Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / clang / 572bbec2e0e23f684108e3a410a67c8c60ff9179 / . / docs / LanguageExtensions.html
blob: 551e17f22058994c7b071dd68303e4eccd0cf87d [file] [log] [blame]
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1<html>
2<head>
3<title>Clang Language Extensions</title>
4<link type="text/css" rel="stylesheet" href="../menu.css" />
5<link type="text/css" rel="stylesheet" href="../content.css" />
6<style type="text/css">
7td {
8 vertical-align: top;
9}
10</style>
11</head>
12<body>
13
14<!--#include virtual="../menu.html.incl"-->
15
16<div id="content">
17
18<h1>Clang Language Extensions</h1>
19
20<ul>
21<li><a href="#intro">Introduction</a></li>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]22<li><a href="#feature_check">Feature Checking Macros</a></li>
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]23<li><a href="#has_include">Include File Checking Macros</a></li>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]24<li><a href="#builtinmacros">Builtin Macros</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]25<li><a href="#vectors">Vectors and Extended Vectors</a></li>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]26<li><a href="#deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> attributes</a></li>
27<li><a href="#attributes-on-enumerators">Attributes on enumerators</a></li>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]28<li><a href="#checking_language_features">Checks for Standard Language Features</a></li>
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]29 <ul>
30 <li><a href="#cxx_exceptions">C++ exceptions</a></li>
31 <li><a href="#cxx_rtti">C++ RTTI</a></li>
32 </ul>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]33<li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a></li>
34 <ul>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]35 <li><a href="#cxx0x">C++0x</a>
36 <ul>
37 <li><a href="#cxx_decltype">C++0x <tt>decltype()</tt></a></li>
38 <li><a href="#cxx_access_control_sfinae">C++0x SFINAE includes access control</a></li>
39 <li><a href="#cxx_alias_templates">C++0x alias templates</a></li>
40 <li><a href="#cxx_attributes">C++0x attributes</a></li>
41 <li><a href="#cxx_default_function_template_args">C++0x default template arguments in function templates</a></li>
42 <li><a href="#cxx_deleted_functions">C++0x deleted functions</a></li>
43 <li><a href="#cxx_lambdas">C++0x lambdas</a></li>
44 <li><a href="#cxx_nullptr">C++0x nullptr</a></li>
45 <li><a href="#cxx_override_control">C++0x override control</a></li>
46 <li><a href="#cxx_range_for">C++0x range-based for loop</a></li>
47 <li><a href="#cxx_rvalue_references">C++0x rvalue references</a></li>
48 <li><a href="#cxx_reference_qualified_functions">C++0x reference-qualified functions</a></li>
49 <li><a href="#cxx_static_assert">C++0x <tt>static_assert()</tt></a></li>
50 <li><a href="#cxx_auto_type">C++0x type inference</a></li>
51 <li><a href="#cxx_variadic_templates">C++0x variadic templates</a></li>
52 <li><a href="#cxx_inline_namespaces">C++0x inline namespaces</a></li>
53 <li><a href="#cxx_strong_enums">C++0x strongly-typed enumerations</a></li>
54 <li><a href="#cxx_trailing_return">C++0x trailing return type</a></li>
55 <li><a href="#cxx_noexcept">C++0x noexcept specification</a></li>
56 </ul>
57 <li><a href="#c1x">C1X</a>
58 <ul>
59 <li><a href="#c_generic_selections">C1X generic selections</a></li>
60 <li><a href="#c_static_assert">C1X <tt>_Static_assert()</tt></a></li>
61 </ul>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]62 </ul>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]63<li><a href="#checking_type_traits">Checks for Type Traits</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]64<li><a href="#blocks">Blocks</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]65<li><a href="#objc_features">Objective-C Features</a>
66 <ul>
67 <li><a href="#objc_instancetype">Related result types</a></li>
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]68 <li><a href="#objc_arc">Automatic reference counting</a></li>
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]69 </ul>
70</li>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]71<li><a href="#overloading-in-c">Function Overloading in C</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]72<li><a href="#builtins">Builtin Functions</a>
73 <ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]74 <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li>
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]75 <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li>
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]76 <li><a href="#__sync_swap">__sync_swap</a></li>
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]77 </ul>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]78</li>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]79<li><a href="#targetspecific">Target-Specific Extensions</a>
80 <ul>
81 <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li>
82 </ul>
83</li>
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]84<li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]85</ul>
86
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]87<!-- ======================================================================= -->
88<h2 id="intro">Introduction</h2>
89<!-- ======================================================================= -->
90
91<p>This document describes the language extensions provided by Clang. In
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]92addition to the language extensions listed here, Clang aims to support a broad
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]93range of GCC extensions. Please see the <a
94href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for
95more information on these extensions.</p>
96
97<!-- ======================================================================= -->
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]98<h2 id="feature_check">Feature Checking Macros</h2>
99<!-- ======================================================================= -->
100
101<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]102on them. In order to allow fine-grain features checks, we support three builtin
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]103function-like macros. This allows you to directly test for a feature in your
104code without having to resort to something like autoconf or fragile "compiler
105version checks".</p>
106
107<!-- ======================================================================= -->
108<h3 id="__has_builtin">__has_builtin</h3>
109<!-- ======================================================================= -->
110
111<p>This function-like macro takes a single identifier argument that is the name
112of a builtin function. It evaluates to 1 if the builtin is supported or 0 if
113not. It can be used like this:</p>
114
115<blockquote>
116<pre>
117#ifndef __has_builtin // Optional of course.
118 #define __has_builtin(x) 0 // Compatibility with non-clang compilers.
119#endif
120
121...
122#if __has_builtin(__builtin_trap)
123 __builtin_trap();
124#else
125 abort();
126#endif
127...
128</pre>
129</blockquote>
130
131
132<!-- ======================================================================= -->
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]133<h3 id="__has_feature_extension">__has_feature and __has_extension</h3>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]134<!-- ======================================================================= -->
135
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]136<p>These function-like macros take a single identifier argument that is the
137name of a feature. <code>__has_feature</code> evaluates to 1 if the feature
138is both supported by Clang and standardized in the current language standard
139or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while
140<code>__has_extension</code> evaluates to 1 if the feature is supported by
141Clang in the current language (either as a language extension or a standard
142language feature) or 0 if not. They can be used like this:</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]143
144<blockquote>
145<pre>
146#ifndef __has_feature // Optional of course.
147 #define __has_feature(x) 0 // Compatibility with non-clang compilers.
148#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]149#ifndef __has_extension
150 #define __has_extension __has_feature // Compatibility with pre-3.0 compilers.
151#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]152
153...
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]154#if __has_feature(cxx_rvalue_references)
155// This code will only be compiled with the -std=c++0x and -std=gnu++0x
156// options, because rvalue references are only standardized in C++0x.
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]157#endif
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]158
159#if __has_extension(cxx_rvalue_references)
160// This code will be compiled with the -std=c++0x, -std=gnu++0x, -std=c++98
161// and -std=gnu++98 options, because rvalue references are supported as a
162// language extension in C++98.
163#endif
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]164</pre>
165</blockquote>
166
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]167<p id="has_feature_back_compat">For backwards compatibility reasons,
168<code>__has_feature</code> can also be used to test for support for
169non-standardized features, i.e. features not prefixed <code>c_</code>,
170<code>cxx_</code> or <code>objc_</code>.</p>
171
172<p>If the <code>-pedantic-errors</code> option is given,
173<code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p>
174
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]175<p>The feature tag is described along with the language feature below.</p>
176
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]177<!-- ======================================================================= -->
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]178<h3 id="__has_attribute">__has_attribute</h3>
179<!-- ======================================================================= -->
180
181<p>This function-like macro takes a single identifier argument that is the name
182of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It
183can be used like this:</p>
184
185<blockquote>
186<pre>
187#ifndef __has_attribute // Optional of course.
188 #define __has_attribute(x) 0 // Compatibility with non-clang compilers.
189#endif
190
191...
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]192#if __has_attribute(always_inline)
193#define ALWAYS_INLINE __attribute__((always_inline))
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]194#else
Anders Carlsson961003d2011-01-24 03:54:51 +0000[diff] [blame]195#define ALWAYS_INLINE
Anders Carlssoncae50952010-10-20 02:31:43 +0000[diff] [blame]196#endif
197...
198</pre>
199</blockquote>
200
201<!-- ======================================================================= -->
John Thompson92bd8c72009-11-02 22:28:12 +0000[diff] [blame]202<h2 id="has_include">Include File Checking Macros</h2>
203<!-- ======================================================================= -->
204
205<p>Not all developments systems have the same include files.
206The <a href="#__has_include">__has_include</a> and
207<a href="#__has_include_next">__has_include_next</a> macros allow you to
208check for the existence of an include file before doing
209a possibly failing #include directive.</p>
210
211<!-- ======================================================================= -->
212<h3 id="__has_include">__has_include</h3>
213<!-- ======================================================================= -->
214
215<p>This function-like macro takes a single file name string argument that
216is the name of an include file. It evaluates to 1 if the file can
217be found using the include paths, or 0 otherwise:</p>
218
219<blockquote>
220<pre>
221// Note the two possible file name string formats.
222#if __has_include("myinclude.h") && __has_include(&lt;stdint.h&gt;)
223# include "myinclude.h"
224#endif
225
226// To avoid problem with non-clang compilers not having this macro.
227#if defined(__has_include) && __has_include("myinclude.h")
228# include "myinclude.h"
229#endif
230</pre>
231</blockquote>
232
233<p>To test for this feature, use #if defined(__has_include).</p>
234
235<!-- ======================================================================= -->
236<h3 id="__has_include_next">__has_include_next</h3>
237<!-- ======================================================================= -->
238
239<p>This function-like macro takes a single file name string argument that
240is the name of an include file. It is like __has_include except that it
241looks for the second instance of the given file found in the include
242paths. It evaluates to 1 if the second instance of the file can
243be found using the include paths, or 0 otherwise:</p>
244
245<blockquote>
246<pre>
247// Note the two possible file name string formats.
248#if __has_include_next("myinclude.h") && __has_include_next(&lt;stdint.h&gt;)
249# include_next "myinclude.h"
250#endif
251
252// To avoid problem with non-clang compilers not having this macro.
253#if defined(__has_include_next) && __has_include_next("myinclude.h")
254# include_next "myinclude.h"
255#endif
256</pre>
257</blockquote>
258
259<p>Note that __has_include_next, like the GNU extension
260#include_next directive, is intended for use in headers only,
261and will issue a warning if used in the top-level compilation
262file. A warning will also be issued if an absolute path
263is used in the file argument.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]264
265<!-- ======================================================================= -->
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]266<h2 id="builtinmacros">Builtin Macros</h2>
267<!-- ======================================================================= -->
268
Douglas Gregor4290fbd2010-04-30 02:51:06 +0000[diff] [blame]269<dl>
270 <dt><code>__BASE_FILE__</code></dt>
271 <dd>Defined to a string that contains the name of the main input
272 file passed to Clang.</dd>
273
274 <dt><code>__COUNTER__</code></dt>
275 <dd>Defined to an integer value that starts at zero and is
276 incremented each time the <code>__COUNTER__</code> macro is
277 expanded.</dd>
278
279 <dt><code>__INCLUDE_LEVEL__</code></dt>
280 <dd>Defined to an integral value that is the include depth of the
281 file currently being translated. For the main file, this value is
282 zero.</dd>
283
284 <dt><code>__TIMESTAMP__</code></dt>
285 <dd>Defined to the date and time of the last modification of the
286 current source file.</dd>
287
288 <dt><code>__clang__</code></dt>
289 <dd>Defined when compiling with Clang</dd>
290
291 <dt><code>__clang_major__</code></dt>
292 <dd>Defined to the major version number of Clang (e.g., the 2 in
293 2.0.1).</dd>
294
295 <dt><code>__clang_minor__</code></dt>
296 <dd>Defined to the minor version number of Clang (e.g., the 0 in
297 2.0.1).</dd>
298
299 <dt><code>__clang_patchlevel__</code></dt>
300 <dd>Defined to the patch level of Clang (e.g., the 1 in 2.0.1).</dd>
301
302 <dt><code>__clang_version__</code></dt>
303 <dd>Defined to a string that captures the Clang version, including
304 the Subversion tag or revision number, e.g., "1.5 (trunk
305 102332)".</dd>
306</dl>
Chris Lattner81edc9f2009-04-13 02:45:46 +0000[diff] [blame]307
308<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]309<h2 id="vectors">Vectors and Extended Vectors</h2>
310<!-- ======================================================================= -->
311
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]312<p>Supports the GCC vector extensions, plus some stuff like V[1].</p>
313
314<p>Also supports <tt>ext_vector</tt>, which additionally support for V.xyzw
315syntax and other tidbits as seen in OpenCL. An example is:</p>
316
317<blockquote>
318<pre>
319typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>;
320typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>;
321
322float4 foo(float2 a, float2 b) {
323 float4 c;
324 c.xz = a;
325 c.yw = b;
326 return c;
327}
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]328</pre>
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]329</blockquote>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]330
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]331<p>Query for this feature with __has_extension(attribute_ext_vector_type).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]332
Owen Andersond2bf0cd2010-01-27 01:22:36 +0000[diff] [blame]333<p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p>
334
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]335<!-- ======================================================================= -->
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]336<h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]337<!-- ======================================================================= -->
338
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]339<p>An optional string message can be added to the <tt>deprecated</tt>
340and <tt>unavailable</tt> attributes. For example:</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]341
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]342<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]343<pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]344</blockquote>
345
346<p>If the deprecated or unavailable declaration is used, the message
347will be incorporated into the appropriate diagnostic:</p>
348
349<blockquote>
Chris Lattner4836d6a2010-11-09 19:43:35 +0000[diff] [blame]350<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]351 explode();
352 ^</pre>
353</blockquote>
354
355<p>Query for this feature
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]356with <tt>__has_extension(attribute_deprecated_with_message)</tt>
357and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p>
John McCall48209082010-11-08 19:48:17 +0000[diff] [blame]358
359<!-- ======================================================================= -->
360<h2 id="attributes-on-enumerators">Attributes on Enumerators</h2>
361<!-- ======================================================================= -->
362
363<p>Clang allows attributes to be written on individual enumerators.
364This allows enumerators to be deprecated, made unavailable, etc. The
365attribute must appear after the enumerator name and before any
366initializer, like so:</p>
367
368<blockquote>
369<pre>enum OperationMode {
370 OM_Invalid,
371 OM_Normal,
372 OM_Terrified __attribute__((deprecated)),
373 OM_AbortOnError __attribute__((deprecated)) = 4
374};</pre>
375</blockquote>
376
377<p>Attributes on the <tt>enum</tt> declaration do not apply to
378individual enumerators.</p>
379
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]380<p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p>
Fariborz Jahanianc784dc12010-10-06 23:12:32 +0000[diff] [blame]381
382<!-- ======================================================================= -->
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]383<h2 id="checking_language_features">Checks for Standard Language Features</h2>
384<!-- ======================================================================= -->
385
386<p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are
387enabled. Those features are listed here.</p>
388
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]389<h3 id="cxx_exceptions">C++ exceptions</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]390
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]391<p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For
Sean Hunt4b4ba8b2011-06-22 23:49:12 +0000[diff] [blame]392example, compiling code with <tt>-fexceptions</tt> and <tt>-fc++-exceptions</tt> enables C++ exceptions.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]393
Ted Kremenek22c34102009-12-03 02:05:57 +0000[diff] [blame]394<h3 id="cxx_rtti">C++ RTTI</h3>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]395
Ted Kremenek0eb95602009-12-03 02:06:43 +0000[diff] [blame]396<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]397compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p>
Ted Kremenek87774fd2009-12-03 02:04:01 +0000[diff] [blame]398
399<!-- ======================================================================= -->
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]400<h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2>
401<!-- ======================================================================= -->
402
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]403<p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used
404to query if certain upcoming standard language features are enabled. Those
405features are listed here. Features that are not yet implemented will be
406noted.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]407
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]408<h3 id="cxx0x">C++0x</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]409
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]410<p>The features listed below are slated for inclusion in the upcoming
411C++0x standard. As a result, all these features are enabled
412with the <tt>-std=c++0x</tt> option when compiling C++ code.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]413
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]414<h4 id="cxx_decltype">C++0x <tt>decltype()</tt></h3>
415
416<p>Use <tt>__has_feature(cxx_decltype)</tt> or
417<tt>__has_extension(cxx_decltype)</tt> to determine if support for the
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]418<tt>decltype()</tt> specifier is enabled.</p>
419
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]420<h4 id="cxx_access_control_sfinae">C++0x SFINAE includes access control</h3>
Douglas Gregor7822ee32011-05-11 23:45:11 +0000[diff] [blame]421
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]422<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]423
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]424<h4 id="cxx_alias_templates">C++0x alias templates</h3>
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]425
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]426<p>Use <tt>__has_feature(cxx_alias_templates)</tt> or
427<tt>__has_extension(cxx_alias_templates)</tt> to determine if support for
Richard Smith3e4c6c42011-05-05 21:57:07 +0000[diff] [blame]428C++0x's alias declarations and alias templates is enabled.</p>
429
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]430<h4 id="cxx_attributes">C++0x attributes</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]431
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]432<p>Use <tt>__has_feature(cxx_attributes)</tt> or
433<tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute
434parsing with C++0x's square bracket notation is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]435
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]436<h4 id="cxx_default_function_template_args">C++0x default template arguments in function templates</h3>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]437
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]438<p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or
439<tt>__has_extension(cxx_default_function_template_args)</tt> to determine
440if support for default template arguments in function templates is enabled.</p>
Douglas Gregor07508002011-02-05 20:35:30 +0000[diff] [blame]441
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]442<h4 id="cxx_deleted_functions">C++0x deleted functions</tt></h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]443
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]444<p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or
445<tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]446deleted function definitions (with <tt>= delete</tt>) is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]447
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]448<h4 id="cxx_lambdas">C++0x lambdas</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]449
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]450<p>Use <tt>__has_feature(cxx_lambdas)</tt> or
451<tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas
452is enabled. clang does not currently implement this feature.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]453
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]454<h4 id="cxx_nullptr">C++0x <tt>nullptr</tt></h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]455
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]456<p>Use <tt>__has_feature(cxx_nullptr)</tt> or
457<tt>__has_extension(cxx_nullptr)</tt> to determine if support for
Douglas Gregor84ee2ee2011-05-21 23:15:46 +0000[diff] [blame]458<tt>nullptr</tt> is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]459
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]460<h4 id="cxx_override_control">C++0x <tt>override control</tt></h3>
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]461
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]462<p>Use <tt>__has_feature(cxx_override_control)</tt> or
463<tt>__has_extension(cxx_override_control)</tt> to determine if support for
Anders Carlssonc8b9f792011-03-25 15:04:23 +0000[diff] [blame]464the override control keywords is enabled.</p>
465
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]466<h4 id="cxx_reference_qualified_functions">C++0x reference-qualified functions</h3>
467<p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or
468<tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine
469if support for reference-qualified functions (e.g., member functions with
470<code>&amp;</code> or <code>&amp;&amp;</code> applied to <code>*this</code>)
471is enabled.</p>
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]472
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]473<h4 id="cxx_range_for">C++0x range-based for loop</tt></h3>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]474
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]475<p>Use <tt>__has_feature(cxx_range_for)</tt> or
476<tt>__has_extension(cxx_range_for)</tt> to determine if support for the
477range-based for loop is enabled. </p>
Richard Smitha391a462011-04-15 15:14:40 +0000[diff] [blame]478
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]479<h4 id="cxx_rvalue_references">C++0x rvalue references</tt></h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]480
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]481<p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or
482<tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for
Douglas Gregor56209ff2011-01-26 21:25:54 +0000[diff] [blame]483rvalue references is enabled. </p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]484
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]485<h4 id="cxx_static_assert">C++0x <tt>static_assert()</tt></h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]486
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]487<p>Use <tt>__has_feature(cxx_static_assert)</tt> or
488<tt>__has_extension(cxx_static_assert)</tt> to determine if support for
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]489compile-time assertions using <tt>static_assert</tt> is enabled.</p>
490
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]491<h4 id="cxx_auto_type">C++0x type inference</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]492
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]493<p>Use <tt>__has_feature(cxx_auto_type)</tt> or
494<tt>__has_extension(cxx_auto_type)</tt> to determine C++0x type inference is
495supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt>
496will instead be a storage class specifier, as in C or C++98.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]497
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]498<h4 id="cxx_variadic_templates">C++0x variadic templates</h3>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]499
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]500<p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or
501<tt>__has_extension(cxx_variadic_templates)</tt> to determine if support
Douglas Gregor83d77812011-01-19 23:15:20 +0000[diff] [blame]502for variadic templates is enabled.</p>
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]503
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]504<h4 id="cxx_inline_namespaces">C++0x inline namespaces</h3>
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]505
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]506<p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or
507<tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for
Sebastian Redlf6c09772010-08-31 23:28:47 +0000[diff] [blame]508inline namespaces is enabled.</p>
509
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]510<h4 id="cxx_trailing_return">C++0x trailing return type</h3>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]511
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]512<p>Use <tt>__has_feature(cxx_trailing_return)</tt> or
513<tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the
514alternate function declaration syntax with trailing return type is enabled.</p>
Douglas Gregordab60ad2010-10-01 18:44:50 +0000[diff] [blame]515
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]516<h4 id="cxx_noexcept">C++0x noexcept</h3>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]517
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]518<p>Use <tt>__has_feature(cxx_noexcept)</tt> or
519<tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept
520exception specifications is enabled.</p>
Sebastian Redl4561ecd2011-03-15 21:17:12 +0000[diff] [blame]521
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]522<h4 id="cxx_strong_enums">C++0x strongly typed enumerations</h3>
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]523
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]524<p>Use <tt>__has_feature(cxx_strong_enums)</tt> or
525<tt>__has_extension(cxx_strong_enums)</tt> to determine if support for
Douglas Gregor1274ccd2010-10-08 23:50:27 +0000[diff] [blame]526strongly typed, scoped enumerations is enabled.</p>
527
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]528<h3 id="c1x">C1X</h3>
529
530<p>The features listed below are slated for inclusion in the upcoming
531C1X standard. As a result, all these features are enabled
532with the <tt>-std=c1x</tt> option when compiling C code.</p>
533
534<h4 id="c_generic_selections">C1X generic selections</h2>
535
536<p>Use <tt>__has_feature(c_generic_selections)</tt> or
537<tt>__has_extension(c_generic_selections)</tt> to determine if support for
538generic selections is enabled.</p>
539
540<p>As an extension, the C1X generic selection expression is available in all
541languages supported by Clang. The syntax is the same as that given in the
542C1X draft standard.</p>
543
544<p>In C, type compatibility is decided according to the rules given in the
545appropriate standard, but in C++, which lacks the type compatibility rules
546used in C, types are considered compatible only if they are equivalent.</p>
547
548<h4 id="c_static_assert">C1X <tt>_Static_assert()</tt></h3>
549
550<p>Use <tt>__has_feature(c_static_assert)</tt> or
551<tt>__has_extension(c_static_assert)</tt> to determine if support for
552compile-time assertions using <tt>_Static_assert</tt> is enabled.</p>
553
Sean Hunt4ef4c6b2010-01-13 08:31:49 +0000[diff] [blame]554<!-- ======================================================================= -->
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]555<h2 id="checking_type_traits">Checks for Type Traits</h2>
556<!-- ======================================================================= -->
557
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]558<p>Clang supports the <a hef="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]559<blockquote>
560<pre>
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]561#if __has_extension(is_convertible_to)
Douglas Gregorafdf1372011-02-03 21:57:35 +0000[diff] [blame]562template&lt;typename From, typename To&gt;
563struct is_convertible_to {
564 static const bool value = __is_convertible_to(From, To);
565};
566#else
567// Emulate type trait
568#endif
569</pre>
570</blockquote>
571
572<p>The following type traits are supported by Clang:</p>
573<ul>
574 <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li>
575 <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li>
576 <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li>
577 <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li>
578 <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li>
579 <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li>
580 <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li>
581 <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li>
582 <li><code>__is_abstract</code> (GNU, Microsoft)</li>
583 <li><code>__is_base_of</code> (GNU, Microsoft)</li>
584 <li><code>__is_class</code> (GNU, Microsoft)</li>
585 <li><code>__is_convertible_to</code> (Microsoft)</li>
586 <li><code>__is_empty</code> (GNU, Microsoft)</li>
587 <li><code>__is_enum</code> (GNU, Microsoft)</li>
588 <li><code>__is_pod</code> (GNU, Microsoft)</li>
589 <li><code>__is_polymorphic</code> (GNU, Microsoft)</li>
590 <li><code>__is_union</code> (GNU, Microsoft)</li>
591 <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li>
592</ul>
593
594<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]595<h2 id="blocks">Blocks</h2>
596<!-- ======================================================================= -->
597
Chris Lattnera7dbdf52009-03-09 07:03:22 +0000[diff] [blame]598<p>The syntax and high level language feature description is in <a
599href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI
600details for the clang implementation are in <a
Chris Lattner5d7650b2010-03-16 21:43:03 +0000[diff] [blame]601href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]602
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]603
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]604<p>Query for this feature with __has_extension(blocks).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]605
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]606<!-- ======================================================================= -->
Douglas Gregor926df6c2011-06-11 01:09:30 +0000[diff] [blame]607<h2 id="objc_features">Objective-C Features</h2>
608<!-- ======================================================================= -->
609
610<h3 id="objc_instancetype">Related result types</h3>
611
612<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>
613
614<blockquote>
615<pre>
616@interface NSObject
617+ (id)alloc;
618- (id)init;
619@end
620
621@interface NSArray : NSObject
622@end
623</pre>
624</blockquote>
625
626<p>and this common initialization pattern</p>
627
628<blockquote>
629<pre>
630NSArray *array = [[NSArray alloc] init];
631</pre>
632</blockquote>
633
634<p>the type of the expression <code>[NSArray alloc]</code> is
635<code>NSArray*</code> because <code>alloc</code> implicitly has a
636related result type. Similarly, the type of the expression
637<code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since
638<code>init</code> has a related result type and its receiver is known
639to 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>
640
641<p>To determine whether a method has a related result type, the first
642word in the camel-case selector (e.g., "init" in "initWithObjects") is
643considered, and the method will a related result type if its return
644type is compatible with the type of its class and if
645
646<ul>
647
648 <li>the first word is "alloc" or "new", and the method is a class
649 method, or</li>
650
651 <li>the first word is "autorelease", "init", "retain", or "self",
652 and the method is an instance method.</li>
653
654</ul></p>
655
656<p>If a method with a related result type is overridden by a subclass
657method, the subclass method must also return a type that is compatible
658with the subclass type. For example:</p>
659
660<blockquote>
661<pre>
662@interface NSString : NSObject
663- (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString
664@end
665</pre>
666</blockquote>
667
668<p>Related result types only affect the type of a message send or
669property access via the given method. In all other respects, a method
670with a related result type is treated the same way as method without a
671related result type.</p>
672
673<!-- ======================================================================= -->
John McCallf85e1932011-06-15 23:02:42 +0000[diff] [blame]674<h2 id="objc_arc">Automatic reference counting </h2>
675<!-- ======================================================================= -->
676
677<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>
678
679<!-- ======================================================================= -->
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]680<h2 id="overloading-in-c">Function Overloading in C</h2>
681<!-- ======================================================================= -->
682
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]683<p>Clang provides support for C++ function overloading in C. Function
684overloading in C is introduced using the <tt>overloadable</tt> attribute. For
685example, one might provide several overloaded versions of a <tt>tgsin</tt>
686function that invokes the appropriate standard function computing the sine of a
687value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt>
688precision:</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]689
690<blockquote>
691<pre>
692#include &lt;math.h&gt;
693float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); }
694double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); }
695long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); }
696</pre>
697</blockquote>
698
699<p>Given these declarations, one can call <tt>tgsin</tt> with a
700<tt>float</tt> value to receive a <tt>float</tt> result, with a
701<tt>double</tt> to receive a <tt>double</tt> result, etc. Function
702overloading in C follows the rules of C++ function overloading to pick
703the best overload given the call arguments, with a few C-specific
704semantics:</p>
705<ul>
706 <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long
707 double</tt> is ranked as a floating-point promotion (per C99) rather
708 than as a floating-point conversion (as in C++).</li>
709
710 <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type
711 <tt>U*</tt> is considered a pointer conversion (with conversion
712 rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li>
713
714 <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt>
715 is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This
716 conversion is given "conversion" rank.</li>
717</ul>
718
719<p>The declaration of <tt>overloadable</tt> functions is restricted to
720function declarations and definitions. Most importantly, if any
721function with a given name is given the <tt>overloadable</tt>
722attribute, then all function declarations and definitions with that
723name (and in that scope) must have the <tt>overloadable</tt>
Chris Lattnerf161d412009-02-13 21:51:45 +0000[diff] [blame]724attribute. This rule even applies to redeclarations of functions whose original
725declaration had the <tt>overloadable</tt> attribute, e.g.,</p>
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]726
727<blockquote>
728<pre>
729int f(int) __attribute__((overloadable));
730float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i>
731
732int g(int) __attribute__((overloadable));
733int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i>
734</pre>
735</blockquote>
736
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]737<p>Functions marked <tt>overloadable</tt> must have
738prototypes. Therefore, the following code is ill-formed:</p>
739
740<blockquote>
741<pre>
742int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i>
743</pre>
744</blockquote>
745
746<p>However, <tt>overloadable</tt> functions are allowed to use a
747ellipsis 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>
748
749<blockquote>
750<pre>
Chris Lattner02246802009-02-18 22:27:46 +0000[diff] [blame]751void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i>
Douglas Gregor965acbb2009-02-18 07:07:28 +0000[diff] [blame]752</pre>
753</blockquote>
754
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]755<p>Functions declared with the <tt>overloadable</tt> attribute have
756their names mangled according to the same rules as C++ function
757names. For example, the three <tt>tgsin</tt> functions in our
758motivating example get the mangled names <tt>_Z5tgsinf</tt>,
Chris Lattner71b48d62010-11-28 18:19:13 +0000[diff] [blame]759<tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]760caveats to this use of name mangling:</p>
761
762<ul>
763
764 <li>Future versions of Clang may change the name mangling of
765 functions overloaded in C, so you should not depend on an specific
766 mangling. To be completely safe, we strongly urge the use of
767 <tt>static inline</tt> with <tt>overloadable</tt> functions.</li>
768
769 <li>The <tt>overloadable</tt> attribute has almost no meaning when
770 used in C++, because names will already be mangled and functions are
771 already overloadable. However, when an <tt>overloadable</tt>
772 function occurs within an <tt>extern "C"</tt> linkage specification,
773 it's name <i>will</i> be mangled in the same way as it would in
774 C.</li>
775</ul>
776
Peter Collingbournec1b5fa42011-05-13 20:54:45 +0000[diff] [blame]777<p>Query for this feature with __has_extension(attribute_overloadable).</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]778
779
Douglas Gregorcb54d432009-02-13 00:57:04 +0000[diff] [blame]780<!-- ======================================================================= -->
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]781<h2 id="builtins">Builtin Functions</h2>
782<!-- ======================================================================= -->
783
784<p>Clang supports a number of builtin library functions with the same syntax as
785GCC, including things like <tt>__builtin_nan</tt>,
786<tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>,
787<tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In
788addition to the GCC builtins, Clang supports a number of builtins that GCC does
789not, which are listed here.</p>
790
791<p>Please note that Clang does not and will not support all of the GCC builtins
792for vector operations. Instead of using builtins, you should use the functions
793defined in target-specific header files like <tt>&lt;xmmintrin.h&gt;</tt>, which
794define portable wrappers for these. Many of the Clang versions of these
795functions are implemented directly in terms of <a href="#vectors">extended
796vector support</a> instead of builtins, in order to reduce the number of
797builtins that we need to implement.</p>
798
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]799<!-- ======================================================================= -->
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]800<h3 id="__builtin_shufflevector">__builtin_shufflevector</h3>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]801<!-- ======================================================================= -->
802
Chris Lattneraad826b2009-09-16 18:56:12 +0000[diff] [blame]803<p><tt>__builtin_shufflevector</tt> is used to express generic vector
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]804permutation/shuffle/swizzle operations. This builtin is also very important for
805the implementation of various target-specific header files like
806<tt>&lt;xmmintrin.h&gt;</tt>.
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]807</p>
808
809<p><b>Syntax:</b></p>
810
811<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]812__builtin_shufflevector(vec1, vec2, index1, index2, ...)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]813</pre>
814
815<p><b>Examples:</b></p>
816
817<pre>
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]818 // Identity operation - return 4-element vector V1.
819 __builtin_shufflevector(V1, V1, 0, 1, 2, 3)
820
821 // "Splat" element 0 of V1 into a 4-element result.
822 __builtin_shufflevector(V1, V1, 0, 0, 0, 0)
823
824 // Reverse 4-element vector V1.
825 __builtin_shufflevector(V1, V1, 3, 2, 1, 0)
826
827 // Concatenate every other element of 4-element vectors V1 and V2.
828 __builtin_shufflevector(V1, V2, 0, 2, 4, 6)
829
830 // Concatenate every other element of 8-element vectors V1 and V2.
831 __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14)
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]832</pre>
833
834<p><b>Description:</b></p>
835
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]836<p>The first two arguments to __builtin_shufflevector are vectors that have the
837same element type. The remaining arguments are a list of integers that specify
838the elements indices of the first two vectors that should be extracted and
839returned in a new vector. These element indices are numbered sequentially
840starting with the first vector, continuing into the second vector. Thus, if
841vec1 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]842</p>
843
Chris Lattner6f72da52009-02-13 20:00:20 +0000[diff] [blame]844<p>The result of __builtin_shufflevector is a vector
845with the same element type as vec1/vec2 but that has an element count equal to
846the number of indices specified.
847</p>
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]848
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]849<p>Query for this feature with __has_builtin(__builtin_shufflevector).</p>
850
851<!-- ======================================================================= -->
852<h3 id="__builtin_unreachable">__builtin_unreachable</h3>
853<!-- ======================================================================= -->
854
855<p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in
856the program cannot be reached, even if the compiler might otherwise think it
857can. This is useful to improve optimization and eliminates certain warnings.
858For example, without the <tt>__builtin_unreachable</tt> in the example below,
859the compiler assumes that the inline asm can fall through and prints a "function
860declared 'noreturn' should not return" warning.
861</p>
862
863<p><b>Syntax:</b></p>
864
865<pre>
866__builtin_unreachable()
867</pre>
868
869<p><b>Example of Use:</b></p>
870
871<pre>
872void myabort(void) __attribute__((noreturn));
873void myabort(void) {
874 asm("int3");
875 __builtin_unreachable();
876}
877</pre>
878
879<p><b>Description:</b></p>
880
881<p>The __builtin_unreachable() builtin has completely undefined behavior. Since
882it has undefined behavior, it is a statement that it is never reached and the
883optimizer can take advantage of this to produce better code. This builtin takes
884no arguments and produces a void result.
885</p>
886
887<p>Query for this feature with __has_builtin(__builtin_unreachable).</p>
888
Chris Lattner23aa9c82011-04-09 03:57:26 +0000[diff] [blame]889<!-- ======================================================================= -->
890<h3 id="__sync_swap">__sync_swap</h3>
891<!-- ======================================================================= -->
892
893<p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in
894memory.
895</p>
896
897<p><b>Syntax:</b></p>
898
899<pre>
900<i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...)
901</pre>
902
903<p><b>Example of Use:</b></p>
904
905<pre>
906int old_value = __sync_swap(&value, new_value);
907</pre>
908
909<p><b>Description:</b></p>
910
911<p>The __sync_swap() builtin extends the existing __sync_*() family of atomic
912intrinsics to allow code to atomically swap the current value with the new
913value. More importantly, it helps developers write more efficient and correct
914code by avoiding expensive loops around __sync_bool_compare_and_swap() or
915relying on the platform specific implementation details of
916__sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier.
917</p>
918
Chris Lattner21190d52009-09-21 03:09:59 +0000[diff] [blame]919
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]920<!-- ======================================================================= -->
921<h2 id="targetspecific">Target-Specific Extensions</h2>
922<!-- ======================================================================= -->
923
924<p>Clang supports some language features conditionally on some targets.</p>
925
926<!-- ======================================================================= -->
927<h3 id="x86-specific">X86/X86-64 Language Extensions</h3>
928<!-- ======================================================================= -->
929
930<p>The X86 backend has these language extensions:</p>
931
932<!-- ======================================================================= -->
933<h4 id="x86-gs-segment">Memory references off the GS segment</h4>
934<!-- ======================================================================= -->
935
936<p>Annotating a pointer with address space #256 causes it to be code generated
Chris Lattnera021e7c2009-05-05 18:54:47 +0000[diff] [blame]937relative to the X86 GS segment register, and address space #257 causes it to be
938relative to the X86 FS segment. Note that this is a very very low-level
939feature that should only be used if you know what you're doing (for example in
940an OS kernel).</p>
Chris Lattner1177f912009-04-09 19:58:15 +0000[diff] [blame]941
942<p>Here is an example:</p>
943
944<pre>
945#define GS_RELATIVE __attribute__((address_space(256)))
946int foo(int GS_RELATIVE *P) {
947 return *P;
948}
949</pre>
950
951<p>Which compiles to (on X86-32):</p>
952
953<pre>
954_foo:
955 movl 4(%esp), %eax
956 movl %gs:(%eax), %eax
957 ret
958</pre>
959
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]960<!-- ======================================================================= -->
961<h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2>
962<!-- ======================================================================= -->
963
964<p>Clang supports additional attributes that are useful for documenting program
965invariants and rules for static analysis tools. The extensions documented here
966are used by the <a
967href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer
968engine</a> that is part of Clang's Analysis library.</p>
969
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]970<h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]971
972<p>Clang's static analysis engine understands the standard <tt>noreturn</tt>
Ted Kremenek4df21142009-04-10 05:04:22 +0000[diff] [blame]973attribute. This attribute, which is typically affixed to a function prototype,
974indicates that a call to a given function never returns. Function prototypes for
975common functions like <tt>exit</tt> are typically annotated with this attribute,
976as well as a variety of common assertion handlers. Users can educate the static
977analyzer about their own custom assertion handles (thus cutting down on false
978positives due to false paths) by marking their own &quot;panic&quot; functions
979with this attribute.</p>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]980
981<p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes
Nick Lewycky625b5862009-06-14 04:08:08 +0000[diff] [blame]982there are special functions that for all intents and purposes should be
983considered panic functions (i.e., they are only called when an internal program
984error occurs) but may actually return so that the program can fail gracefully.
985The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions
986as being interpreted as &quot;no return&quot; functions by the analyzer (thus
Chris Lattner28935892009-04-10 05:54:56 +0000[diff] [blame]987pruning bogus paths) but will not affect compilation (as in the case of
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]988<tt>noreturn</tt>).</p>
989
990<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]991same places where the <tt>noreturn</tt> attribute can be placed. It is commonly
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]992placed at the end of function prototypes:</p>
993
994<pre>
995 void foo() <b>__attribute__((analyzer_noreturn))</b>;
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]996</pre>
997
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]998<p>Query for this feature with
999<tt>__has_attribute(analyzer_noreturn)</tt>.</p>
Chris Lattner148772a2009-06-13 07:13:28 +0000[diff] [blame]1000
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1001<h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3>
1002
1003<p>Many methods in Objective-C have conventional meanings determined
1004by their selectors. For the purposes of static analysis, it is
1005sometimes useful to be able to mark a method as having a particular
1006conventional meaning despite not having the right selector, or as not
1007having the conventional meaning that its selector would suggest.
1008For these use cases, we provide an attribute to specifically describe
1009the <q>method family</q> that a method belongs to.</p>
1010
1011<p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>,
1012where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>,
1013<tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute
1014can only be placed at the end of a method declaration:</p>
1015
1016<pre>
1017 - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>;
1018</pre>
1019
1020<p>Users who do not wish to change the conventional meaning of a
1021method, and who merely want to document its non-standard retain and
1022release semantics, should use the
1023<a href="#attr_retain_release">retaining behavior attributes</a>
1024described below.</p>
1025
1026<p>Query for this feature with
1027<tt>__has_attribute(objc_method_family)</tt>.</p>
1028
1029<h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1030
1031<p>In Objective-C, functions and methods are generally assumed to take
1032and return objects with +0 retain counts, with some exceptions for
1033special methods like <tt>+alloc</tt> and <tt>init</tt>. However,
1034there are exceptions, and so Clang provides attributes to allow these
1035exceptions to be documented, which helps the analyzer find leaks (and
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1036ignore non-leaks). Some exceptions may be better described using
1037the <a href="#attr_method_family"><tt>objc_method_family</tt></a>
1038attribute instead.</p>
John McCall630b7ae2011-01-25 04:26:21 +0000[diff] [blame]1039
1040<p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>,
1041<tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>,
1042and <tt>cf_returns_not_retained</tt> attributes can be placed on
1043methods and functions that return Objective-C or CoreFoundation
1044objects. They are commonly placed at the end of a function prototype
1045or method declaration:</p>
1046
1047<pre>
1048 id foo() <b>__attribute__((ns_returns_retained))</b>;
1049
1050 - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>;
1051</pre>
1052
1053<p>The <tt>*_returns_retained</tt> attributes specify that the
1054returned object has a +1 retain count.
1055The <tt>*_returns_not_retained</tt> attributes specify that the return
1056object has a +0 retain count, even if the normal convention for its
1057selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the
1058returned object is +0, but is guaranteed to live at least as long as the
1059next flush of an autorelease pool.</p>
1060
1061<p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt>
1062attributes can be placed on an parameter declaration; they specify
1063that the argument is expected to have a +1 retain count, which will be
1064balanced in some way by the function or method.
1065The <tt>ns_consumes_self</tt> attribute can only be placed on an
1066Objective-C method; it specifies that the method expects
1067its <tt>self</tt> parameter to have a +1 retain count, which it will
1068balance in some way.</p>
1069
1070<pre>
1071 void <b>foo(__attribute__((ns_consumed))</b> NSString *string);
1072
1073 - (void) bar <b>__attribute__((ns_consumes_self))</b>;
1074 - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x;
1075</pre>
Ted Kremeneked869312009-04-10 05:03:33 +0000[diff] [blame]1076
John McCall87494012011-03-18 03:51:49 +0000[diff] [blame]1077<p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>,
1078<tt>__has_attribute(ns_returns_retained)</tt>, etc.</p>
1079
Chris Lattner5ce933f2009-02-09 08:46:11 +0000[diff] [blame]1080</div>
1081</body>
1082</html>