Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 815eb6b80739ffca9879b9833208932dee55d0cb / . / docs / UsersManual.html
blob: 9d7981919c2805c0de691694ea54c733ae476639 [file] [log] [blame]
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]1<html>
2<head>
3<title>Clang Compiler User's Manual</title>
4<link type="text/css" rel="stylesheet" href="../menu.css" />
5<link type="text/css" rel="stylesheet" href="../content.css" />
6<style type="text/css">
7td {
8 vertical-align: top;
9}
10</style>
11</head>
12<body>
13
14<!--#include virtual="../menu.html.incl"-->
15
16<div id="content">
17
18<h1>Clang Compiler User's Manual</h1>
19
20<ul>
21<li><a href="#intro">Introduction</a>
22 <ul>
23 <li><a href="#terminology">Terminology</a></li>
24 <li><a href="#basicusage">Basic Usage</a></li>
25 </ul>
26</li>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]27<li><a href="#commandline">Command Line Options</a>
28 <ul>
29 <li><a href="#cl_diagnostics">Options to Control Error and Warning
30 Messages</a></li>
31 </ul>
32</li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]33<li><a href="#general_features">Language and Target-Independent Features</a>
34 <ul>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]35 <li><a href="#diagnostics">Controlling Errors and Warnings</a>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]36 <ul>
37 <li><a href="#diagnostics_display">Controlling How Clang Displays Diagnostics</a></li>
38 <li><a href="#diagnostics_mappings">Diagnostic Mappings</a></li>
Chris Lattner67db8cd2010-05-30 23:42:51 +0000[diff] [blame]39 <li><a href="#diagnostics_categories">Diagnostic Categories</a></li>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]40 <li><a href="#diagnostics_commandline">Controlling Diagnostics via Command Line Flags</a></li>
41 <li><a href="#diagnostics_pragmas">Controlling Diagnostics via Pragmas</a></li>
Ted Kremenekf391fa72011-08-18 01:17:05 +0000[diff] [blame]42 <li><a href="#diagnostics_enable_everything">Enabling All Warnings</a></li>
Ted Kremenek2fb11eb2010-08-24 18:12:35 +0000[diff] [blame]43 <li><a href="#analyzer_diagnositics">Controlling Static Analyzer Diagnostics</a></li>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]44 </ul>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]45 </li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]46 <li><a href="#precompiledheaders">Precompiled Headers</a></li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]47 <li><a href="#codegen">Controlling Code Generation</a></li>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]48 </ul>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]49</li>
50<li><a href="#c">C Language Features</a>
51 <ul>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]52 <li><a href="#c_ext">Extensions supported by clang</a></li>
53 <li><a href="#c_modes">Differences between various standard modes</a></li>
54 <li><a href="#c_unimpl_gcc">GCC extensions not implemented yet</a></li>
55 <li><a href="#c_unsupp_gcc">Intentionally unsupported GCC extensions</a></li>
56 <li><a href="#c_ms">Microsoft extensions</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]57 </ul>
58</li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]59<li><a href="#target_features">Target-Specific Features and Limitations</a>
60 <ul>
61 <li><a href="#target_arch">CPU Architectures Features and Limitations</a>
62 <ul>
63 <li><a href="#target_arch_x86">X86</a></li>
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]64 <li><a href="#target_arch_arm">ARM</a></li>
65 <li><a href="#target_arch_other">Other platforms</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]66 </ul>
67 </li>
68 <li><a href="#target_os">Operating System Features and Limitations</a>
69 <ul>
70 <li><a href="#target_os_darwin">Darwin (Mac OS/X)</a></li>
71 <li>Linux, etc.</li>
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]72 <li><a href="#target_os_win32">Windows</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]73 </ul>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]74 </li>
75 </ul>
76</li>
77</ul>
78
79
80<!-- ======================================================================= -->
81<h2 id="intro">Introduction</h2>
82<!-- ======================================================================= -->
83
84<p>The Clang Compiler is an open-source compiler for the C family of programming
85languages, aiming to be the best in class implementation of these languages.
86Clang builds on the LLVM optimizer and code generator, allowing it to provide
87high-quality optimization and code generation support for many targets. For
88more general information, please see the <a href="http://clang.llvm.org">Clang
89Web Site</a> or the <a href="http://llvm.org">LLVM Web Site</a>.</p>
90
91<p>This document describes important notes about using Clang as a compiler for
92an end-user, documenting the supported features, command line options, etc. If
93you are interested in using Clang to build a tool that processes code, please
94see <a href="InternalsManual.html">the Clang Internals Manual</a>. If you are
95interested in the <a href="http://clang.llvm.org/StaticAnalysis.html">Clang
96Static Analyzer</a>, please see its web page.</p>
97
98<p>Clang is designed to support the C family of programming languages, which
99includes <a href="#c">C</a>, <a href="#objc">Objective-C</a>, <a
100href="#cxx">C++</a>, and <a href="#objcxx">Objective-C++</a> as well as many
101dialects of those. For language-specific information, please see the
102corresponding language specific section:</p>
103
104<ul>
105<li><a href="#c">C Language</a>: K&amp;R C, ANSI C89, ISO C90, ISO C94
106 (C89+AMD1), ISO C99 (+TC1, TC2, TC3). </li>
107<li><a href="#objc">Objective-C Language</a>: ObjC 1, ObjC 2, ObjC 2.1, plus
108 variants depending on base language.</li>
109<li><a href="#cxx">C++ Language Features</a></li>
110<li><a href="#objcxx">Objective C++ Language</a></li>
111</ul>
112
113<p>In addition to these base languages and their dialects, Clang supports a
114broad variety of language extensions, which are documented in the corresponding
115language section. These extensions are provided to be compatible with the GCC,
116Microsoft, and other popular compilers as well as to improve functionality
117through Clang-specific features. The Clang driver and language features are
118intentionally designed to be as compatible with the GNU GCC compiler as
119reasonably possible, easing migration from GCC to Clang. In most cases, code
120"just works".</p>
121
122<p>In addition to language specific features, Clang has a variety of features
123that depend on what CPU architecture or operating system is being compiled for.
Douglas Gregorcd5a5052009-11-09 15:15:41 +0000[diff] [blame]124Please see the <a href="#target_features">Target-Specific Features and
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]125Limitations</a> section for more details.</p>
126
127<p>The rest of the introduction introduces some basic <a
128href="#terminology">compiler terminology</a> that is used throughout this manual
129and contains a basic <a href="#basicusage">introduction to using Clang</a>
130as a command line compiler.</p>
131
132<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
133<h3 id="terminology">Terminology</h3>
134<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
135
136<p>Front end, parser, backend, preprocessor, undefined behavior, diagnostic,
137 optimizer</p>
138
139<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
140<h3 id="basicusage">Basic Usage</h3>
141<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
142
143<p>Intro to how to use a C compiler for newbies.</p>
144<p>
145compile + link
146
147compile then link
148
149debug info
150
151enabling optimizations
152
153picking a language to use, defaults to C99 by default. Autosenses based on
154extension.
155
156using a makefile
157</p>
158
159
160<!-- ======================================================================= -->
161<h2 id="commandline">Command Line Options</h2>
162<!-- ======================================================================= -->
163
164<p>
165This section is generally an index into other sections. It does not go into
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]166depth on the ones that are covered by other sections. However, the first part
167introduces the language selection and other high level options like -c, -g, etc.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]168</p>
169
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]170
171<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
172<h3 id="cl_diagnostics">Options to Control Error and Warning Messages</h3>
173<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
174
175<p><b>-Werror</b>: Turn warnings into errors.</p>
176<p><b>-Werror=foo</b>: Turn warning "foo" into an error.</p>
177<p><b>-Wno-error=foo</b>: Turn warning "foo" into an warning even if -Werror is
178 specified.</p>
179<p><b>-Wfoo</b>: Enable warning foo</p>
180<p><b>-Wno-foo</b>: Disable warning foo</p>
181<p><b>-w</b>: Disable all warnings.</p>
182<p><b>-pedantic</b>: Warn on language extensions.</p>
183<p><b>-pedantic-errors</b>: Error on language extensions.</p>
184<p><b>-Wsystem-headers</b>: Enable warnings from system headers.</p>
185
Chris Lattner0f0c9632010-04-07 20:49:23 +0000[diff] [blame]186<p><b>-ferror-limit=123</b>: Stop emitting diagnostics after 123 errors have
187 been produced. The default is 20, and the error limit can be disabled with
188 -ferror-limit=0.</p>
189
Douglas Gregor575cf372010-04-20 07:18:24 +0000[diff] [blame]190<p><b>-ftemplate-backtrace-limit=123</b>: Only emit up to 123 template instantiation notes within the template instantiation backtrace for a single warning or error. The default is 10, and the limit can be disabled with -ftemplate-backtrace-limit=0.</p>
191
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]192<!-- ================================================= -->
193<h4 id="cl_diag_formatting">Formatting of Diagnostics</h4>
194<!-- ================================================= -->
195
196<p>Clang aims to produce beautiful diagnostics by default, particularly for new
197users that first come to Clang. However, different people have different
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]198preferences, and sometimes Clang is driven by another program that wants to
199parse simple and consistent output, not a person. For these cases, Clang
200provides a wide range of options to control the exact output format of the
201diagnostics that it generates.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]202
203<dl>
204
205<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
206<dt id="opt_fshow-column"><b>-f[no-]show-column</b>: Print column number in
207diagnostic.</dt>
208<dd>This option, which defaults to on, controls whether or not Clang prints the
209column number of a diagnostic. For example, when this is enabled, Clang will
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]210print something like:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]211
212<pre>
213 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
214 #endif bad
215 ^
216 //
217</pre>
218
219<p>When this is disabled, Clang will print "test.c:28: warning..." with no
220column number.</p>
221</dd>
222
223<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
224<dt id="opt_fshow-source-location"><b>-f[no-]show-source-location</b>: Print
225source file/line/column information in diagnostic.</dt>
226<dd>This option, which defaults to on, controls whether or not Clang prints the
227filename, line number and column number of a diagnostic. For example,
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]228when this is enabled, Clang will print something like:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]229
230<pre>
231 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
232 #endif bad
233 ^
234 //
235</pre>
236
237<p>When this is disabled, Clang will not print the "test.c:28:8: " part.</p>
238</dd>
239
240<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
241<dt id="opt_fcaret-diagnostics"><b>-f[no-]caret-diagnostics</b>: Print source
242line and ranges from source code in diagnostic.</dt>
243<dd>This option, which defaults to on, controls whether or not Clang prints the
244source line, source ranges, and caret when emitting a diagnostic. For example,
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]245when this is enabled, Clang will print something like:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]246
247<pre>
248 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
249 #endif bad
250 ^
251 //
252</pre>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]253</dd>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]254<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Douglas Gregore8d44dd2010-07-09 16:31:58 +0000[diff] [blame]255<dt id="opt_fcolor_diagnostics"><b>-f[no-]color-diagnostics</b>: </dt>
256<dd>This option, which defaults to on when a color-capable terminal is
257 detected, controls whether or not Clang prints diagnostics in color.
258 When this option is enabled, Clang will use colors to highlight
259 specific parts of the diagnostic, e.g.,
Chris Lattner4e1c53d2011-01-24 03:47:34 +0000[diff] [blame]260 <pre>
261 <b><font color="black">test.c:28:8: <font color="magenta">warning</font>: extra tokens at end of #endif directive [-Wextra-tokens]</font></b>
262 #endif bad
263 <font color="green">^</font>
264 <font color="green">//</font>
Douglas Gregore8d44dd2010-07-09 16:31:58 +0000[diff] [blame]265</pre>
Chris Lattner4e1c53d2011-01-24 03:47:34 +0000[diff] [blame]266
267<p>When this is disabled, Clang will just print:</p>
268
269<pre>
270 test.c:2:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
271 #endif bad
272 ^
273 //
274</pre>
275</dd>
Douglas Gregore8d44dd2010-07-09 16:31:58 +0000[diff] [blame]276<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Douglas Gregorc9471b02011-05-21 17:07:29 +0000[diff] [blame]277<dt id="opt_fdiagnostics-format"><b>-fdiagnostics-format=clang/msvc/vi</b>:
278Changes diagnostic output format to better match IDEs and command line tools.</dt>
279<dd>This option controls the output format of the filename, line number, and column printed in diagnostic messages. The options, and their affect on formatting a simple conversion diagnostic, follow:
280
281 <dl>
282 <dt><b>clang</b> (default)</dt>
283 <dd>
284 <pre>t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int'</pre>
285 </dd>
286
287 <dt><b>msvc</b></dt>
288 <dd>
289 <pre>t.c(3,11) : warning: conversion specifies type 'char *' but the argument has type 'int'</pre>
290 </dd>
291
292 <dt><b>vi</b></dt>
293 <dd>
294 <pre>t.c +3:11: warning: conversion specifies type 'char *' but the argument has type 'int'</pre>
295 </dd>
296 </dl>
297</dd>
298
299<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Argyrios Kyrtzidis477aab62011-05-25 05:05:01 +0000[diff] [blame]300<dt id="opt_fdiagnostics-show-name"><b>-f[no-]diagnostics-show-name</b>:
301Enable the display of the diagnostic name.</dt>
302<dd>This option, which defaults to off, controls whether or not
303Clang prints the associated name.</dd>
304<br>
305<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]306<dt id="opt_fdiagnostics-show-option"><b>-f[no-]diagnostics-show-option</b>:
307Enable <tt>[-Woption]</tt> information in diagnostic line.</dt>
308<dd>This option, which defaults to on,
309controls whether or not Clang prints the associated <A
310href="#cl_diag_warning_groups">warning group</a> option name when outputting
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]311a warning diagnostic. For example, in this output:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]312
313<pre>
314 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
315 #endif bad
316 ^
317 //
318</pre>
319
320<p>Passing <b>-fno-diagnostics-show-option</b> will prevent Clang from printing
321the [<a href="#opt_Wextra-tokens">-Wextra-tokens</a>] information in the
322diagnostic. This information tells you the flag needed to enable or disable the
323diagnostic, either from the command line or through <a
324href="#pragma_GCC_diagnostic">#pragma GCC diagnostic</a>.</dd>
325
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]326<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
327<dt id="opt_fdiagnostics-show-category"><b>-fdiagnostics-show-category=none/id/name</b>:
328Enable printing category information in diagnostic line.</dt>
329<dd>This option, which defaults to "none",
330controls whether or not Clang prints the category associated with a diagnostic
331when emitting it. Each diagnostic may or many not have an associated category,
332if it has one, it is listed in the diagnostic categorization field of the
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]333diagnostic line (in the []'s).
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]334
335<p>For example, a format string warning will produce these three renditions
336based on the setting of this option:</p>
337
338<pre>
339 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat]
340 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,1</b>]
341 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,Format String</b>]
342</pre>
343
344<p>This category can be used by clients that want to group diagnostics by
345category, so it should be a high level category. We want dozens of these, not
346hundreds or thousands of them.</p>
347</dd>
348
349
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]350
351<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
352<dt id="opt_fdiagnostics-fixit-info"><b>-f[no-]diagnostics-fixit-info</b>:
353Enable "FixIt" information in the diagnostics output.</dt>
354<dd>This option, which defaults to on, controls whether or not Clang prints the
355information on how to fix a specific diagnostic underneath it when it knows.
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]356For example, in this output:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]357
358<pre>
359 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
360 #endif bad
361 ^
362 //
363</pre>
364
365<p>Passing <b>-fno-diagnostics-fixit-info</b> will prevent Clang from printing
366the "//" line at the end of the message. This information is useful for users
367who may not understand what is wrong, but can be confusing for machine
368parsing.</p>
369</dd>
370
371<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]372<dt id="opt_fdiagnostics-print-source-range-info">
373<b>-f[no-]diagnostics-print-source-range-info</b>:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]374Print machine parsable information about source ranges.</dt>
375<dd>This option, which defaults to off, controls whether or not Clang prints
376information about source ranges in a machine parsable format after the
377file/line/column number information. The information is a simple sequence of
378brace enclosed ranges, where each range lists the start and end line/column
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]379locations. For example, in this output:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]380
381<pre>
382exprs.c:47:15:{47:8-47:14}{47:17-47:24}: error: invalid operands to binary expression ('int *' and '_Complex float')
383 P = (P-42) + Gamma*4;
384 ~~~~~~ ^ ~~~~~~~
385</pre>
386
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]387<p>The {}'s are generated by -fdiagnostics-print-source-range-info.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]388</dd>
389
Douglas Gregor4786c152010-08-19 20:24:43 +0000[diff] [blame]390<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
391<dt id="opt_fdiagnostics-parseable-fixits">
392<b>-fdiagnostics-parseable-fixits</b>:
393Print Fix-Its in a machine parseable form.</dt>
394<dd><p>This option makes Clang print available Fix-Its in a machine parseable format at the end of diagnostics. The following example illustrates the format:</p>
395
396<pre>
Douglas Gregorbf5e09d2010-08-20 03:17:33 +0000[diff] [blame]397 fix-it:"t.cpp":{7:25-7:29}:"Gamma"
Douglas Gregor4786c152010-08-19 20:24:43 +0000[diff] [blame]398</pre>
399
NAKAMURA Takumi44626362011-04-05 00:57:02 +0000[diff] [blame]400<p>The range printed is a half-open range, so in this example the characters at
401column 25 up to but not including column 29 on line 7 in t.cpp should be
402replaced with the string &quot;Gamma&quot;. Either the range or the replacement
403string may be empty (representing strict insertions and strict erasures,
404respectively). Both the file name and the insertion string escape backslash (as
405&quot;\\&quot;), tabs (as &quot;\t&quot;), newlines (as &quot;\n&quot;), double
406quotes(as &quot;\&quot;&quot;) and non-printable characters (as octal
407&quot;\xxx&quot;).</p>
Douglas Gregor4786c152010-08-19 20:24:43 +0000[diff] [blame]408</dd>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]409
410</dl>
411
412
413
414
415<!-- ===================================================== -->
416<h4 id="cl_diag_warning_groups">Individual Warning Groups</h4>
417<!-- ===================================================== -->
418
419<p>TODO: Generate this from tblgen. Define one anchor per warning group.</p>
420
421
422<dl>
423
424
425<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
426<dt id="opt_Wextra-tokens"><b>-Wextra-tokens</b>: Warn about excess tokens at
427 the end of a preprocessor directive.</dt>
428<dd>This option, which defaults to on, enables warnings about extra tokens at
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]429the end of preprocessor directives. For example:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]430
431<pre>
432 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
433 #endif bad
434 ^
435</pre>
436
437<p>These extra tokens are not strictly conforming, and are usually best handled
438by commenting them out.</p>
439
440<p>This option is also enabled by <a href="">-Wfoo</a>, <a href="">-Wbar</a>,
441 and <a href="">-Wbaz</a>.</p>
442</dd>
443
Jeffrey Yasskin21d07e42010-06-05 01:39:57 +0000[diff] [blame]444<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
445<dt id="opt_Wambiguous-member-template"><b>-Wambiguous-member-template</b>:
446Warn about unqualified uses of a member template whose name resolves
447to another template at the location of the use.</dt>
448<dd>This option, which defaults to on, enables a warning in the
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]449following code:
Jeffrey Yasskin21d07e42010-06-05 01:39:57 +0000[diff] [blame]450
451<pre>
452template&lt;typename T> struct set{};
453template&lt;typename T> struct trait { typedef const T& type; };
454struct Value {
455 template&lt;typename T> void set(typename trait&lt;T>::type value) {}
456};
457void foo() {
458 Value v;
459 v.set&lt;double>(3.2);
460}
461</pre>
462
463<p>C++ [basic.lookup.classref] requires this to be an error, but,
464because it's hard to work around, Clang downgrades it to a warning as
465an extension.</p>
466</dd>
467
Jeffrey Yasskin57d12fd2010-06-07 15:58:05 +0000[diff] [blame]468<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
469<dt id="opt_Wbind-to-temporary-copy"><b>-Wbind-to-temporary-copy</b>: Warn about
470an unusable copy constructor when binding a reference to a temporary.</dt>
471<dd>This option, which defaults to on, enables warnings about binding a
472reference to a temporary when the temporary doesn't have a usable copy
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]473constructor. For example:
Jeffrey Yasskin57d12fd2010-06-07 15:58:05 +0000[diff] [blame]474
475<pre>
476 struct NonCopyable {
477 NonCopyable();
478 private:
479 NonCopyable(const NonCopyable&);
480 };
481 void foo(const NonCopyable&);
482 void bar() {
483 foo(NonCopyable()); // Disallowed in C++98; allowed in C++0x.
484 }
485</pre>
486<pre>
487 struct NonCopyable2 {
488 NonCopyable2();
489 NonCopyable2(NonCopyable2&);
490 };
491 void foo(const NonCopyable2&);
492 void bar() {
493 foo(NonCopyable2()); // Disallowed in C++98; allowed in C++0x.
494 }
495</pre>
496
497<p>Note that if <tt>NonCopyable2::NonCopyable2()</tt> has a default
498argument whose instantiation produces a compile error, that error will
499still be a hard error in C++98 mode even if this warning is turned
500off.</p>
501
502</dd>
503
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]504</dl>
505
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]506<!-- ======================================================================= -->
507<h2 id="general_features">Language and Target-Independent Features</h2>
508<!-- ======================================================================= -->
509
510
511<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
512<h3 id="diagnostics">Controlling Errors and Warnings</h3>
513<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
514
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]515<p>Clang provides a number of ways to control which code constructs cause it to
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]516emit errors and warning messages, and how they are displayed to the console.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]517
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]518<h4 id="diagnostics_display">Controlling How Clang Displays Diagnostics</h4>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]519
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]520<p>When Clang emits a diagnostic, it includes rich information in the output,
521and gives you fine-grain control over which information is printed. Clang has
522the ability to print this information, and these are the options that control
523it:</p>
524
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]525<ol>
526<li>A file/line/column indicator that shows exactly where the diagnostic occurs
527 in your code [<a href="#opt_fshow-column">-fshow-column</a>, <a
528 href="#opt_fshow-source-location">-fshow-source-location</a>].</li>
529<li>A categorization of the diagnostic as a note, warning, error, or fatal
530 error.</li>
531<li>A text string that describes what the problem is.</li>
Argyrios Kyrtzidis477aab62011-05-25 05:05:01 +0000[diff] [blame]532<li>An option that indicates whether to print the diagnostic name [<a
533 href="#opt_fdiagnostics-show-name">-fdiagnostics-show-name</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]534<li>An option that indicates how to control the diagnostic (for diagnostics that
535 support it) [<a
536 href="#opt_fdiagnostics-show-option">-fdiagnostics-show-option</a>].</li>
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]537<li>A <a href="#diagnostics_categories">high-level category</a> for the
538 diagnostic for clients that want to group diagnostics by class (for
539 diagnostics that support it) [<a
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]540 href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]541<li>The line of source code that the issue occurs on, along with a caret and
542 ranges that indicate the important locations [<a
543 href="opt_fcaret-diagnostics">-fcaret-diagnostics</a>].</li>
544<li>"FixIt" information, which is a concise explanation of how to fix the
545 problem (when Clang is certain it knows) [<a
546 href="opt_fdiagnostics-fixit-info">-fdiagnostics-fixit-info</a>].</li>
547<li>A machine-parsable representation of the ranges involved (off by
548 default) [<a
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]549 href="opt_fdiagnostics-print-source-range-info">-fdiagnostics-print-source-range-info</a>].</li>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]550</ol>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]551
552<p>For more information please see <a href="#cl_diag_formatting">Formatting of
553Diagnostics</a>.</p>
554
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]555
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]556<h4 id="diagnostics_mappings">Diagnostic Mappings</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]557
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]558<p>All diagnostics are mapped into one of these 5 classes:</p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]559
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]560<ul>
561<li>Ignored</li>
562<li>Note</li>
563<li>Warning</li>
564<li>Error</li>
565<li>Fatal</li>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]566</ul>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]567
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]568<h4 id="diagnostics_categories">Diagnostic Categories</h4>
569
570<p>Though not shown by default, diagnostics may each be associated with a
571 high-level category. This category is intended to make it possible to triage
572 builds that produce a large number of errors or warnings in a grouped way.
573</p>
574
575<p>Categories are not shown by default, but they can be turned on with the
576<a href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a> option.
577When set to "<tt>name</tt>", the category is printed textually in the diagnostic
578output. When it is set to "<tt>id</tt>", a category number is printed. The
579mapping of category names to category id's can be obtained by running '<tt>clang
580 --print-diagnostic-categories</tt>'.
581</p>
582
583<h4 id="diagnostics_commandline">Controlling Diagnostics via Command Line
584 Flags</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]585
586<p>-W flags, -pedantic, etc</p>
587
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]588<h4 id="diagnostics_pragmas">Controlling Diagnostics via Pragmas</h4>
589
590<p>Clang can also control what diagnostics are enabled through the use of
591pragmas in the source code. This is useful for turning off specific warnings
592in a section of source code. Clang supports GCC's pragma for compatibility
593with existing source code, as well as several extensions. </p>
594
595<p>The pragma may control any warning that can be used from the command line.
596Warnings may be set to ignored, warning, error, or fatal. The following
597example code will tell Clang or GCC to ignore the -Wall warnings:</p>
598
599<pre>
600#pragma GCC diagnostic ignored "-Wall"
601</pre>
602
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]603<p>In addition to all of the functionality provided by GCC's pragma, Clang
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]604also allows you to push and pop the current warning state. This is particularly
605useful when writing a header file that will be compiled by other people, because
606you don't know what warning flags they build with.</p>
607
608<p>In the below example
609-Wmultichar is ignored for only a single line of code, after which the
610diagnostics return to whatever state had previously existed.</p>
611
612<pre>
613#pragma clang diagnostic push
614#pragma clang diagnostic ignored "-Wmultichar"
615
616char b = 'df'; // no warning.
617
618#pragma clang diagnostic pop
619</pre>
620
621<p>The push and pop pragmas will save and restore the full diagnostic state of
622the compiler, regardless of how it was set. That means that it is possible to
623use push and pop around GCC compatible diagnostics and Clang will push and pop
624them appropriately, while GCC will ignore the pushes and pops as unknown
625pragmas. It should be noted that while Clang supports the GCC pragma, Clang and
626GCC do not support the exact same set of warnings, so even when using GCC
627compatible #pragmas there is no guarantee that they will have identical behaviour
628on both compilers. </p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]629
Ted Kremenekf391fa72011-08-18 01:17:05 +0000[diff] [blame]630<h4 id="diagnostics_enable_everything">Enabling All Warnings</h4>
631
632<p>In addition to the traditional <tt>-W</tt> flags, one can enable <b>all</b>
633 warnings by passing <tt>-Weverything</tt>.
634 This works as expected with <tt>-Werror</tt>,
635 and also includes the warnings from <tt>-pedantic</tt>.</p>
636
637<p>Note that when combined with <tt>-w</tt> (which disables all warnings), that
638 flag wins.</p>
639
Ted Kremenek2fb11eb2010-08-24 18:12:35 +0000[diff] [blame]640<h4 id="analyzer_diagnositics">Controlling Static Analyzer Diagnostics</h4>
641
642<p>While not strictly part of the compiler, the diagnostics from Clang's <a
643href="http://clang-analyzer.llvm.org">static analyzer</a> can also be influenced
644by the user via changes to the source code. This can be done in two ways:
645
646<ul>
647
648<li id="analyzer_annotations"><b>Annotations</b>: The static analyzer recognizes various GCC-style
649attributes (e.g., <tt>__attribute__((nonnull)))</tt>) that can either suppress
650static analyzer warnings or teach the analyzer about code invariants which
651enable it to find more bugs. While many of these attributes are standard GCC
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]652attributes, additional ones have been added to Clang to specifically support the
Ted Kremenek2fb11eb2010-08-24 18:12:35 +0000[diff] [blame]653static analyzer. Detailed information on these annotations can be found in the
654<a href="http://clang-analyzer.llvm.org/annotations.html">analyzer's
655documentation</a>.</li>
656
657<li><b><tt>__clang_analyzer__</tt></b>: When the static analyzer is using Clang
658to parse source files, it implicitly defines the preprocessor macro
659<tt>__clang_analyzer__</tt>. While discouraged, code can use this macro to
660selectively exclude code the analyzer examines. Here is an example:
661
662<pre>
663#ifndef __clang_analyzer__
664// Code not to be analyzed
665#endif
666</pre>
667
668In general, this usage is discouraged. Instead, we prefer that users file bugs
669against the analyzer when it flags false positives. There is also active
670discussion of allowing users in the future to selectively silence specific
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]671analyzer warnings (some of which can already be done using <a
Ted Kremenek2fb11eb2010-08-24 18:12:35 +0000[diff] [blame]672href="analyzer_annotations">annotations</a>).</li>
673
674</ul>
675
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]676<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
677<h3 id="precompiledheaders">Precompiled Headers</h3>
678<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
679
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]680<p><a href="http://en.wikipedia.org/wiki/Precompiled_header">Precompiled
681headers</a> are a general approach employed by many compilers to reduce
682compilation time. The underlying motivation of the approach is that it is
683common for the same (and often large) header files to be included by
684multiple source files. Consequently, compile times can often be greatly improved
685by caching some of the (redundant) work done by a compiler to process headers.
686Precompiled header files, which represent one of many ways to implement
687this optimization, are literally files that represent an on-disk cache that
688contains the vital information necessary to reduce some of the work
689needed to process a corresponding header file. While details of precompiled
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]690headers vary between compilers, precompiled headers have been shown to be
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]691highly effective at speeding up program compilation on systems with very large
692system headers (e.g., Mac OS/X).</p>
693
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]694<h4>Generating a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]695
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]696<p>To generate a PCH file using Clang, one invokes Clang with
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]697the <b><tt>-x <i>&lt;language&gt;</i>-header</tt></b> option. This mirrors the
698interface in GCC for generating PCH files:</p>
699
700<pre>
701 $ gcc -x c-header test.h -o test.h.gch
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]702 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]703</pre>
704
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]705<h4>Using a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]706
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]707<p>A PCH file can then be used as a prefix header when a
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]708<b><tt>-include</tt></b> option is passed to <tt>clang</tt>:</p>
709
710<pre>
711 $ clang -include test.h test.c -o test
712</pre>
713
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]714<p>The <tt>clang</tt> driver will first check if a PCH file for <tt>test.h</tt>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]715is available; if so, the contents of <tt>test.h</tt> (and the files it includes)
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]716will be processed from the PCH file. Otherwise, Clang falls back to
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]717directly processing the content of <tt>test.h</tt>. This mirrors the behavior of
718GCC.</p>
719
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]720<p><b>NOTE:</b> Clang does <em>not</em> automatically use PCH files
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]721for headers that are directly included within a source file. For example:</p>
722
723<pre>
Chris Lattnere42ec542009-06-13 20:35:58 +0000[diff] [blame]724 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]725 $ cat test.c
726 #include "test.h"
727 $ clang test.c -o test
728</pre>
729
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]730<p>In this example, <tt>clang</tt> will not automatically use the PCH file for
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]731<tt>test.h</tt> since <tt>test.h</tt> was included directly in the source file
732and not specified on the command line using <tt>-include</tt>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]733
Douglas Gregore650c8c2009-07-07 00:12:59 +0000[diff] [blame]734<h4>Relocatable PCH Files</h4>
735<p>It is sometimes necessary to build a precompiled header from headers that
736are not yet in their final, installed locations. For example, one might build a
737precompiled header within the build tree that is then meant to be installed
738alongside the headers. Clang permits the creation of "relocatable" precompiled
739headers, which are built with a given path (into the build directory) and can
740later be used from an installed location.</p>
741
742<p>To build a relocatable precompiled header, place your headers into a
743subdirectory whose structure mimics the installed location. For example, if you
744want to build a precompiled header for the header <code>mylib.h</code> that
745will be installed into <code>/usr/include</code>, create a subdirectory
746<code>build/usr/include</code> and place the header <code>mylib.h</code> into
747that subdirectory. If <code>mylib.h</code> depends on other headers, then
748they can be stored within <code>build/usr/include</code> in a way that mimics
749the installed location.</p>
750
751<p>Building a relocatable precompiled header requires two additional arguments.
752First, pass the <code>--relocatable-pch</code> flag to indicate that the
753resulting PCH file should be relocatable. Second, pass
754<code>-isysroot /path/to/build</code>, which makes all includes for your
755library relative to the build directory. For example:</p>
756
757<pre>
758 # clang -x c-header --relocatable-pch -isysroot /path/to/build /path/to/build/mylib.h mylib.h.pch
759</pre>
760
761<p>When loading the relocatable PCH file, the various headers used in the PCH
762file are found from the system header root. For example, <code>mylib.h</code>
763can be found in <code>/usr/include/mylib.h</code>. If the headers are installed
764in some other system root, the <code>-isysroot</code> option can be used provide
765a different system root from which the headers will be based. For example,
766<code>-isysroot /Developer/SDKs/MacOSX10.4u.sdk</code> will look for
767<code>mylib.h</code> in
768<code>/Developer/SDKs/MacOSX10.4u.sdk/usr/include/mylib.h</code>.</p>
769
770<p>Relocatable precompiled headers are intended to be used in a limited number
771of cases where the compilation environment is tightly controlled and the
772precompiled header cannot be generated after headers have been installed.
773Relocatable precompiled headers also have some performance impact, because
774the difference in location between the header locations at PCH build time vs.
775at the time of PCH use requires one of the PCH optimizations,
776<code>stat()</code> caching, to be disabled. However, this change is only
777likely to affect PCH files that reference a large number of headers.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]778
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]779<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
780<h3 id="codegen">Controlling Code Generation</h3>
781<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
782
783<p>Clang provides a number of ways to control code generation. The options are listed below.</p>
784
785<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]786<dl>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]787<dt id="opt_fcatch-undefined-behavior"><b>-fcatch-undefined-behavior</b>: Turn
788on runtime code generation to check for undefined behavior.</dt>
789
790<dd>This option, which defaults to off, controls whether or not Clang
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]791adds runtime checks for undefined runtime behavior. If a check fails,
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]792<tt>__builtin_trap()</tt> is used to indicate failure.
793The checks are:
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]794<ul>
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]795<li>Subscripting where the static type of one operand is a variable
Mike Stump88b2a172009-12-16 03:25:12 +0000[diff] [blame]796 which is decayed from an array type and the other operand is
797 greater than the size of the array or less than zero.</li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]798<li>Shift operators where the amount shifted is greater or equal to the
799 promoted bit-width of the left-hand-side or less than zero.</li>
Mike Stump8f6a3ed2009-12-16 03:18:14 +0000[diff] [blame]800<li>If control flow reaches __builtin_unreachable.
801<li>When llvm implements more __builtin_object_size support, reads and
802 writes for objects that __builtin_object_size indicates we aren't
803 accessing valid memory. Bit-fields and vectors are not yet checked.
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]804</ul>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]805</dd>
806
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]807<dt id="opt_fno-assume-sane-operator-new"><b>-fno-assume-sane-operator-new</b>:
808Don't assume that the C++'s new operator is sane.</dt>
Nuno Lopesb23f20d2009-12-17 10:15:49 +0000[diff] [blame]809<dd>This option tells the compiler to do not assume that C++'s global new
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]810operator will always return a pointer that does not
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]811alias any other pointer when the function returns.</dd>
Evan Chengfda026b2011-04-08 22:18:01 +0000[diff] [blame]812
Evan Cheng4cf4b592011-04-08 22:34:21 +0000[diff] [blame]813<dt id="opt_ftrap-function"><b>-ftrap-function=[name]</b>: Instruct code
Evan Chengfda026b2011-04-08 22:18:01 +0000[diff] [blame]814generator to emit a function call to the specified function name for
815<tt>__builtin_trap()</tt>.</dt>
816
817<dd>LLVM code generator translates <tt>__builtin_trap()</tt> to a trap
818instruction if it is supported by the target ISA. Otherwise, the builtin is
819translated into a call to <tt>abort</tt>. If this option is set, then the code
820generator will always lower the builtin to a call to the specified function
821regardless of whether the target ISA has a trap instruction. This option is
Chris Lattner02e266b2011-04-08 22:58:43 +0000[diff] [blame]822useful for environments (e.g. deeply embedded) where a trap cannot be properly
823handled, or when some custom behavior is desired.</dd>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]824</dl>
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]825
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]826<!-- ======================================================================= -->
827<h2 id="c">C Language Features</h2>
828<!-- ======================================================================= -->
829
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]830<p>The support for standard C in clang is feature-complete except for the C99
831floating-point pragmas.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]832
833<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]834<h3 id="c_ext">Extensions supported by clang</h3>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]835<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
836
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]837<p>See <a href="LanguageExtensions.html">clang language extensions</a>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]838
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]839<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
840<h3 id="c_modes">Differences between various standard modes</h3>
841<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]842
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]843<p>clang supports the -std option, which changes what language mode clang uses.
844The supported modes for C are c89, gnu89, c94, c99, gnu99 and various aliases
845for those modes. If no -std option is specified, clang defaults to gnu99 mode.
846</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]847
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]848<p>Differences between all c* and gnu* modes:</p>
849<ul>
850<li>c* modes define "__STRICT_ANSI__".</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]851<li>Target-specific defines not prefixed by underscores, like "linux", are
852defined in gnu* modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]853<li>Trigraphs default to being off in gnu* modes; they can be enabled by the
854-trigraphs option.</li>
855<li>The parser recognizes "asm" and "typeof" as keywords in gnu* modes; the
856variants "__asm__" and "__typeof__" are recognized in all modes.</li>
Eli Friedmanb0ac5452009-05-16 23:17:30 +0000[diff] [blame]857<li>The Apple "blocks" extension is recognized by default in gnu* modes
858on some platforms; it can be enabled in any mode with the "-fblocks"
859option.</li>
Chris Lattner741be6a2011-06-15 00:41:40 +0000[diff] [blame]860<li>Arrays that are VLA's according to the standard, but which can be constant
861 folded by the frontend are treated as fixed size arrays. This occurs for
862 things like "int X[(1, 2)];", which is technically a VLA. c* modes are
863 strictly compliant and treat these as VLAs.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]864</ul>
865
866<p>Differences between *89 and *99 modes:</p>
867<ul>
868<li>The *99 modes default to implementing "inline" as specified in C99, while
869the *89 modes implement the GNU version. This can be overridden for individual
870functions with the __gnu_inline__ attribute.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]871<li>Digraphs are not recognized in c89 mode.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]872<li>The scope of names defined inside a "for", "if", "switch", "while", or "do"
873statement is different. (example: "if ((struct x {int x;}*)0) {}".)</li>
874<li>__STDC_VERSION__ is not defined in *89 modes.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]875<li>"inline" is not recognized as a keyword in c89 mode.</li>
876<li>"restrict" is not recognized as a keyword in *89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]877<li>Commas are allowed in integer constant expressions in *99 modes.</li>
878<li>Arrays which are not lvalues are not implicitly promoted to pointers in
879*89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]880<li>Some warnings are different.</li>
881</ul>
882
883<p>c94 mode is identical to c89 mode except that digraphs are enabled in
884c94 mode (FIXME: And __STDC_VERSION__ should be defined!).</p>
885
886<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
887<h3 id="c_unimpl_gcc">GCC extensions not implemented yet</h3>
888<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
889
890<p>clang tries to be compatible with gcc as much as possible, but some gcc
891extensions are not implemented yet:</p>
892
893<ul>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]894
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]895<li>clang does not support #pragma weak
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]896(<a href="http://llvm.org/bugs/show_bug.cgi?id=3679">bug 3679</a>). Due to
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]897the uses described in the bug, this is likely to be implemented at some
898point, at least partially.</li>
899
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]900<li>clang does not support decimal floating point types (_Decimal32 and
901friends) or fixed-point types (_Fract and friends); nobody has expressed
902interest in these features yet, so it's hard to say when they will be
903implemented.</li>
904
905<li>clang does not support nested functions; this is a complex feature which
906is infrequently used, so it is unlikely to be implemented anytime soon.</li>
907
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]908<li>clang does not support global register variables, this is unlikely
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]909to be implemented soon because it requires additional LLVM backend support.
910</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]911
912<li>clang does not support static initialization of flexible array
913members. This appears to be a rarely used extension, but could be
914implemented pending user demand.</li>
915
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]916<li>clang does not support __builtin_va_arg_pack/__builtin_va_arg_pack_len.
917This is used rarely, but in some potentially interesting places, like the
918glibc headers, so it may be implemented pending user demand. Note that
919because clang pretends to be like GCC 4.2, and this extension was introduced
920in 4.3, the glibc headers will not try to use this extension with clang at
921the moment.</li>
922
923<li>clang does not support the gcc extension for forward-declaring function
924parameters; this has not showed up in any real-world code yet, though, so it
925might never be implemented.</li>
926
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]927</ul>
928
929<p>This is not a complete list; if you find an unsupported extension
930missing from this list, please send an e-mail to cfe-dev. This list
931currently excludes C++; see <a href="#cxx">C++ Language Features</a>.
932Also, this list does not include bugs in mostly-implemented features; please
933see the <a href="http://llvm.org/bugs/buglist.cgi?quicksearch=product%3Aclang+component%3A-New%2BBugs%2CAST%2CBasic%2CDriver%2CHeaders%2CLLVM%2BCodeGen%2Cparser%2Cpreprocessor%2CSemantic%2BAnalyzer">
934bug tracker</a> for known existing bugs (FIXME: Is there a section for
935bug-reporting guidelines somewhere?).</p>
936
937<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
938<h3 id="c_unsupp_gcc">Intentionally unsupported GCC extensions</h3>
939<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
940
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]941<ul>
942
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]943<li>clang does not support the gcc extension that allows variable-length arrays
Chris Lattnerd462b6a2011-01-24 03:47:59 +0000[diff] [blame]944in structures. This is for a few reasons: one, it is tricky
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]945to implement, two, the extension is completely undocumented, and three, the
Chris Lattnerb9d511c2010-10-28 02:20:32 +0000[diff] [blame]946extension appears to be rarely used. Note that clang <em>does</em> support
947flexible array members (arrays with a zero or unspecified size at the end of
948a structure).</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]949
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]950<li>clang does not have an equivalent to gcc's "fold"; this means that
951clang doesn't accept some constructs gcc might accept in contexts where a
Eli Friedman248f9732011-06-13 23:12:01 +0000[diff] [blame]952constant expression is required, like "x-x" where x is a variable.</li>
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]953
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]954<li>clang does not support multiple alternative constraints in inline asm; this
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]955is an extremely obscure feature which would be complicated to implement
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]956correctly.</li>
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]957
958<li>clang does not support __builtin_apply and friends; this extension is
959extremely obscure and difficult to implement reliably.</li>
960
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]961</ul>
962
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]963<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
964<h3 id="c_ms">Microsoft extensions</h3>
965<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
966
967<p>clang has some experimental support for extensions from
968Microsoft Visual C++; to enable it, use the -fms-extensions command-line
Eli Friedmana1821b52009-06-08 06:21:03 +0000[diff] [blame]969option. This is the default for Windows targets. Note that the
970support is incomplete; enabling Microsoft extensions will silently drop
971certain constructs (including __declspec and Microsoft-style asm statements).
972</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]973
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]974<ul>
Douglas Gregord3b227d2010-12-14 16:21:49 +0000[diff] [blame]975<li>clang allows setting _MSC_VER with -fmsc-version=. It defaults to 1300 which
Michael J. Spencerdae4ac42010-10-21 05:21:48 +0000[diff] [blame]976is the same as Visual C/C++ 2003. Any number is supported and can greatly affect
977what Windows SDK and c++stdlib headers clang can compile. This option will be
978removed when clang supports the full set of MS extensions required for these
979headers.</li>
980
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]981<li>clang does not support the Microsoft extension where anonymous
Daniel Dunbar869e0542009-05-13 00:23:53 +0000[diff] [blame]982record members can be declared using user defined typedefs.</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]983
Daniel Dunbar9375ed12009-05-13 21:40:49 +0000[diff] [blame]984<li>clang supports the Microsoft "#pragma pack" feature for
985controlling record layout. GCC also contains support for this feature,
986however where MSVC and GCC are incompatible clang follows the MSVC
987definition.</li>
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]988</ul>
Daniel Dunbar9375ed12009-05-13 21:40:49 +0000[diff] [blame]989
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]990<!-- ======================================================================= -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]991<h2 id="target_features">Target-Specific Features and Limitations</h2>
992<!-- ======================================================================= -->
993
994
995<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
996<h3 id="target_arch">CPU Architectures Features and Limitations</h3>
997<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
998
999<!-- ======================== -->
1000<h4 id="target_arch_x86">X86</h4>
1001<!-- ======================== -->
Daniel Dunbarbcaf7aa2010-09-19 19:26:59 +0000[diff] [blame]1002
1003<p>The support for X86 (both 32-bit and 64-bit) is considered stable on Darwin
1004(Mac OS/X), Linux, FreeBSD, and Dragonfly BSD: it has been tested to correctly
1005compile many large C, C++, Objective-C, and Objective-C++ codebases.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]1006
NAKAMURA Takumiddf68ff2011-04-05 04:31:31 +0000[diff] [blame]1007<p>On x86_64-mingw32, passing i128(by value) is incompatible to Microsoft x64
1008calling conversion. You might need to tweak WinX86_64ABIInfo::classify()
1009in lib/CodeGen/TargetInfo.cpp.</p>
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]1010
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]1011<!-- ======================== -->
1012<h4 id="target_arch_arm">ARM</h4>
1013<!-- ======================== -->
Daniel Dunbarbcaf7aa2010-09-19 19:26:59 +0000[diff] [blame]1014
1015<p>The support for ARM (specifically ARMv6 and ARMv7) is considered stable on
1016Darwin (iOS): it has been tested to correctly compile many large C, C++,
Bob Wilson4ea8dfa2011-01-10 17:55:14 +0000[diff] [blame]1017Objective-C, and Objective-C++ codebases. Clang only supports a limited number
1018of ARM architectures. It does not yet fully support ARMv5, for example.</p>
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]1019
1020<!-- ======================== -->
1021<h4 id="target_arch_other">Other platforms</h4>
1022<!-- ======================== -->
1023clang currently contains some support for PPC and Sparc; however, significant
1024pieces of code generation are still missing, and they haven't undergone
1025significant testing.
1026
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]1027<p>clang contains limited support for the MSP430 embedded processor, but both
1028the clang support and the LLVM backend support are highly experimental.
1029
1030<p>Other platforms are completely unsupported at the moment. Adding the
1031minimal support needed for parsing and semantic analysis on a new platform
1032is quite easy; see lib/Basic/Targets.cpp in the clang source tree. This level
1033of support is also sufficient for conversion to LLVM IR for simple programs.
1034Proper support for conversion to LLVM IR requires adding code to
1035lib/CodeGen/CGCall.cpp at the moment; this is likely to change soon, though.
1036Generating assembly requires a suitable LLVM backend.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]1037
1038<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
1039<h3 id="target_os">Operating System Features and Limitations</h3>
1040<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
1041
1042<!-- ======================================= -->
1043<h4 id="target_os_darwin">Darwin (Mac OS/X)</h4>
1044<!-- ======================================= -->
1045
1046<p>No __thread support, 64-bit ObjC support requires SL tools.</p>
1047
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]1048<!-- ======================================= -->
1049<h4 id="target_os_win32">Windows</h4>
1050<!-- ======================================= -->
1051
1052<p>Experimental supports are on Cygming.</p>
1053
1054<h5>Cygwin</h5>
1055
1056<p>Clang works on Cygwin-1.7.</p>
1057
1058<h5>MinGW32</h5>
1059
1060<p>Clang works on some mingw32 distributions.
1061Clang assumes directories as below;</p>
1062
1063<ul>
1064<li><tt>C:/mingw/include</tt></li>
1065<li><tt>C:/mingw/lib</tt></li>
1066<li><tt>C:/mingw/lib/gcc/mingw32/4.[3-5].0/include/c++</tt></li>
1067</ul>
1068
1069<p>On MSYS, a few tests might fail. It is due to <a href="http://llvm.org/bugs/show_bug.cgi?id=8520">Bug 8520</a> and is fixed in <a href="http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20110314/118106.html">LLVM's r127724</a>.</p>
1070
1071<h5>MinGW-w64</h5>
1072
Douglas Gregordad879a2011-08-30 00:40:12 +0000[diff] [blame]1073<p>For 32-bit (i686-w64-mingw32), and 64-bit (x86_64-w64-mingw32), Clang assumes as below;<p>
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]1074
1075<ul>
Douglas Gregordad879a2011-08-30 00:40:12 +0000[diff] [blame]1076<li><tt>GCC versions 4.5.0 to 4.5.3, 4.6.0 to 4.6.2, or 4.7.0 (for the C++ header search path)</tt></li>
1077<li><tt>some_directory/bin/gcc.exe</tt></li>
1078<li><tt>some_directory/bin/clang.exe</tt></li>
1079<li><tt>some_directory/bin/clang++.exe</tt></li>
1080<li><tt>some_directory/bin/../include/c++/GCC_version</tt></li>
1081<li><tt>some_directory/bin/../include/c++/GCC_version/x86_64-w64-mingw32</tt></li>
1082<li><tt>some_directory/bin/../include/c++/GCC_version/i686-w64-mingw32</tt></li>
1083<li><tt>some_directory/bin/../include/c++/GCC_version/backward</tt></li>
1084<li><tt>some_directory/bin/../x86_64-w64-mingw32/include</tt></li>
1085<li><tt>some_directory/bin/../i686-w64-mingw32/include</tt></li>
1086<li><tt>some_directory/bin/../include</tt></li>
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]1087</ul>
1088
Douglas Gregordad879a2011-08-30 00:40:12 +0000[diff] [blame]1089<p>This directory layout is standard for any toolchain you will find on the official <a href="mingw-w64.sourceforge.net">MinGW-w64 website</a>.
1090
1091<p>Clang expects the GCC executable &quot;gcc.exe&quot; compiled for i686-w64-mingw32 (or x86_64-w64-mingw32) to be present on PATH.</p>
1092
1093<p><a href="http://llvm.org/bugs/show_bug.cgi?id=9072">Some tests might fail</a>
1094on x86_64-w64-mingw32.</p>
NAKAMURA Takumia75fdb22011-04-04 15:02:41 +0000[diff] [blame]1095
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]1096</div>
1097</body>
1098</html>