p><b>Syntax:</b></p>
1005
1006<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +00001007__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +00001008</pre>
1009
1010<p><b>Examples:</b></p>
1011
1012<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +00001013 // Identity operation - return 4-element vector V1.
1014 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
1015
1016 // "Splat" element 0 of V1 into a 4-element result.
1017 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
1018
1019 // Reverse 4-element vector V1.
1020 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
1021
1022 // Concatenate every other element of 4-element vectors V1 and V2.
1023 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
1024
1025 // Concatenate every other element of 8-element vectors V1 and V2.
1026 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +00001027</pre>
1028
1029<p><b>Description:</b></p>
1030
Chris Lattner6f72da52009-02-13 20:00:20 +00001031<p>The first two arguments to __builtin_shufflevector are vectors that have the
1032same element type. The remaining arguments are a list of integers that specify
1033the elements indices of the first two vectors that should be extracted and
1034returned in a new vector. These element indices are numbered sequentially
1035starting with the first vector, continuing into the second vector. Thus, if
1036vec1 is a 4-element vector, index 5 would refer to the second element of vec2.
Chris Lattner5ce933f2009-02-09 08:46:11 +00001037</p>
1038
Chris Lattner6f72da52009-02-13 20:00:20 +00001039<p>The result of __builtin_shufflevector is a vector
1040with the same element type as vec1/vec2 but that has an element count equal to
1041the number of indices specified.
1042</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +00001043
Chris Lattner21190d52009-09-21 03:09:59 +00001044<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
1045
1046<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +00001047<h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3>
Chris Lattner21190d52009-09-21 03:09:59 +00001048<!-- ======================================================================= -->
1049
1050<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
1051the program cannot be reached, even if the compiler might otherwise think it
1052can. This is useful to improve optimization and eliminates certain warnings.
1053For example, without the <tt>__builtin_unreachable</tt> in the example below,
1054the compiler assumes that the inline asm can fall through and prints a "function
1055declared 'noreturn' should not return" warning.
1056</p>
1057
1058<p><b>Syntax:</b></p>
1059
1060<pre>
1061__builtin_unreachable()
1062</pre>
1063
1064<p><b>Example of Use:</b></p>
1065
1066<pre>
1067void myabort(void) __attribute__((noreturn));
1068void myabort(void) {
1069 asm("int3");
1070 __builtin_unreachable();
1071}
1072</pre>
1073
1074<p><b>Description:</b></p>
1075
1076<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
1077it has undefined behavior, it is a statement that it is never reached and the
1078optimizer can take advantage of this to produce better code. This builtin takes
1079no arguments and produces a void result.
1080</p>
1081
1082<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
1083
Chris Lattner23aa9c82011-04-09 03:57:26 +00001084<!-- ======================================================================= -->
Sean Hunt7e98b472011-06-23 01:21:01 +00001085<h3><a name="__sync_swap">__sync_swap</a></h3>
Chris Lattner23aa9c82011-04-09 03:57:26 +00001086<!-- ======================================================================= -->
1087
1088<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
1089memory.
1090</p>
1091
1092<p><b>Syntax:</b></p>
1093
1094<pre>
1095<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
1096</pre>
1097
1098<p><b>Example of Use:</b></p>
1099
1100<pre>
Sean Hunt7e98b472011-06-23 01:21:01 +00001101int old_value = __sync_swap(&amp;value, new_value);
Chris Lattner23aa9c82011-04-09 03:57:26 +00001102</pre>
1103
1104<p><b>Description:</b></p>
1105
1106<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
1107intrinsics to allow code to atomically swap the current value with the new
1108value. More importantly, it helps developers write more efficient and correct
1109code by avoiding expensive loops around __sync_bool_compare_and_swap() or
1110relying on the platform specific implementation details of
1111__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
1112</p>
1113
Chris Lattner21190d52009-09-21 03:09:59 +00001114
Chris Lattner1177f912009-04-09 19:58:15 +00001115<!-- ======================================================================= -->
1116<h2 id="targetspecific">Target-Specific Extensions</h2>
1117<!-- ======================================================================= -->
1118
1119<p>Clang supports some language features conditionally on some targets.</p>
1120
1121<!-- ======================================================================= -->
1122<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
1123<!-- ======================================================================= -->
1124
1125<p>The X86 backend has these language extensions:</p>
1126
1127<!-- ======================================================================= -->
1128<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
1129<!-- ======================================================================= -->
1130
1131<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +00001132relative to the X86 GS segment register, and address space #257 causes it to be
1133relative to the X86 FS segment. Note that this is a very very low-level
1134feature that should only be used if you know what you're doing (for example in
1135an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +00001136
1137<p>Here is an example:</p>
1138
1139<pre>
1140#define GS_RELATIVE __attribute__((address_space(256)))
1141int foo(int GS_RELATIVE *P) {
1142 return *P;
1143}
1144</pre>
1145
1146<p>Which compiles to (on X86-32):</p>
1147
1148<pre>
1149_foo:
1150 movl 4(%esp), %eax
1151 movl %gs:(%eax), %eax
1152 ret
1153</pre>
1154
Ted Kremeneked869312009-04-10 05:03:33 +00001155<!-- ======================================================================= -->
1156<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
1157<!-- ======================================================================= -->
1158
1159<p>Clang supports additional attributes that are useful for documenting program
1160invariants and rules for static analysis tools. The extensions documented here
1161are used by the <a
1162href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
1163engine</a> that is part of Clang's Analysis library.</p>
1164
John McCall87494012011-03-18 03:51:49 +00001165<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +00001166
1167<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +00001168attribute. This attribute, which is typically affixed to a function prototype,
1169indicates that a call to a given function never returns. Function prototypes for
1170common functions like <tt>exit</tt> are typically annotated with this attribute,
1171as well as a variety of common assertion handlers. Users can educate the static
1172analyzer about their own custom assertion handles (thus cutting down on false
1173positives due to false paths) by marking their own &quot;panic&quot; functions
1174with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +00001175
1176<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +00001177there are special functions that for all intents and purposes should be
1178considered panic functions (i.e., they are only called when an internal program
1179error occurs) but may actually return so that the program can fail gracefully.
1180The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
1181as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +00001182pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +00001183<tt>noreturn</tt>).</p>
1184
1185<p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the
Chris Lattner28935892009-04-10 05:54:56 +00001186same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +00001187placed at the end of function prototypes:</p>
1188
1189<pre>
1190 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +00001191</pre>
1192
John McCall87494012011-03-18 03:51:49 +00001193<p>Query for this feature with
1194<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +00001195
John McCall87494012011-03-18 03:51:49 +00001196<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1197
1198<p>Many methods in Objective-C have conventional meanings determined
1199by their selectors. For the purposes of static analysis, it is
1200sometimes useful to be able to mark a method as having a particular
1201conventional meaning despite not having the right selector, or as not
1202having the conventional meaning that its selector would suggest.
1203For these use cases, we provide an attribute to specifically describe
1204the <q>method family</q> that a method belongs to.</p>
1205
1206<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1207where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1208<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1209can only be placed at the end of a method declaration:</p>
1210
1211<pre>
1212 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1213</pre>
1214
1215<p>Users who do not wish to change the conventional meaning of a
1216method, and who merely want to document its non-standard retain and
1217release semantics, should use the
1218<a href="#attr_retain_release">retaining behavior attributes</a>
1219described below.</p>
1220
1221<p>Query for this feature with
1222<tt>__has_attribute(objc_method_family)</tt>.</p>
1223
1224<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +00001225
1226<p>In Objective-C, functions and methods are generally assumed to take
1227and return objects with +0 retain counts, with some exceptions for
1228special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1229there are exceptions, and so Clang provides attributes to allow these
1230exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +00001231ignore non-leaks). Some exceptions may be better described using
1232the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1233attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +00001234
1235<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1236<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1237and <tt>cf_returns_not_retained</tt> attributes can be placed on
1238methods and functions that return Objective-C or CoreFoundation
1239objects. They are commonly placed at the end of a function prototype
1240or method declaration:</p>
1241
1242<pre>
1243 id foo() <b>__attribute__((ns_returns_retained))</b>;
1244
1245 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1246</pre>
1247
1248<p>The <tt>*_returns_retained</tt> attributes specify that the
1249returned object has a +1 retain count.
1250The <tt>*_returns_not_retained</tt> attributes specify that the return
1251object has a +0 retain count, even if the normal convention for its
1252selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1253returned object is +0, but is guaranteed to live at least as long as the
1254next flush of an autorelease pool.</p>
1255
1256<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1257attributes can be placed on an parameter declaration; they specify
1258that the argument is expected to have a +1 retain count, which will be
1259balanced in some way by the function or method.
1260The <tt>ns_consumes_self</tt> attribute can only be placed on an
1261Objective-C method; it specifies that the method expects
1262its <tt>self</tt> parameter to have a +1 retain count, which it will
1263balance in some way.</p>
1264
1265<pre>
1266 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1267
1268 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1269 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1270</pre>
Ted Kremeneked869312009-04-10 05:03:33 +00001271
John McCall87494012011-03-18 03:51:49 +00001272<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1273<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1274
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001275
1276<!-- ======================================================================= -->
Caitlin Sadowski73cbbc82011-07-28 18:38:36 +00001277<h2 id="threadsafety">Thread-Safety Annotation Checking</h2>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001278<!-- ======================================================================= -->
1279
1280<p>Clang supports additional attributes for checking basic locking policies in
1281multithreaded programs.
1282Clang currently parses the following list of attributes, although
1283<b>the implementation for these annotations is currently in development.</b>
1284For more details, see the
1285<a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>.
1286</p>
1287
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001288<h4 id="ts_noanal">no_thread_safety_analysis</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001289
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001290<p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function
1291declaration to specify that the thread safety analysis should not be run on that
1292function. This attribute provides an escape hatch (e.g. for situations when it
1293is difficult to annotate the locking policy). </p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001294
1295<h4 id="ts_lockable">lockable</h4>
1296
1297<p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify
1298that it has a lockable type (e.g. a Mutex class). This annotation is primarily
1299used to check consistency.</p>
1300
1301<h4 id="ts_scopedlockable">scoped_lockable</h4>
1302
1303<p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to
1304specify that it has a "scoped" lockable type. Objects of this type will acquire
1305the lock upon construction and release it upon going out of scope.
1306 This annotation is primarily used to check
1307consistency.</p>
1308
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001309<h4 id="ts_guardedvar">guarded_var</h4>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001310
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001311<p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to
1312specify that the variable must be accessed while holding some lock.</p>
1313
1314<h4 id="ts_ptguardedvar">pt_guarded_var</h4>
1315
1316<p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to
1317specify that the pointer must be dereferenced while holding some lock.</p>
1318
1319<h4 id="ts_guardedby">guarded_by(l)</h4>
1320
1321<p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001322specify that the variable must be accessed while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001323
1324<h4 id="ts_ptguardedby">pt_guarded_by(l)</h4>
1325
1326<p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001327specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001328
1329<h4 id="ts_acquiredbefore">acquired_before(...)</h4>
1330
1331<p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration
1332of a lockable variable to specify that the lock must be acquired before all
1333attribute arguments. Arguments must be lockable type, and there must be at
1334least one argument.</p>
1335
1336<h4 id="ts_acquiredafter">acquired_after(...)</h4>
1337
1338<p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration
1339of a lockable variable to specify that the lock must be acquired after all
1340attribute arguments. Arguments must be lockable type, and there must be at
1341least one argument.</p>
1342
1343<h4 id="ts_elf">exclusive_lock_function(...)</h4>
1344
1345<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1346declaration to specify that the function acquires all listed locks
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001347exclusively. This attribute takes zero or more arguments: either of lockable
1348type or integers indexing into function parameters of lockable type. If no
1349arguments are given, the acquired lock is implicitly <tt>this</tt> of the
1350enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001351
1352<h4 id="ts_slf">shared_lock_function(...)</h4>
1353
1354<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1355declaration to specify that the function acquires all listed locks, although
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001356 the locks may be shared (e.g. read locks). This attribute takes zero or more
1357arguments: either of lockable type or integers indexing into function
1358parameters of lockable type. If no arguments are given, the acquired lock is
1359implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001360
1361<h4 id="ts_etf">exclusive_trylock_function(...)</h4>
1362
1363<p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function
1364declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001365all listed locks exclusively. This attribute takes one or more arguments. The
1366first argument is an integer or boolean value specifying the return value of a
1367successful lock acquisition. The remaining arugments are either of lockable type
1368or integers indexing into function parameters of lockable type. If only one
1369argument is given, the acquired lock is implicitly <tt>this</tt> of the
1370enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001371
1372<h4 id="ts_stf">shared_trylock_function(...)</h4>
1373
1374<p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function
1375declaration to specify that the function will try (without blocking) to acquire
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001376all listed locks, although the locks may be shared (e.g. read locks). This
1377attribute takes one or more arguments. The first argument is an integer or
1378boolean value specifying the return value of a successful lock acquisition. The
1379remaining arugments are either of lockable type or integers indexing into
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001380function parameters of lockable type. If only one argument is given, the
1381acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
1382
1383<h4 id="ts_uf">unlock_function(...)</h4>
1384
1385<p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001386declaration to specify that the function release all listed locks. This
1387attribute takes zero or more arguments: either of lockable type or integers
1388indexing into function parameters of lockable type. If no arguments are given,
1389the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001390
1391<h4 id="ts_lr">lock_returned(l)</h4>
1392
1393<p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001394declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt>
1395must be of lockable type). This annotation is used to aid in resolving lock
1396expressions.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001397
1398<h4 id="ts_le">locks_excluded(...)</h4>
1399
1400<p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration
Caitlin Sadowskib51e0312011-08-09 17:59:31 +00001401to specify that the function must not be called with the listed locks. Arguments
1402must be lockable type, and there must be at least one argument.</p>
Caitlin Sadowskidb33e142011-07-28 20:12:35 +00001403
1404<h4 id="ts_elr">exclusive_locks_required(...)</h4>
1405
1406<p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function
1407declaration to specify that the function must be called while holding the listed
1408exclusive locks. Arguments must be lockable type, and there must be at
1409least one argument.</p>
1410
1411<h4 id="ts_slr">shared_locks_required(...)</h4>
1412
1413<p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function
1414declaration to specify that the function must be called while holding the listed
1415shared locks. Arguments must be lockable type, and there must be at
1416least one argument.</p>
Caitlin Sadowskifdde9e72011-07-28 17:21:07 +00001417
Chris Lattner5ce933f2009-02-09 08:46:11 +00001418</div>
1419</body>
1420</html>