Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 1 | <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" |
| 2 | "http://www.w3.org/TR/html4/strict.dtd"> |
| 3 | <!-- Material used from: HTML 4.01 specs: http://www.w3.org/TR/html401/ --> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 4 | <html> |
| 5 | <head> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 6 | <META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1"> |
Eli Friedman | 0c706c2 | 2011-09-19 23:17:44 +0000 | [diff] [blame] | 7 | <title>Clang Language Extensions</title> |
Sean Hunt | 64f857b | 2011-06-23 01:22:53 +0000 | [diff] [blame] | 8 | <link type="text/css" rel="stylesheet" href="../menu.css"> |
| 9 | <link type="text/css" rel="stylesheet" href="../content.css"> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 10 | <style type="text/css"> |
| 11 | td { |
| 12 | vertical-align: top; |
| 13 | } |
| 14 | </style> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 15 | </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 Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 26 | <li><a href="#feature_check">Feature Checking Macros</a></li> |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 27 | <li><a href="#has_include">Include File Checking Macros</a></li> |
Chris Lattner | 81edc9f | 2009-04-13 02:45:46 +0000 | [diff] [blame] | 28 | <li><a href="#builtinmacros">Builtin Macros</a></li> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 29 | <li><a href="#vectors">Vectors and Extended Vectors</a></li> |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 30 | <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 Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 32 | <li><a href="#checking_language_features">Checks for Standard Language Features</a> |
Ted Kremenek | 22c3410 | 2009-12-03 02:05:57 +0000 | [diff] [blame] | 33 | <ul> |
| 34 | <li><a href="#cxx_exceptions">C++ exceptions</a></li> |
| 35 | <li><a href="#cxx_rtti">C++ RTTI</a></li> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 36 | </ul></li> |
| 37 | <li><a href="#checking_upcoming_features">Checks for Upcoming Standard Language Features</a> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 38 | <ul> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 39 | <li><a href="#cxx0x">C++0x</a> |
| 40 | <ul> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 41 | <li><a href="#cxx_access_control_sfinae">C++0x SFINAE includes access control</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 42 | <li><a href="#cxx_alias_templates">C++0x alias templates</a></li> |
Peter Collingbourne | fd5f686 | 2011-10-14 23:44:46 +0000 | [diff] [blame^] | 43 | <li><a href="#cxx_alignas">C++0x alignment specifiers</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 44 | <li><a href="#cxx_attributes">C++0x attributes</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 45 | <li><a href="#cxx_constexpr">C++0x generalized constant expressions</a></li> |
| 46 | <li><a href="#cxx_decltype">C++0x <tt>decltype()</tt></a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 47 | <li><a href="#cxx_default_function_template_args">C++0x default template arguments in function templates</a></li> |
Sean Hunt | d962499 | 2011-06-23 06:11:37 +0000 | [diff] [blame] | 48 | <li><a href="#cxx_delegating_constructor">C++0x delegating constructors</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 49 | <li><a href="#cxx_deleted_functions">C++0x deleted functions</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 50 | <li><a href="#cxx_explicit_conversions">C++0x explicit conversion functions</a></li> |
Sean Hunt | e1f6dea | 2011-08-07 00:34:32 +0000 | [diff] [blame] | 51 | <li><a href="#cxx_generalized_initializers">C++0x generalized initializers</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 52 | <li><a href="#cxx_implicit_moves">C++0x implicit move constructors/assignment operators</a></li> |
| 53 | <li><a href="#cxx_inheriting_constructors">C++0x inheriting constructors</a></li> |
| 54 | <li><a href="#cxx_inline_namespaces">C++0x inline namespaces</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 55 | <li><a href="#cxx_lambdas">C++0x lambdas</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 56 | <li><a href="#cxx_noexcept">C++0x noexcept specification</a></li> |
| 57 | <li><a href="#cxx_nonstatic_member_init">C++0x in-class non-static data member initialization</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 58 | <li><a href="#cxx_nullptr">C++0x nullptr</a></li> |
| 59 | <li><a href="#cxx_override_control">C++0x override control</a></li> |
| 60 | <li><a href="#cxx_range_for">C++0x range-based for loop</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 61 | <li><a href="#cxx_raw_string_literals">C++0x raw string literals</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 62 | <li><a href="#cxx_rvalue_references">C++0x rvalue references</a></li> |
| 63 | <li><a href="#cxx_reference_qualified_functions">C++0x reference-qualified functions</a></li> |
| 64 | <li><a href="#cxx_static_assert">C++0x <tt>static_assert()</tt></a></li> |
| 65 | <li><a href="#cxx_auto_type">C++0x type inference</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 66 | <li><a href="#cxx_strong_enums">C++0x strongly-typed enumerations</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 67 | <li><a href="#cxx_trailing_return">C++0x trailing return type</a></li> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 68 | <li><a href="#cxx_unicode_literals">C++0x Unicode string literals</a></li> |
| 69 | <li><a href="#cxx_unrestricted_unions">C++0x unrestricted unions</a></li> |
| 70 | <li><a href="#cxx_user_literals">C++0x user-defined literals</a></li> |
| 71 | <li><a href="#cxx_variadic_templates">C++0x variadic templates</a></li> |
| 72 | </ul></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 73 | <li><a href="#c1x">C1X</a> |
| 74 | <ul> |
Peter Collingbourne | fd5f686 | 2011-10-14 23:44:46 +0000 | [diff] [blame^] | 75 | <li><a href="#c_alignas">C1X alignment specifiers</a></li> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 76 | <li><a href="#c_generic_selections">C1X generic selections</a></li> |
| 77 | <li><a href="#c_static_assert">C1X <tt>_Static_assert()</tt></a></li> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 78 | </ul></li> |
| 79 | </ul> </li> |
Douglas Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 80 | <li><a href="#checking_type_traits">Checks for Type Traits</a></li> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 81 | <li><a href="#blocks">Blocks</a></li> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 82 | <li><a href="#objc_features">Objective-C Features</a> |
| 83 | <ul> |
| 84 | <li><a href="#objc_instancetype">Related result types</a></li> |
John McCall | f85e193 | 2011-06-15 23:02:42 +0000 | [diff] [blame] | 85 | <li><a href="#objc_arc">Automatic reference counting</a></li> |
Douglas Gregor | 5471bc8 | 2011-09-08 17:18:35 +0000 | [diff] [blame] | 86 | <li><a href="#objc_fixed_enum">Enumerations with a fixed underlying type</a></li> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 87 | </ul> |
| 88 | </li> |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 89 | <li><a href="#overloading-in-c">Function Overloading in C</a></li> |
Eli Friedman | 0c706c2 | 2011-09-19 23:17:44 +0000 | [diff] [blame] | 90 | <li><a href="#complex-list-init">Initializer lists for complex numbers in C</a></li> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 91 | <li><a href="#builtins">Builtin Functions</a> |
| 92 | <ul> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 93 | <li><a href="#__builtin_shufflevector">__builtin_shufflevector</a></li> |
Chris Lattner | 21190d5 | 2009-09-21 03:09:59 +0000 | [diff] [blame] | 94 | <li><a href="#__builtin_unreachable">__builtin_unreachable</a></li> |
Chris Lattner | 23aa9c8 | 2011-04-09 03:57:26 +0000 | [diff] [blame] | 95 | <li><a href="#__sync_swap">__sync_swap</a></li> |
Douglas Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 96 | </ul> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 97 | </li> |
Chris Lattner | 1177f91 | 2009-04-09 19:58:15 +0000 | [diff] [blame] | 98 | <li><a href="#targetspecific">Target-Specific Extensions</a> |
| 99 | <ul> |
| 100 | <li><a href="#x86-specific">X86/X86-64 Language Extensions</a></li> |
| 101 | </ul> |
| 102 | </li> |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 103 | <li><a href="#analyzerspecific">Static Analysis-Specific Extensions</a></li> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 104 | <li><a href="#threadsafety">Thread Safety Annotation Checking</a></li> |
| 105 | <ul> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 106 | <li><a href="#ts_noanal"><tt>no_thread_safety_analysis</tt></a></li> |
| 107 | <li><a href="#ts_lockable"><tt>lockable</tt></a></li> |
| 108 | <li><a href="#ts_scopedlockable"><tt>scoped_lockable</tt></a></li> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 109 | <li><a href="#ts_guardedvar"><tt>guarded_var</tt></a></li> |
| 110 | <li><a href="#ts_ptguardedvar"><tt>pt_guarded_var</tt></a></li> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 111 | <li><a href="#ts_guardedby"><tt>guarded_by(l)</tt></a></li> |
| 112 | <li><a href="#ts_ptguardedby"><tt>pt_guarded_by(l)</tt></a></li> |
| 113 | <li><a href="#ts_acquiredbefore"><tt>acquired_before(...)</tt></a></li> |
| 114 | <li><a href="#ts_acquiredafter"><tt>acquired_after(...)</tt></a></li> |
| 115 | <li><a href="#ts_elf"><tt>exclusive_lock_function(...)</tt></a></li> |
| 116 | <li><a href="#ts_slf"><tt>shared_lock_function(...)</tt></a></li> |
| 117 | <li><a href="#ts_etf"><tt>exclusive_trylock_function(...)</tt></a></li> |
| 118 | <li><a href="#ts_stf"><tt>shared_trylock_function(...)</tt></a></li> |
| 119 | <li><a href="#ts_uf"><tt>unlock_function(...)</tt></a></li> |
| 120 | <li><a href="#ts_lr"><tt>lock_returned(l)</tt></a></li> |
| 121 | <li><a href="#ts_le"><tt>locks_excluded(...)</tt></a></li> |
| 122 | <li><a href="#ts_elr"><tt>exclusive_locks_required(...)</tt></a></li> |
| 123 | <li><a href="#ts_slr"><tt>shared_locks_required(...)</tt></a></li> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 124 | </ul> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 125 | </ul> |
| 126 | |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 127 | <!-- ======================================================================= --> |
| 128 | <h2 id="intro">Introduction</h2> |
| 129 | <!-- ======================================================================= --> |
| 130 | |
| 131 | <p>This document describes the language extensions provided by Clang. In |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 132 | addition to the language extensions listed here, Clang aims to support a broad |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 133 | range of GCC extensions. Please see the <a |
| 134 | href="http://gcc.gnu.org/onlinedocs/gcc/C-Extensions.html">GCC manual</a> for |
| 135 | more information on these extensions.</p> |
| 136 | |
| 137 | <!-- ======================================================================= --> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 138 | <h2 id="feature_check">Feature Checking Macros</h2> |
| 139 | <!-- ======================================================================= --> |
| 140 | |
| 141 | <p>Language extensions can be very useful, but only if you know you can depend |
Chris Lattner | c70e193 | 2011-03-21 16:25:11 +0000 | [diff] [blame] | 142 | on them. In order to allow fine-grain features checks, we support three builtin |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 143 | function-like macros. This allows you to directly test for a feature in your |
| 144 | code without having to resort to something like autoconf or fragile "compiler |
| 145 | version checks".</p> |
| 146 | |
| 147 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 148 | <h3><a name="__has_builtin">__has_builtin</a></h3> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 149 | <!-- ======================================================================= --> |
| 150 | |
| 151 | <p>This function-like macro takes a single identifier argument that is the name |
| 152 | of a builtin function. It evaluates to 1 if the builtin is supported or 0 if |
| 153 | not. It can be used like this:</p> |
| 154 | |
| 155 | <blockquote> |
| 156 | <pre> |
| 157 | #ifndef __has_builtin // Optional of course. |
| 158 | #define __has_builtin(x) 0 // Compatibility with non-clang compilers. |
| 159 | #endif |
| 160 | |
| 161 | ... |
| 162 | #if __has_builtin(__builtin_trap) |
| 163 | __builtin_trap(); |
| 164 | #else |
| 165 | abort(); |
| 166 | #endif |
| 167 | ... |
| 168 | </pre> |
| 169 | </blockquote> |
| 170 | |
| 171 | |
| 172 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 173 | <h3><a name="__has_feature_extension"> __has_feature and __has_extension</a></h3> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 174 | <!-- ======================================================================= --> |
| 175 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 176 | <p>These function-like macros take a single identifier argument that is the |
| 177 | name of a feature. <code>__has_feature</code> evaluates to 1 if the feature |
| 178 | is both supported by Clang and standardized in the current language standard |
| 179 | or 0 if not (but see <a href="#has_feature_back_compat">below</a>), while |
| 180 | <code>__has_extension</code> evaluates to 1 if the feature is supported by |
| 181 | Clang in the current language (either as a language extension or a standard |
| 182 | language feature) or 0 if not. They can be used like this:</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 183 | |
| 184 | <blockquote> |
| 185 | <pre> |
| 186 | #ifndef __has_feature // Optional of course. |
| 187 | #define __has_feature(x) 0 // Compatibility with non-clang compilers. |
| 188 | #endif |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 189 | #ifndef __has_extension |
| 190 | #define __has_extension __has_feature // Compatibility with pre-3.0 compilers. |
| 191 | #endif |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 192 | |
| 193 | ... |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 194 | #if __has_feature(cxx_rvalue_references) |
| 195 | // This code will only be compiled with the -std=c++0x and -std=gnu++0x |
| 196 | // options, because rvalue references are only standardized in C++0x. |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 197 | #endif |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 198 | |
| 199 | #if __has_extension(cxx_rvalue_references) |
| 200 | // This code will be compiled with the -std=c++0x, -std=gnu++0x, -std=c++98 |
| 201 | // and -std=gnu++98 options, because rvalue references are supported as a |
| 202 | // language extension in C++98. |
| 203 | #endif |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 204 | </pre> |
| 205 | </blockquote> |
| 206 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 207 | <p id="has_feature_back_compat">For backwards compatibility reasons, |
| 208 | <code>__has_feature</code> can also be used to test for support for |
| 209 | non-standardized features, i.e. features not prefixed <code>c_</code>, |
| 210 | <code>cxx_</code> or <code>objc_</code>.</p> |
| 211 | |
| 212 | <p>If the <code>-pedantic-errors</code> option is given, |
| 213 | <code>__has_extension</code> is equivalent to <code>__has_feature</code>.</p> |
| 214 | |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 215 | <p>The feature tag is described along with the language feature below.</p> |
| 216 | |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 217 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 218 | <h3><a name="__has_attribute">__has_attribute</a></h3> |
Anders Carlsson | cae5095 | 2010-10-20 02:31:43 +0000 | [diff] [blame] | 219 | <!-- ======================================================================= --> |
| 220 | |
| 221 | <p>This function-like macro takes a single identifier argument that is the name |
| 222 | of an attribute. It evaluates to 1 if the attribute is supported or 0 if not. It |
| 223 | can be used like this:</p> |
| 224 | |
| 225 | <blockquote> |
| 226 | <pre> |
| 227 | #ifndef __has_attribute // Optional of course. |
| 228 | #define __has_attribute(x) 0 // Compatibility with non-clang compilers. |
| 229 | #endif |
| 230 | |
| 231 | ... |
Anders Carlsson | 961003d | 2011-01-24 03:54:51 +0000 | [diff] [blame] | 232 | #if __has_attribute(always_inline) |
| 233 | #define ALWAYS_INLINE __attribute__((always_inline)) |
Anders Carlsson | cae5095 | 2010-10-20 02:31:43 +0000 | [diff] [blame] | 234 | #else |
Anders Carlsson | 961003d | 2011-01-24 03:54:51 +0000 | [diff] [blame] | 235 | #define ALWAYS_INLINE |
Anders Carlsson | cae5095 | 2010-10-20 02:31:43 +0000 | [diff] [blame] | 236 | #endif |
| 237 | ... |
| 238 | </pre> |
| 239 | </blockquote> |
| 240 | |
| 241 | <!-- ======================================================================= --> |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 242 | <h2 id="has_include">Include File Checking Macros</h2> |
| 243 | <!-- ======================================================================= --> |
| 244 | |
| 245 | <p>Not all developments systems have the same include files. |
| 246 | The <a href="#__has_include">__has_include</a> and |
| 247 | <a href="#__has_include_next">__has_include_next</a> macros allow you to |
| 248 | check for the existence of an include file before doing |
| 249 | a possibly failing #include directive.</p> |
| 250 | |
| 251 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 252 | <h3><a name="__has_include">__has_include</a></h3> |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 253 | <!-- ======================================================================= --> |
| 254 | |
| 255 | <p>This function-like macro takes a single file name string argument that |
| 256 | is the name of an include file. It evaluates to 1 if the file can |
| 257 | be found using the include paths, or 0 otherwise:</p> |
| 258 | |
| 259 | <blockquote> |
| 260 | <pre> |
| 261 | // Note the two possible file name string formats. |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 262 | #if __has_include("myinclude.h") && __has_include(<stdint.h>) |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 263 | # include "myinclude.h" |
| 264 | #endif |
| 265 | |
| 266 | // To avoid problem with non-clang compilers not having this macro. |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 267 | #if defined(__has_include) && __has_include("myinclude.h") |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 268 | # include "myinclude.h" |
| 269 | #endif |
| 270 | </pre> |
| 271 | </blockquote> |
| 272 | |
| 273 | <p>To test for this feature, use #if defined(__has_include).</p> |
| 274 | |
| 275 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 276 | <h3><a name="__has_include_next">__has_include_next</a></h3> |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 277 | <!-- ======================================================================= --> |
| 278 | |
| 279 | <p>This function-like macro takes a single file name string argument that |
| 280 | is the name of an include file. It is like __has_include except that it |
| 281 | looks for the second instance of the given file found in the include |
| 282 | paths. It evaluates to 1 if the second instance of the file can |
| 283 | be found using the include paths, or 0 otherwise:</p> |
| 284 | |
| 285 | <blockquote> |
| 286 | <pre> |
| 287 | // Note the two possible file name string formats. |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 288 | #if __has_include_next("myinclude.h") && __has_include_next(<stdint.h>) |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 289 | # include_next "myinclude.h" |
| 290 | #endif |
| 291 | |
| 292 | // To avoid problem with non-clang compilers not having this macro. |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 293 | #if defined(__has_include_next) && __has_include_next("myinclude.h") |
John Thompson | 92bd8c7 | 2009-11-02 22:28:12 +0000 | [diff] [blame] | 294 | # include_next "myinclude.h" |
| 295 | #endif |
| 296 | </pre> |
| 297 | </blockquote> |
| 298 | |
| 299 | <p>Note that __has_include_next, like the GNU extension |
| 300 | #include_next directive, is intended for use in headers only, |
| 301 | and will issue a warning if used in the top-level compilation |
| 302 | file. A warning will also be issued if an absolute path |
| 303 | is used in the file argument.</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 304 | |
Ted Kremenek | d768150 | 2011-10-12 19:46:30 +0000 | [diff] [blame] | 305 | |
| 306 | <!-- ======================================================================= --> |
| 307 | <h3><a name="__has_warning">__has_warning</a></h3> |
| 308 | <!-- ======================================================================= --> |
| 309 | |
| 310 | <p>This function-like macro takes a string literal that represents a command |
| 311 | line option for a warning and returns true if that is a valid warning |
| 312 | option.</p> |
| 313 | |
| 314 | <blockquote> |
| 315 | <pre> |
| 316 | #if __has_warning("-Wformat") |
| 317 | ... |
| 318 | #endif |
| 319 | </pre> |
| 320 | </blockquote> |
| 321 | |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 322 | <!-- ======================================================================= --> |
Chris Lattner | 81edc9f | 2009-04-13 02:45:46 +0000 | [diff] [blame] | 323 | <h2 id="builtinmacros">Builtin Macros</h2> |
| 324 | <!-- ======================================================================= --> |
| 325 | |
Douglas Gregor | 4290fbd | 2010-04-30 02:51:06 +0000 | [diff] [blame] | 326 | <dl> |
| 327 | <dt><code>__BASE_FILE__</code></dt> |
| 328 | <dd>Defined to a string that contains the name of the main input |
| 329 | file passed to Clang.</dd> |
| 330 | |
| 331 | <dt><code>__COUNTER__</code></dt> |
| 332 | <dd>Defined to an integer value that starts at zero and is |
| 333 | incremented each time the <code>__COUNTER__</code> macro is |
| 334 | expanded.</dd> |
| 335 | |
| 336 | <dt><code>__INCLUDE_LEVEL__</code></dt> |
| 337 | <dd>Defined to an integral value that is the include depth of the |
| 338 | file currently being translated. For the main file, this value is |
| 339 | zero.</dd> |
| 340 | |
| 341 | <dt><code>__TIMESTAMP__</code></dt> |
| 342 | <dd>Defined to the date and time of the last modification of the |
| 343 | current source file.</dd> |
| 344 | |
| 345 | <dt><code>__clang__</code></dt> |
| 346 | <dd>Defined when compiling with Clang</dd> |
| 347 | |
| 348 | <dt><code>__clang_major__</code></dt> |
| 349 | <dd>Defined to the major version number of Clang (e.g., the 2 in |
| 350 | 2.0.1).</dd> |
| 351 | |
| 352 | <dt><code>__clang_minor__</code></dt> |
| 353 | <dd>Defined to the minor version number of Clang (e.g., the 0 in |
| 354 | 2.0.1).</dd> |
| 355 | |
| 356 | <dt><code>__clang_patchlevel__</code></dt> |
| 357 | <dd>Defined to the patch level of Clang (e.g., the 1 in 2.0.1).</dd> |
| 358 | |
| 359 | <dt><code>__clang_version__</code></dt> |
| 360 | <dd>Defined to a string that captures the Clang version, including |
| 361 | the Subversion tag or revision number, e.g., "1.5 (trunk |
| 362 | 102332)".</dd> |
| 363 | </dl> |
Chris Lattner | 81edc9f | 2009-04-13 02:45:46 +0000 | [diff] [blame] | 364 | |
| 365 | <!-- ======================================================================= --> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 366 | <h2 id="vectors">Vectors and Extended Vectors</h2> |
| 367 | <!-- ======================================================================= --> |
| 368 | |
Owen Anderson | d2bf0cd | 2010-01-27 01:22:36 +0000 | [diff] [blame] | 369 | <p>Supports the GCC vector extensions, plus some stuff like V[1].</p> |
| 370 | |
| 371 | <p>Also supports <tt>ext_vector</tt>, which additionally support for V.xyzw |
| 372 | syntax and other tidbits as seen in OpenCL. An example is:</p> |
| 373 | |
| 374 | <blockquote> |
| 375 | <pre> |
| 376 | typedef float float4 <b>__attribute__((ext_vector_type(4)))</b>; |
| 377 | typedef float float2 <b>__attribute__((ext_vector_type(2)))</b>; |
| 378 | |
| 379 | float4 foo(float2 a, float2 b) { |
| 380 | float4 c; |
| 381 | c.xz = a; |
| 382 | c.yw = b; |
| 383 | return c; |
| 384 | } |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 385 | </pre> |
Owen Anderson | d2bf0cd | 2010-01-27 01:22:36 +0000 | [diff] [blame] | 386 | </blockquote> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 387 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 388 | <p>Query for this feature with __has_extension(attribute_ext_vector_type).</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 389 | |
Owen Anderson | d2bf0cd | 2010-01-27 01:22:36 +0000 | [diff] [blame] | 390 | <p>See also <a href="#__builtin_shufflevector">__builtin_shufflevector</a>.</p> |
| 391 | |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 392 | <!-- ======================================================================= --> |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 393 | <h2 id="deprecated">Messages on <tt>deprecated</tt> and <tt>unavailable</tt> Attributes</h2> |
Fariborz Jahanian | c784dc1 | 2010-10-06 23:12:32 +0000 | [diff] [blame] | 394 | <!-- ======================================================================= --> |
| 395 | |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 396 | <p>An optional string message can be added to the <tt>deprecated</tt> |
| 397 | and <tt>unavailable</tt> attributes. For example:</p> |
Fariborz Jahanian | c784dc1 | 2010-10-06 23:12:32 +0000 | [diff] [blame] | 398 | |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 399 | <blockquote> |
Chris Lattner | 4836d6a | 2010-11-09 19:43:35 +0000 | [diff] [blame] | 400 | <pre>void explode(void) __attribute__((deprecated("extremely unsafe, use 'combust' instead!!!")));</pre> |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 401 | </blockquote> |
| 402 | |
| 403 | <p>If the deprecated or unavailable declaration is used, the message |
| 404 | will be incorporated into the appropriate diagnostic:</p> |
| 405 | |
| 406 | <blockquote> |
Chris Lattner | 4836d6a | 2010-11-09 19:43:35 +0000 | [diff] [blame] | 407 | <pre>harmless.c:4:3: warning: 'explode' is deprecated: extremely unsafe, use 'combust' instead!!! [-Wdeprecated-declarations] |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 408 | explode(); |
| 409 | ^</pre> |
| 410 | </blockquote> |
| 411 | |
| 412 | <p>Query for this feature |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 413 | with <tt>__has_extension(attribute_deprecated_with_message)</tt> |
| 414 | and <tt>__has_extension(attribute_unavailable_with_message)</tt>.</p> |
John McCall | 4820908 | 2010-11-08 19:48:17 +0000 | [diff] [blame] | 415 | |
| 416 | <!-- ======================================================================= --> |
| 417 | <h2 id="attributes-on-enumerators">Attributes on Enumerators</h2> |
| 418 | <!-- ======================================================================= --> |
| 419 | |
| 420 | <p>Clang allows attributes to be written on individual enumerators. |
| 421 | This allows enumerators to be deprecated, made unavailable, etc. The |
| 422 | attribute must appear after the enumerator name and before any |
| 423 | initializer, like so:</p> |
| 424 | |
| 425 | <blockquote> |
| 426 | <pre>enum OperationMode { |
| 427 | OM_Invalid, |
| 428 | OM_Normal, |
| 429 | OM_Terrified __attribute__((deprecated)), |
| 430 | OM_AbortOnError __attribute__((deprecated)) = 4 |
| 431 | };</pre> |
| 432 | </blockquote> |
| 433 | |
| 434 | <p>Attributes on the <tt>enum</tt> declaration do not apply to |
| 435 | individual enumerators.</p> |
| 436 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 437 | <p>Query for this feature with <tt>__has_extension(enumerator_attributes)</tt>.</p> |
Fariborz Jahanian | c784dc1 | 2010-10-06 23:12:32 +0000 | [diff] [blame] | 438 | |
| 439 | <!-- ======================================================================= --> |
Ted Kremenek | 87774fd | 2009-12-03 02:04:01 +0000 | [diff] [blame] | 440 | <h2 id="checking_language_features">Checks for Standard Language Features</h2> |
| 441 | <!-- ======================================================================= --> |
| 442 | |
| 443 | <p>The <tt>__has_feature</tt> macro can be used to query if certain standard language features are |
| 444 | enabled. Those features are listed here.</p> |
| 445 | |
Ted Kremenek | 22c3410 | 2009-12-03 02:05:57 +0000 | [diff] [blame] | 446 | <h3 id="cxx_exceptions">C++ exceptions</h3> |
Ted Kremenek | 87774fd | 2009-12-03 02:04:01 +0000 | [diff] [blame] | 447 | |
Ted Kremenek | 22c3410 | 2009-12-03 02:05:57 +0000 | [diff] [blame] | 448 | <p>Use <tt>__has_feature(cxx_exceptions)</tt> to determine if C++ exceptions have been enabled. For |
Sean Hunt | 647ba1b | 2011-06-23 00:42:53 +0000 | [diff] [blame] | 449 | example, compiling code with <tt>-fexceptions</tt> enables C++ exceptions.</p> |
Ted Kremenek | 87774fd | 2009-12-03 02:04:01 +0000 | [diff] [blame] | 450 | |
Ted Kremenek | 22c3410 | 2009-12-03 02:05:57 +0000 | [diff] [blame] | 451 | <h3 id="cxx_rtti">C++ RTTI</h3> |
Ted Kremenek | 87774fd | 2009-12-03 02:04:01 +0000 | [diff] [blame] | 452 | |
Ted Kremenek | 0eb9560 | 2009-12-03 02:06:43 +0000 | [diff] [blame] | 453 | <p>Use <tt>__has_feature(cxx_rtti)</tt> to determine if C++ RTTI has been enabled. For example, |
Ted Kremenek | 22c3410 | 2009-12-03 02:05:57 +0000 | [diff] [blame] | 454 | compiling code with <tt>-fno-rtti</tt> disables the use of RTTI.</p> |
Ted Kremenek | 87774fd | 2009-12-03 02:04:01 +0000 | [diff] [blame] | 455 | |
| 456 | <!-- ======================================================================= --> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 457 | <h2 id="checking_upcoming_features">Checks for Upcoming Standard Language Features</h2> |
| 458 | <!-- ======================================================================= --> |
| 459 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 460 | <p>The <tt>__has_feature</tt> or <tt>__has_extension</tt> macros can be used |
| 461 | to query if certain upcoming standard language features are enabled. Those |
| 462 | features are listed here. Features that are not yet implemented will be |
| 463 | noted.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 464 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 465 | <h3 id="cxx0x">C++0x</h3> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 466 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 467 | <p>The features listed below are slated for inclusion in the upcoming |
| 468 | C++0x standard. As a result, all these features are enabled |
| 469 | with the <tt>-std=c++0x</tt> option when compiling C++ code.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 470 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 471 | <h4 id="cxx_access_control_sfinae">C++0x SFINAE includes access control</h4> |
Douglas Gregor | 7822ee3 | 2011-05-11 23:45:11 +0000 | [diff] [blame] | 472 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 473 | <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 Gregor | 7822ee3 | 2011-05-11 23:45:11 +0000 | [diff] [blame] | 474 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 475 | <h4 id="cxx_alias_templates">C++0x alias templates</h4> |
Richard Smith | 3e4c6c4 | 2011-05-05 21:57:07 +0000 | [diff] [blame] | 476 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 477 | <p>Use <tt>__has_feature(cxx_alias_templates)</tt> or |
| 478 | <tt>__has_extension(cxx_alias_templates)</tt> to determine if support for |
Richard Smith | 3e4c6c4 | 2011-05-05 21:57:07 +0000 | [diff] [blame] | 479 | C++0x's alias declarations and alias templates is enabled.</p> |
| 480 | |
Peter Collingbourne | fd5f686 | 2011-10-14 23:44:46 +0000 | [diff] [blame^] | 481 | <h4 id="cxx_alignas">C++0x alignment specifiers</h4> |
| 482 | |
| 483 | <p>Use <tt>__has_feature(cxx_alignas)</tt> or |
| 484 | <tt>__has_extension(cxx_alignas)</tt> to determine if support for alignment |
| 485 | specifiers using <tt>alignas</tt> is enabled.</p> |
| 486 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 487 | <h4 id="cxx_attributes">C++0x attributes</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 488 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 489 | <p>Use <tt>__has_feature(cxx_attributes)</tt> or |
| 490 | <tt>__has_extension(cxx_attributes)</tt> to determine if support for attribute |
| 491 | parsing with C++0x's square bracket notation is enabled.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 492 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 493 | <h4 id="cxx_constexpr">C++0x generalized constant expressions</h4> |
| 494 | |
| 495 | <p>Use <tt>__has_feature(cxx_constexpr)</tt> to determine if support |
| 496 | for generalized constant expressions (e.g., <tt>constexpr</tt>) is |
| 497 | enabled. Clang does not currently implement this feature.</p> |
| 498 | |
| 499 | <h4 id="cxx_decltype">C++0x <tt>decltype()</tt></h4> |
| 500 | |
| 501 | <p>Use <tt>__has_feature(cxx_decltype)</tt> or |
| 502 | <tt>__has_extension(cxx_decltype)</tt> to determine if support for the |
| 503 | <tt>decltype()</tt> specifier is enabled.</p> |
| 504 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 505 | <h4 id="cxx_default_function_template_args">C++0x default template arguments in function templates</h4> |
Douglas Gregor | 0750800 | 2011-02-05 20:35:30 +0000 | [diff] [blame] | 506 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 507 | <p>Use <tt>__has_feature(cxx_default_function_template_args)</tt> or |
| 508 | <tt>__has_extension(cxx_default_function_template_args)</tt> to determine |
| 509 | if support for default template arguments in function templates is enabled.</p> |
Douglas Gregor | 0750800 | 2011-02-05 20:35:30 +0000 | [diff] [blame] | 510 | |
Sean Hunt | d962499 | 2011-06-23 06:11:37 +0000 | [diff] [blame] | 511 | <h4 id="cxx_delegating_constructors">C++0x delegating constructors</h4> |
| 512 | |
| 513 | <p>Use <tt>__has_feature(cxx_delegating_constructors)</tt> to determine if |
| 514 | support for delegating constructors is enabled.</p> |
| 515 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 516 | <h4 id="cxx_deleted_functions">C++0x <tt>delete</tt>d functions</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 517 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 518 | <p>Use <tt>__has_feature(cxx_deleted_functions)</tt> or |
| 519 | <tt>__has_extension(cxx_deleted_functions)</tt> to determine if support for |
Sebastian Redl | f6c0977 | 2010-08-31 23:28:47 +0000 | [diff] [blame] | 520 | deleted function definitions (with <tt>= delete</tt>) is enabled.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 521 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 522 | <h4 id="cxx_explicit_conversions">C++0x explicit conversion functions</h3> |
| 523 | <p>Use <tt>__has_feature(cxx_explicit_conversions)</tt> to determine if support for <tt>explicit</tt> conversion functions is enabled.</p> |
| 524 | |
Sean Hunt | e1f6dea | 2011-08-07 00:34:32 +0000 | [diff] [blame] | 525 | <h4 id="cxx_generalized_initializers">C++0x generalized initializers</h4> |
| 526 | |
| 527 | <p>Use <tt>__has_feature(cxx_generalized_initializers)</tt> to determine if |
| 528 | support for generalized initializers (using braced lists and |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 529 | <tt>std::initializer_list</tt>) is enabled. Clang does not currently implement |
| 530 | this feature.</p> |
| 531 | |
| 532 | <h4 id="cxx_implicit_moves">C++0x implicit move constructors/assignment operators</h4> |
| 533 | |
Sebastian Redl | 72a81d2 | 2011-10-10 18:10:00 +0000 | [diff] [blame] | 534 | <p>Use <tt>__has_feature(cxx_implicit_moves)</tt> to determine if Clang will |
| 535 | implicitly generate move constructors and move assignment operators where needed.</p> |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 536 | |
| 537 | <h4 id="cxx_inheriting_constructors">C++0x inheriting constructors</h4> |
| 538 | |
| 539 | <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> |
| 540 | |
| 541 | <h4 id="cxx_inline_namespaces">C++0x inline namespaces</h4> |
| 542 | |
| 543 | <p>Use <tt>__has_feature(cxx_inline_namespaces)</tt> or |
| 544 | <tt>__has_extension(cxx_inline_namespaces)</tt> to determine if support for |
| 545 | inline namespaces is enabled.</p> |
Sean Hunt | e1f6dea | 2011-08-07 00:34:32 +0000 | [diff] [blame] | 546 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 547 | <h4 id="cxx_lambdas">C++0x lambdas</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 548 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 549 | <p>Use <tt>__has_feature(cxx_lambdas)</tt> or |
| 550 | <tt>__has_extension(cxx_lambdas)</tt> to determine if support for lambdas |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 551 | is enabled. Clang does not currently implement this feature.</p> |
| 552 | |
| 553 | <h4 id="cxx_noexcept">C++0x noexcept</h4> |
| 554 | |
| 555 | <p>Use <tt>__has_feature(cxx_noexcept)</tt> or |
| 556 | <tt>__has_extension(cxx_noexcept)</tt> to determine if support for noexcept |
| 557 | exception specifications is enabled.</p> |
| 558 | |
| 559 | <h4 id="cxx_nonstatic_member_init">C++0x in-class non-static data member initialization</h4> |
| 560 | |
| 561 | <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 Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 562 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 563 | <h4 id="cxx_nullptr">C++0x <tt>nullptr</tt></h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 564 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 565 | <p>Use <tt>__has_feature(cxx_nullptr)</tt> or |
| 566 | <tt>__has_extension(cxx_nullptr)</tt> to determine if support for |
Douglas Gregor | 84ee2ee | 2011-05-21 23:15:46 +0000 | [diff] [blame] | 567 | <tt>nullptr</tt> is enabled.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 568 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 569 | <h4 id="cxx_override_control">C++0x <tt>override control</tt></h4> |
Anders Carlsson | c8b9f79 | 2011-03-25 15:04:23 +0000 | [diff] [blame] | 570 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 571 | <p>Use <tt>__has_feature(cxx_override_control)</tt> or |
| 572 | <tt>__has_extension(cxx_override_control)</tt> to determine if support for |
Anders Carlsson | c8b9f79 | 2011-03-25 15:04:23 +0000 | [diff] [blame] | 573 | the override control keywords is enabled.</p> |
| 574 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 575 | <h4 id="cxx_reference_qualified_functions">C++0x reference-qualified functions</h4> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 576 | <p>Use <tt>__has_feature(cxx_reference_qualified_functions)</tt> or |
| 577 | <tt>__has_extension(cxx_reference_qualified_functions)</tt> to determine |
| 578 | if support for reference-qualified functions (e.g., member functions with |
| 579 | <code>&</code> or <code>&&</code> applied to <code>*this</code>) |
| 580 | is enabled.</p> |
Douglas Gregor | 56209ff | 2011-01-26 21:25:54 +0000 | [diff] [blame] | 581 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 582 | <h4 id="cxx_range_for">C++0x range-based <tt>for</tt> loop</h4> |
Richard Smith | a391a46 | 2011-04-15 15:14:40 +0000 | [diff] [blame] | 583 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 584 | <p>Use <tt>__has_feature(cxx_range_for)</tt> or |
| 585 | <tt>__has_extension(cxx_range_for)</tt> to determine if support for the |
| 586 | range-based for loop is enabled. </p> |
Richard Smith | a391a46 | 2011-04-15 15:14:40 +0000 | [diff] [blame] | 587 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 588 | <h4 id="cxx_raw_string_literals">C++0x raw string literals</h4> |
| 589 | <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> |
| 590 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 591 | <h4 id="cxx_rvalue_references">C++0x rvalue references</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 592 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 593 | <p>Use <tt>__has_feature(cxx_rvalue_references)</tt> or |
| 594 | <tt>__has_extension(cxx_rvalue_references)</tt> to determine if support for |
Douglas Gregor | 56209ff | 2011-01-26 21:25:54 +0000 | [diff] [blame] | 595 | rvalue references is enabled. </p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 596 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 597 | <h4 id="cxx_static_assert">C++0x <tt>static_assert()</tt></h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 598 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 599 | <p>Use <tt>__has_feature(cxx_static_assert)</tt> or |
| 600 | <tt>__has_extension(cxx_static_assert)</tt> to determine if support for |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 601 | compile-time assertions using <tt>static_assert</tt> is enabled.</p> |
| 602 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 603 | <h4 id="cxx_auto_type">C++0x type inference</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 604 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 605 | <p>Use <tt>__has_feature(cxx_auto_type)</tt> or |
| 606 | <tt>__has_extension(cxx_auto_type)</tt> to determine C++0x type inference is |
| 607 | supported using the <tt>auto</tt> specifier. If this is disabled, <tt>auto</tt> |
| 608 | will instead be a storage class specifier, as in C or C++98.</p> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 609 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 610 | <h4 id="cxx_strong_enums">C++0x strongly typed enumerations</h4> |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 611 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 612 | <p>Use <tt>__has_feature(cxx_strong_enums)</tt> or |
| 613 | <tt>__has_extension(cxx_strong_enums)</tt> to determine if support for |
| 614 | strongly typed, scoped enumerations is enabled.</p> |
Sebastian Redl | f6c0977 | 2010-08-31 23:28:47 +0000 | [diff] [blame] | 615 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 616 | <h4 id="cxx_trailing_return">C++0x trailing return type</h4> |
Douglas Gregor | dab60ad | 2010-10-01 18:44:50 +0000 | [diff] [blame] | 617 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 618 | <p>Use <tt>__has_feature(cxx_trailing_return)</tt> or |
| 619 | <tt>__has_extension(cxx_trailing_return)</tt> to determine if support for the |
| 620 | alternate function declaration syntax with trailing return type is enabled.</p> |
Douglas Gregor | dab60ad | 2010-10-01 18:44:50 +0000 | [diff] [blame] | 621 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 622 | <h4 id="cxx_unicode_literals">C++0x Unicode string literals</h4> |
| 623 | <p>Use <tt>__has_feature(cxx_unicode_literals)</tt> to determine if |
| 624 | support for Unicode string literals is enabled.</p> |
Sebastian Redl | 4561ecd | 2011-03-15 21:17:12 +0000 | [diff] [blame] | 625 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 626 | <h4 id="cxx_unrestricted_unions">C++0x unrestricted unions</h4> |
Sebastian Redl | 4561ecd | 2011-03-15 21:17:12 +0000 | [diff] [blame] | 627 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 628 | <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 Gregor | 1274ccd | 2010-10-08 23:50:27 +0000 | [diff] [blame] | 629 | |
Douglas Gregor | ece3894 | 2011-08-29 17:28:38 +0000 | [diff] [blame] | 630 | <h4 id="cxx_user_literals">C++0x user-defined literals</h4> |
| 631 | |
| 632 | <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> |
| 633 | |
| 634 | <h4 id="cxx_variadic_templates">C++0x variadic templates</h4> |
| 635 | |
| 636 | <p>Use <tt>__has_feature(cxx_variadic_templates)</tt> or |
| 637 | <tt>__has_extension(cxx_variadic_templates)</tt> to determine if support |
| 638 | for variadic templates is enabled.</p> |
Douglas Gregor | 1274ccd | 2010-10-08 23:50:27 +0000 | [diff] [blame] | 639 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 640 | <h3 id="c1x">C1X</h3> |
| 641 | |
| 642 | <p>The features listed below are slated for inclusion in the upcoming |
| 643 | C1X standard. As a result, all these features are enabled |
| 644 | with the <tt>-std=c1x</tt> option when compiling C code.</p> |
| 645 | |
Peter Collingbourne | fd5f686 | 2011-10-14 23:44:46 +0000 | [diff] [blame^] | 646 | <h4 id="c_alignas">C1X alignment specifiers</h4> |
| 647 | |
| 648 | <p>Use <tt>__has_feature(c_alignas)</tt> or <tt>__has_extension(c_alignas)</tt> |
| 649 | to determine if support for alignment specifiers using <tt>_Alignas</tt> |
| 650 | is enabled.</p> |
| 651 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 652 | <h4 id="c_generic_selections">C1X generic selections</h4> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 653 | |
| 654 | <p>Use <tt>__has_feature(c_generic_selections)</tt> or |
| 655 | <tt>__has_extension(c_generic_selections)</tt> to determine if support for |
| 656 | generic selections is enabled.</p> |
| 657 | |
| 658 | <p>As an extension, the C1X generic selection expression is available in all |
| 659 | languages supported by Clang. The syntax is the same as that given in the |
| 660 | C1X draft standard.</p> |
| 661 | |
| 662 | <p>In C, type compatibility is decided according to the rules given in the |
| 663 | appropriate standard, but in C++, which lacks the type compatibility rules |
| 664 | used in C, types are considered compatible only if they are equivalent.</p> |
| 665 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 666 | <h4 id="c_static_assert">C1X <tt>_Static_assert()</tt></h4> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 667 | |
| 668 | <p>Use <tt>__has_feature(c_static_assert)</tt> or |
| 669 | <tt>__has_extension(c_static_assert)</tt> to determine if support for |
| 670 | compile-time assertions using <tt>_Static_assert</tt> is enabled.</p> |
| 671 | |
Sean Hunt | 4ef4c6b | 2010-01-13 08:31:49 +0000 | [diff] [blame] | 672 | <!-- ======================================================================= --> |
Douglas Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 673 | <h2 id="checking_type_traits">Checks for Type Traits</h2> |
| 674 | <!-- ======================================================================= --> |
| 675 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 676 | <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 Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 677 | <blockquote> |
| 678 | <pre> |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 679 | #if __has_extension(is_convertible_to) |
Douglas Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 680 | template<typename From, typename To> |
| 681 | struct is_convertible_to { |
| 682 | static const bool value = __is_convertible_to(From, To); |
| 683 | }; |
| 684 | #else |
| 685 | // Emulate type trait |
| 686 | #endif |
| 687 | </pre> |
| 688 | </blockquote> |
| 689 | |
| 690 | <p>The following type traits are supported by Clang:</p> |
| 691 | <ul> |
| 692 | <li><code>__has_nothrow_assign</code> (GNU, Microsoft)</li> |
| 693 | <li><code>__has_nothrow_copy</code> (GNU, Microsoft)</li> |
| 694 | <li><code>__has_nothrow_constructor</code> (GNU, Microsoft)</li> |
| 695 | <li><code>__has_trivial_assign</code> (GNU, Microsoft)</li> |
| 696 | <li><code>__has_trivial_copy</code> (GNU, Microsoft)</li> |
| 697 | <li><code>__has_trivial_constructor</code> (GNU, Microsoft)</li> |
| 698 | <li><code>__has_trivial_destructor</code> (GNU, Microsoft)</li> |
| 699 | <li><code>__has_virtual_destructor</code> (GNU, Microsoft)</li> |
| 700 | <li><code>__is_abstract</code> (GNU, Microsoft)</li> |
| 701 | <li><code>__is_base_of</code> (GNU, Microsoft)</li> |
| 702 | <li><code>__is_class</code> (GNU, Microsoft)</li> |
| 703 | <li><code>__is_convertible_to</code> (Microsoft)</li> |
| 704 | <li><code>__is_empty</code> (GNU, Microsoft)</li> |
| 705 | <li><code>__is_enum</code> (GNU, Microsoft)</li> |
| 706 | <li><code>__is_pod</code> (GNU, Microsoft)</li> |
| 707 | <li><code>__is_polymorphic</code> (GNU, Microsoft)</li> |
| 708 | <li><code>__is_union</code> (GNU, Microsoft)</li> |
| 709 | <li><code>__is_literal(type)</code>: Determines whether the given type is a literal type</li> |
Sean Hunt | 1fba828 | 2011-07-18 17:22:33 +0000 | [diff] [blame] | 710 | <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++0x standard library.</li> |
Douglas Gregor | afdf137 | 2011-02-03 21:57:35 +0000 | [diff] [blame] | 711 | </ul> |
| 712 | |
| 713 | <!-- ======================================================================= --> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 714 | <h2 id="blocks">Blocks</h2> |
| 715 | <!-- ======================================================================= --> |
| 716 | |
Chris Lattner | a7dbdf5 | 2009-03-09 07:03:22 +0000 | [diff] [blame] | 717 | <p>The syntax and high level language feature description is in <a |
| 718 | href="BlockLanguageSpec.txt">BlockLanguageSpec.txt</a>. Implementation and ABI |
| 719 | details for the clang implementation are in <a |
Chris Lattner | 5d7650b | 2010-03-16 21:43:03 +0000 | [diff] [blame] | 720 | href="Block-ABI-Apple.txt">Block-ABI-Apple.txt</a>.</p> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 721 | |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 722 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 723 | <p>Query for this feature with __has_extension(blocks).</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 724 | |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 725 | <!-- ======================================================================= --> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 726 | <h2 id="objc_features">Objective-C Features</h2> |
| 727 | <!-- ======================================================================= --> |
| 728 | |
| 729 | <h3 id="objc_instancetype">Related result types</h3> |
| 730 | |
| 731 | <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> |
| 732 | |
| 733 | <blockquote> |
| 734 | <pre> |
| 735 | @interface NSObject |
| 736 | + (id)alloc; |
| 737 | - (id)init; |
| 738 | @end |
| 739 | |
| 740 | @interface NSArray : NSObject |
| 741 | @end |
| 742 | </pre> |
| 743 | </blockquote> |
| 744 | |
| 745 | <p>and this common initialization pattern</p> |
| 746 | |
| 747 | <blockquote> |
| 748 | <pre> |
| 749 | NSArray *array = [[NSArray alloc] init]; |
| 750 | </pre> |
| 751 | </blockquote> |
| 752 | |
| 753 | <p>the type of the expression <code>[NSArray alloc]</code> is |
| 754 | <code>NSArray*</code> because <code>alloc</code> implicitly has a |
| 755 | related result type. Similarly, the type of the expression |
| 756 | <code>[[NSArray alloc] init]</code> is <code>NSArray*</code>, since |
| 757 | <code>init</code> has a related result type and its receiver is known |
| 758 | to 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> |
| 759 | |
Douglas Gregor | e97179c | 2011-09-08 01:46:34 +0000 | [diff] [blame] | 760 | <p>A method with a related result type can be declared by using the |
| 761 | type <tt>instancetype</tt> as its result type. <tt>instancetype</tt> |
| 762 | is a contextual keyword that is only permitted in the result type of |
| 763 | an Objective-C method, e.g.</p> |
| 764 | |
| 765 | <pre> |
| 766 | @interface A |
| 767 | + (<b>instancetype</b>)constructAnA; |
| 768 | @end |
| 769 | </pre> |
| 770 | |
| 771 | <p>The related result type can also be inferred for some methods. |
| 772 | To determine whether a method has an inferred related result type, the first |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 773 | word in the camel-case selector (e.g., "init" in "initWithObjects") is |
| 774 | considered, and the method will a related result type if its return |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 775 | type is compatible with the type of its class and if</p> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 776 | |
| 777 | <ul> |
| 778 | |
| 779 | <li>the first word is "alloc" or "new", and the method is a class |
| 780 | method, or</li> |
| 781 | |
| 782 | <li>the first word is "autorelease", "init", "retain", or "self", |
| 783 | and the method is an instance method.</li> |
| 784 | |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 785 | </ul> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 786 | |
| 787 | <p>If a method with a related result type is overridden by a subclass |
| 788 | method, the subclass method must also return a type that is compatible |
| 789 | with the subclass type. For example:</p> |
| 790 | |
| 791 | <blockquote> |
| 792 | <pre> |
| 793 | @interface NSString : NSObject |
| 794 | - (NSUnrelated *)init; // incorrect usage: NSUnrelated is not NSString or a superclass of NSString |
| 795 | @end |
| 796 | </pre> |
| 797 | </blockquote> |
| 798 | |
| 799 | <p>Related result types only affect the type of a message send or |
| 800 | property access via the given method. In all other respects, a method |
Douglas Gregor | e97179c | 2011-09-08 01:46:34 +0000 | [diff] [blame] | 801 | with a related result type is treated the same way as method that |
| 802 | returns <tt>id</tt>.</p> |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 803 | |
Douglas Gregor | aebb653 | 2011-09-08 17:19:31 +0000 | [diff] [blame] | 804 | <p>Use <tt>__has_feature(objc_instancetype)</tt> to determine whether |
| 805 | the <tt>instancetype</tt> contextual keyword is available.</p> |
| 806 | |
Douglas Gregor | 926df6c | 2011-06-11 01:09:30 +0000 | [diff] [blame] | 807 | <!-- ======================================================================= --> |
John McCall | f85e193 | 2011-06-15 23:02:42 +0000 | [diff] [blame] | 808 | <h2 id="objc_arc">Automatic reference counting </h2> |
| 809 | <!-- ======================================================================= --> |
| 810 | |
| 811 | <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> |
| 812 | |
| 813 | <!-- ======================================================================= --> |
Douglas Gregor | 5471bc8 | 2011-09-08 17:18:35 +0000 | [diff] [blame] | 814 | <h2 id="objc_fixed_enum">Enumerations with a fixed underlying type</h2> |
| 815 | <!-- ======================================================================= --> |
| 816 | |
| 817 | <p>Clang provides support for C++0x enumerations with a fixed |
| 818 | underlying type within Objective-C. For example, one can write an |
| 819 | enumeration type as:</p> |
| 820 | |
| 821 | <pre> |
| 822 | typedef enum : unsigned char { Red, Green, Blue } Color; |
| 823 | </pre> |
| 824 | |
| 825 | <p>This specifies that the underlying type, which is used to store the |
| 826 | enumeration value, is <tt>unsigned char</tt>.</p> |
| 827 | |
| 828 | <p>Use <tt>__has_feature(objc_fixed_enum)</tt> to determine whether |
| 829 | support for fixed underlying types is available in Objective-C.</p> |
| 830 | |
| 831 | <!-- ======================================================================= --> |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 832 | <h2 id="overloading-in-c">Function Overloading in C</h2> |
| 833 | <!-- ======================================================================= --> |
| 834 | |
Chris Lattner | f161d41 | 2009-02-13 21:51:45 +0000 | [diff] [blame] | 835 | <p>Clang provides support for C++ function overloading in C. Function |
| 836 | overloading in C is introduced using the <tt>overloadable</tt> attribute. For |
| 837 | example, one might provide several overloaded versions of a <tt>tgsin</tt> |
| 838 | function that invokes the appropriate standard function computing the sine of a |
| 839 | value with <tt>float</tt>, <tt>double</tt>, or <tt>long double</tt> |
| 840 | precision:</p> |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 841 | |
| 842 | <blockquote> |
| 843 | <pre> |
| 844 | #include <math.h> |
| 845 | float <b>__attribute__((overloadable))</b> tgsin(float x) { return sinf(x); } |
| 846 | double <b>__attribute__((overloadable))</b> tgsin(double x) { return sin(x); } |
| 847 | long double <b>__attribute__((overloadable))</b> tgsin(long double x) { return sinl(x); } |
| 848 | </pre> |
| 849 | </blockquote> |
| 850 | |
| 851 | <p>Given these declarations, one can call <tt>tgsin</tt> with a |
| 852 | <tt>float</tt> value to receive a <tt>float</tt> result, with a |
| 853 | <tt>double</tt> to receive a <tt>double</tt> result, etc. Function |
| 854 | overloading in C follows the rules of C++ function overloading to pick |
| 855 | the best overload given the call arguments, with a few C-specific |
| 856 | semantics:</p> |
| 857 | <ul> |
| 858 | <li>Conversion from <tt>float</tt> or <tt>double</tt> to <tt>long |
| 859 | double</tt> is ranked as a floating-point promotion (per C99) rather |
| 860 | than as a floating-point conversion (as in C++).</li> |
| 861 | |
| 862 | <li>A conversion from a pointer of type <tt>T*</tt> to a pointer of type |
| 863 | <tt>U*</tt> is considered a pointer conversion (with conversion |
| 864 | rank) if <tt>T</tt> and <tt>U</tt> are compatible types.</li> |
| 865 | |
| 866 | <li>A conversion from type <tt>T</tt> to a value of type <tt>U</tt> |
| 867 | is permitted if <tt>T</tt> and <tt>U</tt> are compatible types. This |
| 868 | conversion is given "conversion" rank.</li> |
| 869 | </ul> |
| 870 | |
| 871 | <p>The declaration of <tt>overloadable</tt> functions is restricted to |
| 872 | function declarations and definitions. Most importantly, if any |
| 873 | function with a given name is given the <tt>overloadable</tt> |
| 874 | attribute, then all function declarations and definitions with that |
| 875 | name (and in that scope) must have the <tt>overloadable</tt> |
Chris Lattner | f161d41 | 2009-02-13 21:51:45 +0000 | [diff] [blame] | 876 | attribute. This rule even applies to redeclarations of functions whose original |
| 877 | declaration had the <tt>overloadable</tt> attribute, e.g.,</p> |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 878 | |
| 879 | <blockquote> |
| 880 | <pre> |
| 881 | int f(int) __attribute__((overloadable)); |
| 882 | float f(float); <i>// error: declaration of "f" must have the "overloadable" attribute</i> |
| 883 | |
| 884 | int g(int) __attribute__((overloadable)); |
| 885 | int g(int) { } <i>// error: redeclaration of "g" must also have the "overloadable" attribute</i> |
| 886 | </pre> |
| 887 | </blockquote> |
| 888 | |
Douglas Gregor | 965acbb | 2009-02-18 07:07:28 +0000 | [diff] [blame] | 889 | <p>Functions marked <tt>overloadable</tt> must have |
| 890 | prototypes. Therefore, the following code is ill-formed:</p> |
| 891 | |
| 892 | <blockquote> |
| 893 | <pre> |
| 894 | int h() __attribute__((overloadable)); <i>// error: h does not have a prototype</i> |
| 895 | </pre> |
| 896 | </blockquote> |
| 897 | |
| 898 | <p>However, <tt>overloadable</tt> functions are allowed to use a |
| 899 | ellipsis 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> |
| 900 | |
| 901 | <blockquote> |
| 902 | <pre> |
Chris Lattner | 0224680 | 2009-02-18 22:27:46 +0000 | [diff] [blame] | 903 | void honeypot(...) __attribute__((overloadable, unavailable)); <i>// calling me is an error</i> |
Douglas Gregor | 965acbb | 2009-02-18 07:07:28 +0000 | [diff] [blame] | 904 | </pre> |
| 905 | </blockquote> |
| 906 | |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 907 | <p>Functions declared with the <tt>overloadable</tt> attribute have |
| 908 | their names mangled according to the same rules as C++ function |
| 909 | names. For example, the three <tt>tgsin</tt> functions in our |
| 910 | motivating example get the mangled names <tt>_Z5tgsinf</tt>, |
Chris Lattner | 71b48d6 | 2010-11-28 18:19:13 +0000 | [diff] [blame] | 911 | <tt>_Z5tgsind</tt>, and <tt>_Z5tgsine</tt>, respectively. There are two |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 912 | caveats to this use of name mangling:</p> |
| 913 | |
| 914 | <ul> |
| 915 | |
| 916 | <li>Future versions of Clang may change the name mangling of |
| 917 | functions overloaded in C, so you should not depend on an specific |
| 918 | mangling. To be completely safe, we strongly urge the use of |
| 919 | <tt>static inline</tt> with <tt>overloadable</tt> functions.</li> |
| 920 | |
| 921 | <li>The <tt>overloadable</tt> attribute has almost no meaning when |
| 922 | used in C++, because names will already be mangled and functions are |
| 923 | already overloadable. However, when an <tt>overloadable</tt> |
| 924 | function occurs within an <tt>extern "C"</tt> linkage specification, |
| 925 | it's name <i>will</i> be mangled in the same way as it would in |
| 926 | C.</li> |
| 927 | </ul> |
| 928 | |
Peter Collingbourne | c1b5fa4 | 2011-05-13 20:54:45 +0000 | [diff] [blame] | 929 | <p>Query for this feature with __has_extension(attribute_overloadable).</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 930 | |
Eli Friedman | 0c706c2 | 2011-09-19 23:17:44 +0000 | [diff] [blame] | 931 | <!-- ======================================================================= --> |
| 932 | <h2 id="complex-list-init">Initializer lists for complex numbers in C</h2> |
| 933 | <!-- ======================================================================= --> |
| 934 | |
| 935 | <p>clang supports an extension which allows the following in C:</p> |
| 936 | |
| 937 | <blockquote> |
| 938 | <pre> |
| 939 | #include <math.h> |
| 940 | #include <complex.h> |
| 941 | complex float x = { 1.0f, INFINITY }; // Init to (1, Inf) |
| 942 | </pre> |
| 943 | </blockquote> |
| 944 | |
| 945 | <p>This construct is useful because there is no way to separately |
| 946 | initialize the real and imaginary parts of a complex variable in |
| 947 | standard C, given that clang does not support <code>_Imaginary</code>. |
| 948 | (clang also supports the <code>__real__</code> and <code>__imag__</code> |
| 949 | extensions from gcc, which help in some cases, but are not usable in |
| 950 | static initializers.) |
| 951 | |
| 952 | <p>Note that this extension does not allow eliding the braces; the |
| 953 | meaning of the following two lines is different:</p> |
| 954 | |
| 955 | <blockquote> |
| 956 | <pre> |
| 957 | complex float x[] = { { 1.0f, 1.0f } }; // [0] = (1, 1) |
| 958 | complex float x[] = { 1.0f, 1.0f }; // [0] = (1, 0), [1] = (1, 0) |
| 959 | </pre> |
| 960 | </blockquote> |
| 961 | |
| 962 | <p>This extension also works in C++ mode, as far as that goes, but does not |
| 963 | apply to the C++ <code>std::complex</code>. (In C++11, list |
| 964 | initialization allows the same syntax to be used with |
| 965 | <code>std::complex</code> with the same meaning.) |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 966 | |
Douglas Gregor | cb54d43 | 2009-02-13 00:57:04 +0000 | [diff] [blame] | 967 | <!-- ======================================================================= --> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 968 | <h2 id="builtins">Builtin Functions</h2> |
| 969 | <!-- ======================================================================= --> |
| 970 | |
| 971 | <p>Clang supports a number of builtin library functions with the same syntax as |
| 972 | GCC, including things like <tt>__builtin_nan</tt>, |
| 973 | <tt>__builtin_constant_p</tt>, <tt>__builtin_choose_expr</tt>, |
| 974 | <tt>__builtin_types_compatible_p</tt>, <tt>__sync_fetch_and_add</tt>, etc. In |
| 975 | addition to the GCC builtins, Clang supports a number of builtins that GCC does |
| 976 | not, which are listed here.</p> |
| 977 | |
| 978 | <p>Please note that Clang does not and will not support all of the GCC builtins |
| 979 | for vector operations. Instead of using builtins, you should use the functions |
| 980 | defined in target-specific header files like <tt><xmmintrin.h></tt>, which |
| 981 | define portable wrappers for these. Many of the Clang versions of these |
| 982 | functions are implemented directly in terms of <a href="#vectors">extended |
| 983 | vector support</a> instead of builtins, in order to reduce the number of |
| 984 | builtins that we need to implement.</p> |
| 985 | |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 986 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 987 | <h3><a name="__builtin_shufflevector">__builtin_shufflevector</a></h3> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 988 | <!-- ======================================================================= --> |
| 989 | |
Chris Lattner | aad826b | 2009-09-16 18:56:12 +0000 | [diff] [blame] | 990 | <p><tt>__builtin_shufflevector</tt> is used to express generic vector |
Chris Lattner | 6f72da5 | 2009-02-13 20:00:20 +0000 | [diff] [blame] | 991 | permutation/shuffle/swizzle operations. This builtin is also very important for |
| 992 | the implementation of various target-specific header files like |
| 993 | <tt><xmmintrin.h></tt>. |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 994 | </p> |
| 995 | |
| 996 | <p><b>Syntax:</b></p> |
| 997 | |
| 998 | <pre> |
Chris Lattner | 6f72da5 | 2009-02-13 20:00:20 +0000 | [diff] [blame] | 999 | __builtin_shufflevector(vec1, vec2, index1, index2, ...) |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 1000 | </pre> |
| 1001 | |
| 1002 | <p><b>Examples:</b></p> |
| 1003 | |
| 1004 | <pre> |
Chris Lattner | 6f72da5 | 2009-02-13 20:00:20 +0000 | [diff] [blame] | 1005 | // Identity operation - return 4-element vector V1. |
| 1006 | __builtin_shufflevector(V1, V1, 0, 1, 2, 3) |
| 1007 | |
| 1008 | // "Splat" element 0 of V1 into a 4-element result. |
| 1009 | __builtin_shufflevector(V1, V1, 0, 0, 0, 0) |
| 1010 | |
| 1011 | // Reverse 4-element vector V1. |
| 1012 | __builtin_shufflevector(V1, V1, 3, 2, 1, 0) |
| 1013 | |
| 1014 | // Concatenate every other element of 4-element vectors V1 and V2. |
| 1015 | __builtin_shufflevector(V1, V2, 0, 2, 4, 6) |
| 1016 | |
| 1017 | // Concatenate every other element of 8-element vectors V1 and V2. |
| 1018 | __builtin_shufflevector(V1, V2, 0, 2, 4, 6, 8, 10, 12, 14) |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 1019 | </pre> |
| 1020 | |
| 1021 | <p><b>Description:</b></p> |
| 1022 | |
Chris Lattner | 6f72da5 | 2009-02-13 20:00:20 +0000 | [diff] [blame] | 1023 | <p>The first two arguments to __builtin_shufflevector are vectors that have the |
| 1024 | same element type. The remaining arguments are a list of integers that specify |
| 1025 | the elements indices of the first two vectors that should be extracted and |
| 1026 | returned in a new vector. These element indices are numbered sequentially |
| 1027 | starting with the first vector, continuing into the second vector. Thus, if |
| 1028 | vec1 is a 4-element vector, index 5 would refer to the second element of vec2. |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 1029 | </p> |
| 1030 | |
Chris Lattner | 6f72da5 | 2009-02-13 20:00:20 +0000 | [diff] [blame] | 1031 | <p>The result of __builtin_shufflevector is a vector |
| 1032 | with the same element type as vec1/vec2 but that has an element count equal to |
| 1033 | the number of indices specified. |
| 1034 | </p> |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 1035 | |
Chris Lattner | 21190d5 | 2009-09-21 03:09:59 +0000 | [diff] [blame] | 1036 | <p>Query for this feature with __has_builtin(__builtin_shufflevector).</p> |
| 1037 | |
| 1038 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 1039 | <h3><a name="__builtin_unreachable">__builtin_unreachable</a></h3> |
Chris Lattner | 21190d5 | 2009-09-21 03:09:59 +0000 | [diff] [blame] | 1040 | <!-- ======================================================================= --> |
| 1041 | |
| 1042 | <p><tt>__builtin_unreachable</tt> is used to indicate that a specific point in |
| 1043 | the program cannot be reached, even if the compiler might otherwise think it |
| 1044 | can. This is useful to improve optimization and eliminates certain warnings. |
| 1045 | For example, without the <tt>__builtin_unreachable</tt> in the example below, |
| 1046 | the compiler assumes that the inline asm can fall through and prints a "function |
| 1047 | declared 'noreturn' should not return" warning. |
| 1048 | </p> |
| 1049 | |
| 1050 | <p><b>Syntax:</b></p> |
| 1051 | |
| 1052 | <pre> |
| 1053 | __builtin_unreachable() |
| 1054 | </pre> |
| 1055 | |
| 1056 | <p><b>Example of Use:</b></p> |
| 1057 | |
| 1058 | <pre> |
| 1059 | void myabort(void) __attribute__((noreturn)); |
| 1060 | void myabort(void) { |
| 1061 | asm("int3"); |
| 1062 | __builtin_unreachable(); |
| 1063 | } |
| 1064 | </pre> |
| 1065 | |
| 1066 | <p><b>Description:</b></p> |
| 1067 | |
| 1068 | <p>The __builtin_unreachable() builtin has completely undefined behavior. Since |
| 1069 | it has undefined behavior, it is a statement that it is never reached and the |
| 1070 | optimizer can take advantage of this to produce better code. This builtin takes |
| 1071 | no arguments and produces a void result. |
| 1072 | </p> |
| 1073 | |
| 1074 | <p>Query for this feature with __has_builtin(__builtin_unreachable).</p> |
| 1075 | |
Chris Lattner | 23aa9c8 | 2011-04-09 03:57:26 +0000 | [diff] [blame] | 1076 | <!-- ======================================================================= --> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 1077 | <h3><a name="__sync_swap">__sync_swap</a></h3> |
Chris Lattner | 23aa9c8 | 2011-04-09 03:57:26 +0000 | [diff] [blame] | 1078 | <!-- ======================================================================= --> |
| 1079 | |
| 1080 | <p><tt>__sync_swap</tt> is used to atomically swap integers or pointers in |
| 1081 | memory. |
| 1082 | </p> |
| 1083 | |
| 1084 | <p><b>Syntax:</b></p> |
| 1085 | |
| 1086 | <pre> |
| 1087 | <i>type</i> __sync_swap(<i>type</i> *ptr, <i>type</i> value, ...) |
| 1088 | </pre> |
| 1089 | |
| 1090 | <p><b>Example of Use:</b></p> |
| 1091 | |
| 1092 | <pre> |
Sean Hunt | 7e98b47 | 2011-06-23 01:21:01 +0000 | [diff] [blame] | 1093 | int old_value = __sync_swap(&value, new_value); |
Chris Lattner | 23aa9c8 | 2011-04-09 03:57:26 +0000 | [diff] [blame] | 1094 | </pre> |
| 1095 | |
| 1096 | <p><b>Description:</b></p> |
| 1097 | |
| 1098 | <p>The __sync_swap() builtin extends the existing __sync_*() family of atomic |
| 1099 | intrinsics to allow code to atomically swap the current value with the new |
| 1100 | value. More importantly, it helps developers write more efficient and correct |
| 1101 | code by avoiding expensive loops around __sync_bool_compare_and_swap() or |
| 1102 | relying on the platform specific implementation details of |
| 1103 | __sync_lock_test_and_set(). The __sync_swap() builtin is a full barrier. |
| 1104 | </p> |
| 1105 | |
Chris Lattner | 21190d5 | 2009-09-21 03:09:59 +0000 | [diff] [blame] | 1106 | |
Chris Lattner | 1177f91 | 2009-04-09 19:58:15 +0000 | [diff] [blame] | 1107 | <!-- ======================================================================= --> |
| 1108 | <h2 id="targetspecific">Target-Specific Extensions</h2> |
| 1109 | <!-- ======================================================================= --> |
| 1110 | |
| 1111 | <p>Clang supports some language features conditionally on some targets.</p> |
| 1112 | |
| 1113 | <!-- ======================================================================= --> |
| 1114 | <h3 id="x86-specific">X86/X86-64 Language Extensions</h3> |
| 1115 | <!-- ======================================================================= --> |
| 1116 | |
| 1117 | <p>The X86 backend has these language extensions:</p> |
| 1118 | |
| 1119 | <!-- ======================================================================= --> |
| 1120 | <h4 id="x86-gs-segment">Memory references off the GS segment</h4> |
| 1121 | <!-- ======================================================================= --> |
| 1122 | |
| 1123 | <p>Annotating a pointer with address space #256 causes it to be code generated |
Chris Lattner | a021e7c | 2009-05-05 18:54:47 +0000 | [diff] [blame] | 1124 | relative to the X86 GS segment register, and address space #257 causes it to be |
| 1125 | relative to the X86 FS segment. Note that this is a very very low-level |
| 1126 | feature that should only be used if you know what you're doing (for example in |
| 1127 | an OS kernel).</p> |
Chris Lattner | 1177f91 | 2009-04-09 19:58:15 +0000 | [diff] [blame] | 1128 | |
| 1129 | <p>Here is an example:</p> |
| 1130 | |
| 1131 | <pre> |
| 1132 | #define GS_RELATIVE __attribute__((address_space(256))) |
| 1133 | int foo(int GS_RELATIVE *P) { |
| 1134 | return *P; |
| 1135 | } |
| 1136 | </pre> |
| 1137 | |
| 1138 | <p>Which compiles to (on X86-32):</p> |
| 1139 | |
| 1140 | <pre> |
| 1141 | _foo: |
| 1142 | movl 4(%esp), %eax |
| 1143 | movl %gs:(%eax), %eax |
| 1144 | ret |
| 1145 | </pre> |
| 1146 | |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1147 | <!-- ======================================================================= --> |
| 1148 | <h2 id="analyzerspecific">Static Analysis-Specific Extensions</h2> |
| 1149 | <!-- ======================================================================= --> |
| 1150 | |
| 1151 | <p>Clang supports additional attributes that are useful for documenting program |
| 1152 | invariants and rules for static analysis tools. The extensions documented here |
| 1153 | are used by the <a |
| 1154 | href="http://clang.llvm.org/StaticAnalysis.html">path-sensitive static analyzer |
| 1155 | engine</a> that is part of Clang's Analysis library.</p> |
| 1156 | |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 1157 | <h3 id="attr_analyzer_noreturn">The <tt>analyzer_noreturn</tt> attribute</h3> |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1158 | |
| 1159 | <p>Clang's static analysis engine understands the standard <tt>noreturn</tt> |
Ted Kremenek | 4df2114 | 2009-04-10 05:04:22 +0000 | [diff] [blame] | 1160 | attribute. This attribute, which is typically affixed to a function prototype, |
| 1161 | indicates that a call to a given function never returns. Function prototypes for |
| 1162 | common functions like <tt>exit</tt> are typically annotated with this attribute, |
| 1163 | as well as a variety of common assertion handlers. Users can educate the static |
| 1164 | analyzer about their own custom assertion handles (thus cutting down on false |
| 1165 | positives due to false paths) by marking their own "panic" functions |
| 1166 | with this attribute.</p> |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1167 | |
| 1168 | <p>While useful, <tt>noreturn</tt> is not applicable in all cases. Sometimes |
Nick Lewycky | 625b586 | 2009-06-14 04:08:08 +0000 | [diff] [blame] | 1169 | there are special functions that for all intents and purposes should be |
| 1170 | considered panic functions (i.e., they are only called when an internal program |
| 1171 | error occurs) but may actually return so that the program can fail gracefully. |
| 1172 | The <tt>analyzer_noreturn</tt> attribute allows one to annotate such functions |
| 1173 | as being interpreted as "no return" functions by the analyzer (thus |
Chris Lattner | 2893589 | 2009-04-10 05:54:56 +0000 | [diff] [blame] | 1174 | pruning bogus paths) but will not affect compilation (as in the case of |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1175 | <tt>noreturn</tt>).</p> |
| 1176 | |
| 1177 | <p><b>Usage</b>: The <tt>analyzer_noreturn</tt> attribute can be placed in the |
Chris Lattner | 2893589 | 2009-04-10 05:54:56 +0000 | [diff] [blame] | 1178 | same places where the <tt>noreturn</tt> attribute can be placed. It is commonly |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1179 | placed at the end of function prototypes:</p> |
| 1180 | |
| 1181 | <pre> |
| 1182 | void foo() <b>__attribute__((analyzer_noreturn))</b>; |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 1183 | </pre> |
| 1184 | |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 1185 | <p>Query for this feature with |
| 1186 | <tt>__has_attribute(analyzer_noreturn)</tt>.</p> |
Chris Lattner | 148772a | 2009-06-13 07:13:28 +0000 | [diff] [blame] | 1187 | |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 1188 | <h3 id="attr_method_family">The <tt>objc_method_family</tt> attribute</h3> |
| 1189 | |
| 1190 | <p>Many methods in Objective-C have conventional meanings determined |
| 1191 | by their selectors. For the purposes of static analysis, it is |
| 1192 | sometimes useful to be able to mark a method as having a particular |
| 1193 | conventional meaning despite not having the right selector, or as not |
| 1194 | having the conventional meaning that its selector would suggest. |
| 1195 | For these use cases, we provide an attribute to specifically describe |
| 1196 | the <q>method family</q> that a method belongs to.</p> |
| 1197 | |
| 1198 | <p><b>Usage</b>: <tt>__attribute__((objc_method_family(X)))</tt>, |
| 1199 | where <tt>X</tt> is one of <tt>none</tt>, <tt>alloc</tt>, <tt>copy</tt>, |
| 1200 | <tt>init</tt>, <tt>mutableCopy</tt>, or <tt>new</tt>. This attribute |
| 1201 | can only be placed at the end of a method declaration:</p> |
| 1202 | |
| 1203 | <pre> |
| 1204 | - (NSString*) initMyStringValue <b>__attribute__((objc_method_family(none)))</b>; |
| 1205 | </pre> |
| 1206 | |
| 1207 | <p>Users who do not wish to change the conventional meaning of a |
| 1208 | method, and who merely want to document its non-standard retain and |
| 1209 | release semantics, should use the |
| 1210 | <a href="#attr_retain_release">retaining behavior attributes</a> |
| 1211 | described below.</p> |
| 1212 | |
| 1213 | <p>Query for this feature with |
| 1214 | <tt>__has_attribute(objc_method_family)</tt>.</p> |
| 1215 | |
| 1216 | <h3 id="attr_retain_release">Objective-C retaining behavior attributes</h3> |
John McCall | 630b7ae | 2011-01-25 04:26:21 +0000 | [diff] [blame] | 1217 | |
| 1218 | <p>In Objective-C, functions and methods are generally assumed to take |
| 1219 | and return objects with +0 retain counts, with some exceptions for |
| 1220 | special methods like <tt>+alloc</tt> and <tt>init</tt>. However, |
| 1221 | there are exceptions, and so Clang provides attributes to allow these |
| 1222 | exceptions to be documented, which helps the analyzer find leaks (and |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 1223 | ignore non-leaks). Some exceptions may be better described using |
| 1224 | the <a href="#attr_method_family"><tt>objc_method_family</tt></a> |
| 1225 | attribute instead.</p> |
John McCall | 630b7ae | 2011-01-25 04:26:21 +0000 | [diff] [blame] | 1226 | |
| 1227 | <p><b>Usage</b>: The <tt>ns_returns_retained</tt>, <tt>ns_returns_not_retained</tt>, |
| 1228 | <tt>ns_returns_autoreleased</tt>, <tt>cf_returns_retained</tt>, |
| 1229 | and <tt>cf_returns_not_retained</tt> attributes can be placed on |
| 1230 | methods and functions that return Objective-C or CoreFoundation |
| 1231 | objects. They are commonly placed at the end of a function prototype |
| 1232 | or method declaration:</p> |
| 1233 | |
| 1234 | <pre> |
| 1235 | id foo() <b>__attribute__((ns_returns_retained))</b>; |
| 1236 | |
| 1237 | - (NSString*) bar: (int) x <b>__attribute__((ns_returns_retained))</b>; |
| 1238 | </pre> |
| 1239 | |
| 1240 | <p>The <tt>*_returns_retained</tt> attributes specify that the |
| 1241 | returned object has a +1 retain count. |
| 1242 | The <tt>*_returns_not_retained</tt> attributes specify that the return |
| 1243 | object has a +0 retain count, even if the normal convention for its |
| 1244 | selector would be +1. <tt>ns_returns_autoreleased</tt> specifies that the |
| 1245 | returned object is +0, but is guaranteed to live at least as long as the |
| 1246 | next flush of an autorelease pool.</p> |
| 1247 | |
| 1248 | <p><b>Usage</b>: The <tt>ns_consumed</tt> and <tt>cf_consumed</tt> |
| 1249 | attributes can be placed on an parameter declaration; they specify |
| 1250 | that the argument is expected to have a +1 retain count, which will be |
| 1251 | balanced in some way by the function or method. |
| 1252 | The <tt>ns_consumes_self</tt> attribute can only be placed on an |
| 1253 | Objective-C method; it specifies that the method expects |
| 1254 | its <tt>self</tt> parameter to have a +1 retain count, which it will |
| 1255 | balance in some way.</p> |
| 1256 | |
| 1257 | <pre> |
| 1258 | void <b>foo(__attribute__((ns_consumed))</b> NSString *string); |
| 1259 | |
| 1260 | - (void) bar <b>__attribute__((ns_consumes_self))</b>; |
| 1261 | - (void) baz: (id) <b>__attribute__((ns_consumed))</b> x; |
| 1262 | </pre> |
Ted Kremenek | ed86931 | 2009-04-10 05:03:33 +0000 | [diff] [blame] | 1263 | |
John McCall | 8749401 | 2011-03-18 03:51:49 +0000 | [diff] [blame] | 1264 | <p>Query for these features with <tt>__has_attribute(ns_consumed)</tt>, |
| 1265 | <tt>__has_attribute(ns_returns_retained)</tt>, etc.</p> |
| 1266 | |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1267 | |
| 1268 | <!-- ======================================================================= --> |
Caitlin Sadowski | 73cbbc8 | 2011-07-28 18:38:36 +0000 | [diff] [blame] | 1269 | <h2 id="threadsafety">Thread-Safety Annotation Checking</h2> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1270 | <!-- ======================================================================= --> |
| 1271 | |
| 1272 | <p>Clang supports additional attributes for checking basic locking policies in |
| 1273 | multithreaded programs. |
| 1274 | Clang currently parses the following list of attributes, although |
| 1275 | <b>the implementation for these annotations is currently in development.</b> |
| 1276 | For more details, see the |
| 1277 | <a href="http://gcc.gnu.org/wiki/ThreadSafetyAnnotation">GCC implementation</a>. |
| 1278 | </p> |
| 1279 | |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1280 | <h4 id="ts_noanal">no_thread_safety_analysis</h4> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1281 | |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1282 | <p>Use <tt>__attribute__((no_thread_safety_analysis))</tt> on a function |
| 1283 | declaration to specify that the thread safety analysis should not be run on that |
| 1284 | function. This attribute provides an escape hatch (e.g. for situations when it |
| 1285 | is difficult to annotate the locking policy). </p> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1286 | |
| 1287 | <h4 id="ts_lockable">lockable</h4> |
| 1288 | |
| 1289 | <p>Use <tt>__attribute__((lockable))</tt> on a class definition to specify |
| 1290 | that it has a lockable type (e.g. a Mutex class). This annotation is primarily |
| 1291 | used to check consistency.</p> |
| 1292 | |
| 1293 | <h4 id="ts_scopedlockable">scoped_lockable</h4> |
| 1294 | |
| 1295 | <p>Use <tt>__attribute__((scoped_lockable))</tt> on a class definition to |
| 1296 | specify that it has a "scoped" lockable type. Objects of this type will acquire |
| 1297 | the lock upon construction and release it upon going out of scope. |
| 1298 | This annotation is primarily used to check |
| 1299 | consistency.</p> |
| 1300 | |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1301 | <h4 id="ts_guardedvar">guarded_var</h4> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1302 | |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1303 | <p>Use <tt>__attribute__((guarded_var))</tt> on a variable declaration to |
| 1304 | specify that the variable must be accessed while holding some lock.</p> |
| 1305 | |
| 1306 | <h4 id="ts_ptguardedvar">pt_guarded_var</h4> |
| 1307 | |
| 1308 | <p>Use <tt>__attribute__((pt_guarded_var))</tt> on a pointer declaration to |
| 1309 | specify that the pointer must be dereferenced while holding some lock.</p> |
| 1310 | |
| 1311 | <h4 id="ts_guardedby">guarded_by(l)</h4> |
| 1312 | |
| 1313 | <p>Use <tt>__attribute__((guarded_by(l)))</tt> on a variable declaration to |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1314 | specify that the variable must be accessed while holding lock <tt>l</tt>.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1315 | |
| 1316 | <h4 id="ts_ptguardedby">pt_guarded_by(l)</h4> |
| 1317 | |
| 1318 | <p>Use <tt>__attribute__((pt_guarded_by(l)))</tt> on a pointer declaration to |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1319 | specify that the pointer must be dereferenced while holding lock <tt>l</tt>.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1320 | |
| 1321 | <h4 id="ts_acquiredbefore">acquired_before(...)</h4> |
| 1322 | |
| 1323 | <p>Use <tt>__attribute__((acquired_before(...)))</tt> on a declaration |
| 1324 | of a lockable variable to specify that the lock must be acquired before all |
| 1325 | attribute arguments. Arguments must be lockable type, and there must be at |
| 1326 | least one argument.</p> |
| 1327 | |
| 1328 | <h4 id="ts_acquiredafter">acquired_after(...)</h4> |
| 1329 | |
| 1330 | <p>Use <tt>__attribute__((acquired_after(...)))</tt> on a declaration |
| 1331 | of a lockable variable to specify that the lock must be acquired after all |
| 1332 | attribute arguments. Arguments must be lockable type, and there must be at |
| 1333 | least one argument.</p> |
| 1334 | |
| 1335 | <h4 id="ts_elf">exclusive_lock_function(...)</h4> |
| 1336 | |
| 1337 | <p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function |
| 1338 | declaration to specify that the function acquires all listed locks |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1339 | exclusively. This attribute takes zero or more arguments: either of lockable |
| 1340 | type or integers indexing into function parameters of lockable type. If no |
| 1341 | arguments are given, the acquired lock is implicitly <tt>this</tt> of the |
| 1342 | enclosing object.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1343 | |
| 1344 | <h4 id="ts_slf">shared_lock_function(...)</h4> |
| 1345 | |
| 1346 | <p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function |
| 1347 | declaration to specify that the function acquires all listed locks, although |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1348 | the locks may be shared (e.g. read locks). This attribute takes zero or more |
| 1349 | arguments: either of lockable type or integers indexing into function |
| 1350 | parameters of lockable type. If no arguments are given, the acquired lock is |
| 1351 | implicitly <tt>this</tt> of the enclosing object.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1352 | |
| 1353 | <h4 id="ts_etf">exclusive_trylock_function(...)</h4> |
| 1354 | |
| 1355 | <p>Use <tt>__attribute__((exclusive_lock_function(...)))</tt> on a function |
| 1356 | declaration to specify that the function will try (without blocking) to acquire |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1357 | all listed locks exclusively. This attribute takes one or more arguments. The |
| 1358 | first argument is an integer or boolean value specifying the return value of a |
| 1359 | successful lock acquisition. The remaining arugments are either of lockable type |
| 1360 | or integers indexing into function parameters of lockable type. If only one |
| 1361 | argument is given, the acquired lock is implicitly <tt>this</tt> of the |
| 1362 | enclosing object.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1363 | |
| 1364 | <h4 id="ts_stf">shared_trylock_function(...)</h4> |
| 1365 | |
| 1366 | <p>Use <tt>__attribute__((shared_lock_function(...)))</tt> on a function |
| 1367 | declaration to specify that the function will try (without blocking) to acquire |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1368 | all listed locks, although the locks may be shared (e.g. read locks). This |
| 1369 | attribute takes one or more arguments. The first argument is an integer or |
| 1370 | boolean value specifying the return value of a successful lock acquisition. The |
| 1371 | remaining arugments are either of lockable type or integers indexing into |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1372 | function parameters of lockable type. If only one argument is given, the |
| 1373 | acquired lock is implicitly <tt>this</tt> of the enclosing object.</p> |
| 1374 | |
| 1375 | <h4 id="ts_uf">unlock_function(...)</h4> |
| 1376 | |
| 1377 | <p>Use <tt>__attribute__((unlock_function(...)))</tt> on a function |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1378 | declaration to specify that the function release all listed locks. This |
| 1379 | attribute takes zero or more arguments: either of lockable type or integers |
| 1380 | indexing into function parameters of lockable type. If no arguments are given, |
| 1381 | the acquired lock is implicitly <tt>this</tt> of the enclosing object.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1382 | |
| 1383 | <h4 id="ts_lr">lock_returned(l)</h4> |
| 1384 | |
| 1385 | <p>Use <tt>__attribute__((lock_returned(l)))</tt> on a function |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1386 | declaration to specify that the function returns lock <tt>l</tt> (<tt>l</tt> |
| 1387 | must be of lockable type). This annotation is used to aid in resolving lock |
| 1388 | expressions.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1389 | |
| 1390 | <h4 id="ts_le">locks_excluded(...)</h4> |
| 1391 | |
| 1392 | <p>Use <tt>__attribute__((locks_excluded(...)))</tt> on a function declaration |
Caitlin Sadowski | b51e031 | 2011-08-09 17:59:31 +0000 | [diff] [blame] | 1393 | to specify that the function must not be called with the listed locks. Arguments |
| 1394 | must be lockable type, and there must be at least one argument.</p> |
Caitlin Sadowski | db33e14 | 2011-07-28 20:12:35 +0000 | [diff] [blame] | 1395 | |
| 1396 | <h4 id="ts_elr">exclusive_locks_required(...)</h4> |
| 1397 | |
| 1398 | <p>Use <tt>__attribute__((exclusive_locks_required(...)))</tt> on a function |
| 1399 | declaration to specify that the function must be called while holding the listed |
| 1400 | exclusive locks. Arguments must be lockable type, and there must be at |
| 1401 | least one argument.</p> |
| 1402 | |
| 1403 | <h4 id="ts_slr">shared_locks_required(...)</h4> |
| 1404 | |
| 1405 | <p>Use <tt>__attribute__((shared_locks_required(...)))</tt> on a function |
| 1406 | declaration to specify that the function must be called while holding the listed |
| 1407 | shared locks. Arguments must be lockable type, and there must be at |
| 1408 | least one argument.</p> |
Caitlin Sadowski | fdde9e7 | 2011-07-28 17:21:07 +0000 | [diff] [blame] | 1409 | |
Chris Lattner | 5ce933f | 2009-02-09 08:46:11 +0000 | [diff] [blame] | 1410 | </div> |
| 1411 | </body> |
| 1412 | </html> |