Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 57d12fd4a2bc739c4a4d62a364b7f08cd483c59e / . / docs / UsersManual.html
blob: c19c96b4bcf978da074e7498b49efb92aa8fd626 [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>
35 <li><a href="#diagnostics">Controlling Errors and Warnings</a></li>
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>
42 </ul>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]43 <li><a href="#precompiledheaders">Precompiled Headers</a></li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]44 <li><a href="#codegen">Controlling Code Generation</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]45 </ul>
46</li>
47<li><a href="#c">C Language Features</a>
48 <ul>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]49 <li><a href="#c_ext">Extensions supported by clang</a></li>
50 <li><a href="#c_modes">Differences between various standard modes</a></li>
51 <li><a href="#c_unimpl_gcc">GCC extensions not implemented yet</a></li>
52 <li><a href="#c_unsupp_gcc">Intentionally unsupported GCC extensions</a></li>
53 <li><a href="#c_ms">Microsoft extensions</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]54 </ul>
55</li>
56<li><a href="#objc">Objective-C Language Features</a>
57 <ul>
58 <li><a href="#objc_incompatibilities">Intentional Incompatibilities with
59 GCC</a></li>
60 </ul>
61</li>
62<li><a href="#cxx">C++ Language Features</a>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]63</li>
64<li><a href="#objcxx">Objective C++ Language Features</a>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]65</li>
66<li><a href="#target_features">Target-Specific Features and Limitations</a>
67 <ul>
68 <li><a href="#target_arch">CPU Architectures Features and Limitations</a>
69 <ul>
70 <li><a href="#target_arch_x86">X86</a></li>
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]71 <li><a href="#target_arch_arm">ARM</a></li>
72 <li><a href="#target_arch_other">Other platforms</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]73 </ul>
74 </li>
75 <li><a href="#target_os">Operating System Features and Limitations</a>
76 <ul>
77 <li><a href="#target_os_darwin">Darwin (Mac OS/X)</a></li>
78 <li>Linux, etc.</li>
79 </ul>
80
81 </li>
82 </ul>
83</li>
84</ul>
85
86
87<!-- ======================================================================= -->
88<h2 id="intro">Introduction</h2>
89<!-- ======================================================================= -->
90
91<p>The Clang Compiler is an open-source compiler for the C family of programming
92languages, aiming to be the best in class implementation of these languages.
93Clang builds on the LLVM optimizer and code generator, allowing it to provide
94high-quality optimization and code generation support for many targets. For
95more general information, please see the <a href="http://clang.llvm.org">Clang
96Web Site</a> or the <a href="http://llvm.org">LLVM Web Site</a>.</p>
97
98<p>This document describes important notes about using Clang as a compiler for
99an end-user, documenting the supported features, command line options, etc. If
100you are interested in using Clang to build a tool that processes code, please
101see <a href="InternalsManual.html">the Clang Internals Manual</a>. If you are
102interested in the <a href="http://clang.llvm.org/StaticAnalysis.html">Clang
103Static Analyzer</a>, please see its web page.</p>
104
105<p>Clang is designed to support the C family of programming languages, which
106includes <a href="#c">C</a>, <a href="#objc">Objective-C</a>, <a
107href="#cxx">C++</a>, and <a href="#objcxx">Objective-C++</a> as well as many
108dialects of those. For language-specific information, please see the
109corresponding language specific section:</p>
110
111<ul>
112<li><a href="#c">C Language</a>: K&amp;R C, ANSI C89, ISO C90, ISO C94
113 (C89+AMD1), ISO C99 (+TC1, TC2, TC3). </li>
114<li><a href="#objc">Objective-C Language</a>: ObjC 1, ObjC 2, ObjC 2.1, plus
115 variants depending on base language.</li>
116<li><a href="#cxx">C++ Language Features</a></li>
117<li><a href="#objcxx">Objective C++ Language</a></li>
118</ul>
119
120<p>In addition to these base languages and their dialects, Clang supports a
121broad variety of language extensions, which are documented in the corresponding
122language section. These extensions are provided to be compatible with the GCC,
123Microsoft, and other popular compilers as well as to improve functionality
124through Clang-specific features. The Clang driver and language features are
125intentionally designed to be as compatible with the GNU GCC compiler as
126reasonably possible, easing migration from GCC to Clang. In most cases, code
127"just works".</p>
128
129<p>In addition to language specific features, Clang has a variety of features
130that depend on what CPU architecture or operating system is being compiled for.
Douglas Gregorcd5a5052009-11-09 15:15:41 +0000[diff] [blame]131Please see the <a href="#target_features">Target-Specific Features and
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]132Limitations</a> section for more details.</p>
133
134<p>The rest of the introduction introduces some basic <a
135href="#terminology">compiler terminology</a> that is used throughout this manual
136and contains a basic <a href="#basicusage">introduction to using Clang</a>
137as a command line compiler.</p>
138
139<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
140<h3 id="terminology">Terminology</h3>
141<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
142
143<p>Front end, parser, backend, preprocessor, undefined behavior, diagnostic,
144 optimizer</p>
145
146<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
147<h3 id="basicusage">Basic Usage</h3>
148<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
149
150<p>Intro to how to use a C compiler for newbies.</p>
151<p>
152compile + link
153
154compile then link
155
156debug info
157
158enabling optimizations
159
160picking a language to use, defaults to C99 by default. Autosenses based on
161extension.
162
163using a makefile
164</p>
165
166
167<!-- ======================================================================= -->
168<h2 id="commandline">Command Line Options</h2>
169<!-- ======================================================================= -->
170
171<p>
172This section is generally an index into other sections. It does not go into
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]173depth on the ones that are covered by other sections. However, the first part
174introduces the language selection and other high level options like -c, -g, etc.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]175</p>
176
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]177
178<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
179<h3 id="cl_diagnostics">Options to Control Error and Warning Messages</h3>
180<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
181
182<p><b>-Werror</b>: Turn warnings into errors.</p>
183<p><b>-Werror=foo</b>: Turn warning "foo" into an error.</p>
184<p><b>-Wno-error=foo</b>: Turn warning "foo" into an warning even if -Werror is
185 specified.</p>
186<p><b>-Wfoo</b>: Enable warning foo</p>
187<p><b>-Wno-foo</b>: Disable warning foo</p>
188<p><b>-w</b>: Disable all warnings.</p>
189<p><b>-pedantic</b>: Warn on language extensions.</p>
190<p><b>-pedantic-errors</b>: Error on language extensions.</p>
191<p><b>-Wsystem-headers</b>: Enable warnings from system headers.</p>
192
Chris Lattner0f0c9632010-04-07 20:49:23 +0000[diff] [blame]193<p><b>-ferror-limit=123</b>: Stop emitting diagnostics after 123 errors have
194 been produced. The default is 20, and the error limit can be disabled with
195 -ferror-limit=0.</p>
196
Douglas Gregor575cf372010-04-20 07:18:24 +0000[diff] [blame]197<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>
198
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]199<!-- ================================================= -->
200<h4 id="cl_diag_formatting">Formatting of Diagnostics</h4>
201<!-- ================================================= -->
202
203<p>Clang aims to produce beautiful diagnostics by default, particularly for new
204users that first come to Clang. However, different people have different
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]205preferences, and sometimes Clang is driven by another program that wants to
206parse simple and consistent output, not a person. For these cases, Clang
207provides a wide range of options to control the exact output format of the
208diagnostics that it generates.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]209
210<dl>
211
212<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
213<dt id="opt_fshow-column"><b>-f[no-]show-column</b>: Print column number in
214diagnostic.</dt>
215<dd>This option, which defaults to on, controls whether or not Clang prints the
216column number of a diagnostic. For example, when this is enabled, Clang will
217print something like:</p>
218
219<pre>
220 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
221 #endif bad
222 ^
223 //
224</pre>
225
226<p>When this is disabled, Clang will print "test.c:28: warning..." with no
227column number.</p>
228</dd>
229
230<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
231<dt id="opt_fshow-source-location"><b>-f[no-]show-source-location</b>: Print
232source file/line/column information in diagnostic.</dt>
233<dd>This option, which defaults to on, controls whether or not Clang prints the
234filename, line number and column number of a diagnostic. For example,
235when this is enabled, Clang will print something like:</p>
236
237<pre>
238 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
239 #endif bad
240 ^
241 //
242</pre>
243
244<p>When this is disabled, Clang will not print the "test.c:28:8: " part.</p>
245</dd>
246
247<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
248<dt id="opt_fcaret-diagnostics"><b>-f[no-]caret-diagnostics</b>: Print source
249line and ranges from source code in diagnostic.</dt>
250<dd>This option, which defaults to on, controls whether or not Clang prints the
251source line, source ranges, and caret when emitting a diagnostic. For example,
252when this is enabled, Clang will print something like:</p>
253
254<pre>
255 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
256 #endif bad
257 ^
258 //
259</pre>
260
261<p>When this is disabled, Clang will just print:</p>
262
263<pre>
264 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
265</pre>
266
267</dd>
268
269<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
270<dt id="opt_fdiagnostics-show-option"><b>-f[no-]diagnostics-show-option</b>:
271Enable <tt>[-Woption]</tt> information in diagnostic line.</dt>
272<dd>This option, which defaults to on,
273controls whether or not Clang prints the associated <A
274href="#cl_diag_warning_groups">warning group</a> option name when outputting
275a warning diagnostic. For example, in this output:</p>
276
277<pre>
278 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
279 #endif bad
280 ^
281 //
282</pre>
283
284<p>Passing <b>-fno-diagnostics-show-option</b> will prevent Clang from printing
285the [<a href="#opt_Wextra-tokens">-Wextra-tokens</a>] information in the
286diagnostic. This information tells you the flag needed to enable or disable the
287diagnostic, either from the command line or through <a
288href="#pragma_GCC_diagnostic">#pragma GCC diagnostic</a>.</dd>
289
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]290<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
291<dt id="opt_fdiagnostics-show-category"><b>-fdiagnostics-show-category=none/id/name</b>:
292Enable printing category information in diagnostic line.</dt>
293<dd>This option, which defaults to "none",
294controls whether or not Clang prints the category associated with a diagnostic
295when emitting it. Each diagnostic may or many not have an associated category,
296if it has one, it is listed in the diagnostic categorization field of the
297diagnostic line (in the []'s).</p>
298
299<p>For example, a format string warning will produce these three renditions
300based on the setting of this option:</p>
301
302<pre>
303 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat]
304 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,1</b>]
305 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,Format String</b>]
306</pre>
307
308<p>This category can be used by clients that want to group diagnostics by
309category, so it should be a high level category. We want dozens of these, not
310hundreds or thousands of them.</p>
311</dd>
312
313
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]314
315<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
316<dt id="opt_fdiagnostics-fixit-info"><b>-f[no-]diagnostics-fixit-info</b>:
317Enable "FixIt" information in the diagnostics output.</dt>
318<dd>This option, which defaults to on, controls whether or not Clang prints the
319information on how to fix a specific diagnostic underneath it when it knows.
320For example, in this output:</p>
321
322<pre>
323 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
324 #endif bad
325 ^
326 //
327</pre>
328
329<p>Passing <b>-fno-diagnostics-fixit-info</b> will prevent Clang from printing
330the "//" line at the end of the message. This information is useful for users
331who may not understand what is wrong, but can be confusing for machine
332parsing.</p>
333</dd>
334
335<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]336<dt id="opt_fdiagnostics-print-source-range-info">
337<b>-f[no-]diagnostics-print-source-range-info</b>:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]338Print machine parsable information about source ranges.</dt>
339<dd>This option, which defaults to off, controls whether or not Clang prints
340information about source ranges in a machine parsable format after the
341file/line/column number information. The information is a simple sequence of
342brace enclosed ranges, where each range lists the start and end line/column
343locations. For example, in this output:</p>
344
345<pre>
346exprs.c:47:15:{47:8-47:14}{47:17-47:24}: error: invalid operands to binary expression ('int *' and '_Complex float')
347 P = (P-42) + Gamma*4;
348 ~~~~~~ ^ ~~~~~~~
349</pre>
350
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]351<p>The {}'s are generated by -fdiagnostics-print-source-range-info.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]352</dd>
353
354
355</dl>
356
357
358
359
360<!-- ===================================================== -->
361<h4 id="cl_diag_warning_groups">Individual Warning Groups</h4>
362<!-- ===================================================== -->
363
364<p>TODO: Generate this from tblgen. Define one anchor per warning group.</p>
365
366
367<dl>
368
369
370<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
371<dt id="opt_Wextra-tokens"><b>-Wextra-tokens</b>: Warn about excess tokens at
372 the end of a preprocessor directive.</dt>
373<dd>This option, which defaults to on, enables warnings about extra tokens at
374the end of preprocessor directives. For example:</p>
375
376<pre>
377 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
378 #endif bad
379 ^
380</pre>
381
382<p>These extra tokens are not strictly conforming, and are usually best handled
383by commenting them out.</p>
384
385<p>This option is also enabled by <a href="">-Wfoo</a>, <a href="">-Wbar</a>,
386 and <a href="">-Wbaz</a>.</p>
387</dd>
388
Jeffrey Yasskin21d07e42010-06-05 01:39:57 +0000[diff] [blame]389<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
390<dt id="opt_Wambiguous-member-template"><b>-Wambiguous-member-template</b>:
391Warn about unqualified uses of a member template whose name resolves
392to another template at the location of the use.</dt>
393<dd>This option, which defaults to on, enables a warning in the
394following code:</p>
395
396<pre>
397template&lt;typename T> struct set{};
398template&lt;typename T> struct trait { typedef const T& type; };
399struct Value {
400 template&lt;typename T> void set(typename trait&lt;T>::type value) {}
401};
402void foo() {
403 Value v;
404 v.set&lt;double>(3.2);
405}
406</pre>
407
408<p>C++ [basic.lookup.classref] requires this to be an error, but,
409because it's hard to work around, Clang downgrades it to a warning as
410an extension.</p>
411</dd>
412
Jeffrey Yasskin57d12fd2010-06-07 15:58:05 +0000[diff] [blame^]413<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
414<dt id="opt_Wbind-to-temporary-copy"><b>-Wbind-to-temporary-copy</b>: Warn about
415an unusable copy constructor when binding a reference to a temporary.</dt>
416<dd>This option, which defaults to on, enables warnings about binding a
417reference to a temporary when the temporary doesn't have a usable copy
418constructor. For example:</p>
419
420<pre>
421 struct NonCopyable {
422 NonCopyable();
423 private:
424 NonCopyable(const NonCopyable&);
425 };
426 void foo(const NonCopyable&);
427 void bar() {
428 foo(NonCopyable()); // Disallowed in C++98; allowed in C++0x.
429 }
430</pre>
431<pre>
432 struct NonCopyable2 {
433 NonCopyable2();
434 NonCopyable2(NonCopyable2&);
435 };
436 void foo(const NonCopyable2&);
437 void bar() {
438 foo(NonCopyable2()); // Disallowed in C++98; allowed in C++0x.
439 }
440</pre>
441
442<p>Note that if <tt>NonCopyable2::NonCopyable2()</tt> has a default
443argument whose instantiation produces a compile error, that error will
444still be a hard error in C++98 mode even if this warning is turned
445off.</p>
446
447</dd>
448
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]449</dl>
450
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]451<!-- ======================================================================= -->
452<h2 id="general_features">Language and Target-Independent Features</h2>
453<!-- ======================================================================= -->
454
455
456<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
457<h3 id="diagnostics">Controlling Errors and Warnings</h3>
458<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
459
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]460<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]461emit errors and warning messages, and how they are displayed to the console.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]462
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]463<h4 id="diagnostics_display">Controlling How Clang Displays Diagnostics</h4>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]464
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]465<p>When Clang emits a diagnostic, it includes rich information in the output,
466and gives you fine-grain control over which information is printed. Clang has
467the ability to print this information, and these are the options that control
468it:</p>
469
470<p>
471<ol>
472<li>A file/line/column indicator that shows exactly where the diagnostic occurs
473 in your code [<a href="#opt_fshow-column">-fshow-column</a>, <a
474 href="#opt_fshow-source-location">-fshow-source-location</a>].</li>
475<li>A categorization of the diagnostic as a note, warning, error, or fatal
476 error.</li>
477<li>A text string that describes what the problem is.</li>
478<li>An option that indicates how to control the diagnostic (for diagnostics that
479 support it) [<a
480 href="#opt_fdiagnostics-show-option">-fdiagnostics-show-option</a>].</li>
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]481<li>A <a href="#diagnostics_categories">high-level category</a> for the
482 diagnostic for clients that want to group diagnostics by class (for
483 diagnostics that support it) [<a
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]484 href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]485<li>The line of source code that the issue occurs on, along with a caret and
486 ranges that indicate the important locations [<a
487 href="opt_fcaret-diagnostics">-fcaret-diagnostics</a>].</li>
488<li>"FixIt" information, which is a concise explanation of how to fix the
489 problem (when Clang is certain it knows) [<a
490 href="opt_fdiagnostics-fixit-info">-fdiagnostics-fixit-info</a>].</li>
491<li>A machine-parsable representation of the ranges involved (off by
492 default) [<a
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]493 href="opt_fdiagnostics-print-source-range-info">-fdiagnostics-print-source-range-info</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]494</ol></p>
495
496<p>For more information please see <a href="#cl_diag_formatting">Formatting of
497Diagnostics</a>.</p>
498
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]499
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]500<h4 id="diagnostics_mappings">Diagnostic Mappings</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]501
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]502<p>All diagnostics are mapped into one of these 5 classes:</p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]503
504<p>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]505<ul>
506<li>Ignored</li>
507<li>Note</li>
508<li>Warning</li>
509<li>Error</li>
510<li>Fatal</li>
511</ul></p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]512
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]513<h4 id="diagnostics_categories">Diagnostic Categories</h4>
514
515<p>Though not shown by default, diagnostics may each be associated with a
516 high-level category. This category is intended to make it possible to triage
517 builds that produce a large number of errors or warnings in a grouped way.
518</p>
519
520<p>Categories are not shown by default, but they can be turned on with the
521<a href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a> option.
522When set to "<tt>name</tt>", the category is printed textually in the diagnostic
523output. When it is set to "<tt>id</tt>", a category number is printed. The
524mapping of category names to category id's can be obtained by running '<tt>clang
525 --print-diagnostic-categories</tt>'.
526</p>
527
528<h4 id="diagnostics_commandline">Controlling Diagnostics via Command Line
529 Flags</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]530
531<p>-W flags, -pedantic, etc</p>
532
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]533<h4 id="diagnostics_pragmas">Controlling Diagnostics via Pragmas</h4>
534
535<p>Clang can also control what diagnostics are enabled through the use of
536pragmas in the source code. This is useful for turning off specific warnings
537in a section of source code. Clang supports GCC's pragma for compatibility
538with existing source code, as well as several extensions. </p>
539
540<p>The pragma may control any warning that can be used from the command line.
541Warnings may be set to ignored, warning, error, or fatal. The following
542example code will tell Clang or GCC to ignore the -Wall warnings:</p>
543
544<pre>
545#pragma GCC diagnostic ignored "-Wall"
546</pre>
547
548<p>In addition to all of the functionality of provided by GCC's pragma, Clang
549also allows you to push and pop the current warning state. This is particularly
550useful when writing a header file that will be compiled by other people, because
551you don't know what warning flags they build with.</p>
552
553<p>In the below example
554-Wmultichar is ignored for only a single line of code, after which the
555diagnostics return to whatever state had previously existed.</p>
556
557<pre>
558#pragma clang diagnostic push
559#pragma clang diagnostic ignored "-Wmultichar"
560
561char b = 'df'; // no warning.
562
563#pragma clang diagnostic pop
564</pre>
565
566<p>The push and pop pragmas will save and restore the full diagnostic state of
567the compiler, regardless of how it was set. That means that it is possible to
568use push and pop around GCC compatible diagnostics and Clang will push and pop
569them appropriately, while GCC will ignore the pushes and pops as unknown
570pragmas. It should be noted that while Clang supports the GCC pragma, Clang and
571GCC do not support the exact same set of warnings, so even when using GCC
572compatible #pragmas there is no guarantee that they will have identical behaviour
573on both compilers. </p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]574
575<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
576<h3 id="precompiledheaders">Precompiled Headers</h3>
577<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
578
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]579<p><a href="http://en.wikipedia.org/wiki/Precompiled_header">Precompiled
580headers</a> are a general approach employed by many compilers to reduce
581compilation time. The underlying motivation of the approach is that it is
582common for the same (and often large) header files to be included by
583multiple source files. Consequently, compile times can often be greatly improved
584by caching some of the (redundant) work done by a compiler to process headers.
585Precompiled header files, which represent one of many ways to implement
586this optimization, are literally files that represent an on-disk cache that
587contains the vital information necessary to reduce some of the work
588needed to process a corresponding header file. While details of precompiled
589headers vary between compilers, precompiled headers have been shown to be a
590highly effective at speeding up program compilation on systems with very large
591system headers (e.g., Mac OS/X).</p>
592
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]593<h4>Generating a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]594
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]595<p>To generate a PCH file using Clang, one invokes Clang with
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]596the <b><tt>-x <i>&lt;language&gt;</i>-header</tt></b> option. This mirrors the
597interface in GCC for generating PCH files:</p>
598
599<pre>
600 $ gcc -x c-header test.h -o test.h.gch
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]601 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]602</pre>
603
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]604<h4>Using a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]605
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]606<p>A PCH file can then be used as a prefix header when a
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]607<b><tt>-include</tt></b> option is passed to <tt>clang</tt>:</p>
608
609<pre>
610 $ clang -include test.h test.c -o test
611</pre>
612
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]613<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]614is 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]615will be processed from the PCH file. Otherwise, Clang falls back to
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]616directly processing the content of <tt>test.h</tt>. This mirrors the behavior of
617GCC.</p>
618
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]619<p><b>NOTE:</b> Clang does <em>not</em> automatically use PCH files
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]620for headers that are directly included within a source file. For example:</p>
621
622<pre>
Chris Lattnere42ec542009-06-13 20:35:58 +0000[diff] [blame]623 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]624 $ cat test.c
625 #include "test.h"
626 $ clang test.c -o test
627</pre>
628
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]629<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]630<tt>test.h</tt> since <tt>test.h</tt> was included directly in the source file
631and not specified on the command line using <tt>-include</tt>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]632
Douglas Gregore650c8c2009-07-07 00:12:59 +0000[diff] [blame]633<h4>Relocatable PCH Files</h4>
634<p>It is sometimes necessary to build a precompiled header from headers that
635are not yet in their final, installed locations. For example, one might build a
636precompiled header within the build tree that is then meant to be installed
637alongside the headers. Clang permits the creation of "relocatable" precompiled
638headers, which are built with a given path (into the build directory) and can
639later be used from an installed location.</p>
640
641<p>To build a relocatable precompiled header, place your headers into a
642subdirectory whose structure mimics the installed location. For example, if you
643want to build a precompiled header for the header <code>mylib.h</code> that
644will be installed into <code>/usr/include</code>, create a subdirectory
645<code>build/usr/include</code> and place the header <code>mylib.h</code> into
646that subdirectory. If <code>mylib.h</code> depends on other headers, then
647they can be stored within <code>build/usr/include</code> in a way that mimics
648the installed location.</p>
649
650<p>Building a relocatable precompiled header requires two additional arguments.
651First, pass the <code>--relocatable-pch</code> flag to indicate that the
652resulting PCH file should be relocatable. Second, pass
653<code>-isysroot /path/to/build</code>, which makes all includes for your
654library relative to the build directory. For example:</p>
655
656<pre>
657 # clang -x c-header --relocatable-pch -isysroot /path/to/build /path/to/build/mylib.h mylib.h.pch
658</pre>
659
660<p>When loading the relocatable PCH file, the various headers used in the PCH
661file are found from the system header root. For example, <code>mylib.h</code>
662can be found in <code>/usr/include/mylib.h</code>. If the headers are installed
663in some other system root, the <code>-isysroot</code> option can be used provide
664a different system root from which the headers will be based. For example,
665<code>-isysroot /Developer/SDKs/MacOSX10.4u.sdk</code> will look for
666<code>mylib.h</code> in
667<code>/Developer/SDKs/MacOSX10.4u.sdk/usr/include/mylib.h</code>.</p>
668
669<p>Relocatable precompiled headers are intended to be used in a limited number
670of cases where the compilation environment is tightly controlled and the
671precompiled header cannot be generated after headers have been installed.
672Relocatable precompiled headers also have some performance impact, because
673the difference in location between the header locations at PCH build time vs.
674at the time of PCH use requires one of the PCH optimizations,
675<code>stat()</code> caching, to be disabled. However, this change is only
676likely to affect PCH files that reference a large number of headers.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]677
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]678<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
679<h3 id="codegen">Controlling Code Generation</h3>
680<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
681
682<p>Clang provides a number of ways to control code generation. The options are listed below.</p>
683
684<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
685<dt id="opt_fcatch-undefined-behavior"><b>-fcatch-undefined-behavior</b>: Turn
686on runtime code generation to check for undefined behavior.</dt>
687
688<dd>This option, which defaults to off, controls whether or not Clang
689adds runtime checks for undefined runtime behavior. If the check fails,
690<tt>__builtin_trap()</tt> is used to indicate failure.
691The checks are:
692<p>
Mike Stump88b2a172009-12-16 03:25:12 +0000[diff] [blame]693<li>Subscripting where the static type of one operand is variable
694 which is decayed from an array type and the other operand is
695 greater than the size of the array or less than zero.</li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]696<li>Shift operators where the amount shifted is greater or equal to the
697 promoted bit-width of the left-hand-side or less than zero.</li>
Mike Stump8f6a3ed2009-12-16 03:18:14 +0000[diff] [blame]698<li>If control flow reaches __builtin_unreachable.
699<li>When llvm implements more __builtin_object_size support, reads and
700 writes for objects that __builtin_object_size indicates we aren't
701 accessing valid memory. Bit-fields and vectors are not yet checked.
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]702</p>
703</dd>
704
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]705<dt id="opt_fno-assume-sane-operator-new"><b>-fno-assume-sane-operator-new</b>:
706Don't assume that the C++'s new operator is sane.</dt>
Nuno Lopesb23f20d2009-12-17 10:15:49 +0000[diff] [blame]707<dd>This option tells the compiler to do not assume that C++'s global new
708operator will always return a pointer that do not
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]709alias any other pointer when the function returns.</dd>
710
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]711<!-- ======================================================================= -->
712<h2 id="c">C Language Features</h2>
713<!-- ======================================================================= -->
714
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]715<p>The support for standard C in clang is feature-complete except for the C99
716floating-point pragmas.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]717
718<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]719<h3 id="c_ext">Extensions supported by clang</h3>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]720<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
721
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]722<p>See <a href="LanguageExtensions.html">clang language extensions</a>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]723
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]724<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
725<h3 id="c_modes">Differences between various standard modes</h3>
726<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]727
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]728<p>clang supports the -std option, which changes what language mode clang uses.
729The supported modes for C are c89, gnu89, c94, c99, gnu99 and various aliases
730for those modes. If no -std option is specified, clang defaults to gnu99 mode.
731</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]732
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]733<p>Differences between all c* and gnu* modes:</p>
734<ul>
735<li>c* modes define "__STRICT_ANSI__".</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]736<li>Target-specific defines not prefixed by underscores, like "linux", are
737defined in gnu* modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]738<li>Trigraphs default to being off in gnu* modes; they can be enabled by the
739-trigraphs option.</li>
740<li>The parser recognizes "asm" and "typeof" as keywords in gnu* modes; the
741variants "__asm__" and "__typeof__" are recognized in all modes.</li>
Eli Friedmanb0ac5452009-05-16 23:17:30 +0000[diff] [blame]742<li>The Apple "blocks" extension is recognized by default in gnu* modes
743on some platforms; it can be enabled in any mode with the "-fblocks"
744option.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]745</ul>
746
747<p>Differences between *89 and *99 modes:</p>
748<ul>
749<li>The *99 modes default to implementing "inline" as specified in C99, while
750the *89 modes implement the GNU version. This can be overridden for individual
751functions with the __gnu_inline__ attribute.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]752<li>Digraphs are not recognized in c89 mode.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]753<li>The scope of names defined inside a "for", "if", "switch", "while", or "do"
754statement is different. (example: "if ((struct x {int x;}*)0) {}".)</li>
755<li>__STDC_VERSION__ is not defined in *89 modes.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]756<li>"inline" is not recognized as a keyword in c89 mode.</li>
757<li>"restrict" is not recognized as a keyword in *89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]758<li>Commas are allowed in integer constant expressions in *99 modes.</li>
759<li>Arrays which are not lvalues are not implicitly promoted to pointers in
760*89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]761<li>Some warnings are different.</li>
762</ul>
763
764<p>c94 mode is identical to c89 mode except that digraphs are enabled in
765c94 mode (FIXME: And __STDC_VERSION__ should be defined!).</p>
766
767<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
768<h3 id="c_unimpl_gcc">GCC extensions not implemented yet</h3>
769<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
770
771<p>clang tries to be compatible with gcc as much as possible, but some gcc
772extensions are not implemented yet:</p>
773
774<ul>
775<li>clang does not support __label__
776(<a href="http://llvm.org/bugs/show_bug.cgi?id=3429">bug 3429</a>). This is
777a relatively small feature, so it is likely to be implemented relatively
778soon.</li>
779
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]780<li>clang does not support #pragma weak
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]781(<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]782the uses described in the bug, this is likely to be implemented at some
783point, at least partially.</li>
784
785<li>clang does not support #pragma align
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]786(<a href="http://llvm.org/bugs/show_bug.cgi?id=3811">bug 3811</a>). This is a
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]787relatively small feature, so it is likely to be implemented relatively
788soon.</li>
789
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]790<li>clang does not support code generation for local variables pinned to
791registers (<a href="http://llvm.org/bugs/show_bug.cgi?id=3933">bug 3933</a>).
792This is a relatively small feature, so it is likely to be implemented
793relatively soon.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]794
795<li>clang does not support decimal floating point types (_Decimal32 and
796friends) or fixed-point types (_Fract and friends); nobody has expressed
797interest in these features yet, so it's hard to say when they will be
798implemented.</li>
799
800<li>clang does not support nested functions; this is a complex feature which
801is infrequently used, so it is unlikely to be implemented anytime soon.</li>
802
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]803<li>clang does not support global register variables, this is unlikely
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]804to be implemented soon because it requires additional LLVM backend support.
805</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]806
807<li>clang does not support static initialization of flexible array
808members. This appears to be a rarely used extension, but could be
809implemented pending user demand.</li>
810
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]811<li>clang does not support __builtin_va_arg_pack/__builtin_va_arg_pack_len.
812This is used rarely, but in some potentially interesting places, like the
813glibc headers, so it may be implemented pending user demand. Note that
814because clang pretends to be like GCC 4.2, and this extension was introduced
815in 4.3, the glibc headers will not try to use this extension with clang at
816the moment.</li>
817
818<li>clang does not support the gcc extension for forward-declaring function
819parameters; this has not showed up in any real-world code yet, though, so it
820might never be implemented.</li>
821
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]822</ul>
823
824<p>This is not a complete list; if you find an unsupported extension
825missing from this list, please send an e-mail to cfe-dev. This list
826currently excludes C++; see <a href="#cxx">C++ Language Features</a>.
827Also, this list does not include bugs in mostly-implemented features; please
828see 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">
829bug tracker</a> for known existing bugs (FIXME: Is there a section for
830bug-reporting guidelines somewhere?).</p>
831
832<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
833<h3 id="c_unsupp_gcc">Intentionally unsupported GCC extensions</h3>
834<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
835
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]836<ul>
837
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]838<li>clang does not support the gcc extension that allows variable-length arrays
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]839in structures. This is for a few of reasons: one, it is tricky
840to implement, two, the extension is completely undocumented, and three, the
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]841extension appears to be rarely used.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]842
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]843<li>clang does not support duplicate definitions of a function where one is
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]844inline. This complicates clients of the AST which normally can expect there is
845at most one definition for each function. Source code using this feature should
846be changed to define the inline and out-of-line definitions in separate
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]847translation units.</li>
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]848
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]849<li>clang does not have an equivalent to gcc's "fold"; this means that
850clang doesn't accept some constructs gcc might accept in contexts where a
851constant expression is required, like "x-x" where x is a variable, or calls
852to C library functions like strlen.</li>
853
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]854<li>clang does not support multiple alternative constraints in inline asm; this
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]855is an extremely obscure feature which would be complicated to implement
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]856correctly.</li>
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]857
858<li>clang does not support __builtin_apply and friends; this extension is
859extremely obscure and difficult to implement reliably.</li>
860
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]861</ul>
862
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]863<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
864<h3 id="c_ms">Microsoft extensions</h3>
865<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
866
867<p>clang has some experimental support for extensions from
868Microsoft Visual C++; to enable it, use the -fms-extensions command-line
Eli Friedmana1821b52009-06-08 06:21:03 +0000[diff] [blame]869option. This is the default for Windows targets. Note that the
870support is incomplete; enabling Microsoft extensions will silently drop
871certain constructs (including __declspec and Microsoft-style asm statements).
872</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]873
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]874<li>clang does not support the Microsoft extension where anonymous
Daniel Dunbar869e0542009-05-13 00:23:53 +0000[diff] [blame]875record members can be declared using user defined typedefs.</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]876
Daniel Dunbar9375ed12009-05-13 21:40:49 +0000[diff] [blame]877<li>clang supports the Microsoft "#pragma pack" feature for
878controlling record layout. GCC also contains support for this feature,
879however where MSVC and GCC are incompatible clang follows the MSVC
880definition.</li>
881
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]882<!-- ======================================================================= -->
883<h2 id="objc">Objective-C Language Features</h2>
884<!-- ======================================================================= -->
885
886
887<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
888<h3 id="objc_incompatibilities">Intentional Incompatibilities with GCC</h3>
889<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
890
891<p>No cast of super, no lvalue casts.</p>
892
893
894
895<!-- ======================================================================= -->
896<h2 id="cxx">C++ Language Features</h2>
897<!-- ======================================================================= -->
898
Douglas Gregor57f1a002010-02-05 23:51:14 +0000[diff] [blame]899<p>At this point, Clang C++ is not production-quality and is not recommended for use beyond experimentation. However, Clang C++ support
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]900is under active development and is progressing rapidly. Please see the <a
901href="http://clang.llvm.org/cxx_status.html">C++ Status</a> page for details or
902ask on the mailing list about how you can help.</p>
903
Douglas Gregor57f1a002010-02-05 23:51:14 +0000[diff] [blame]904<p>Note that released Clang compilers will refuse to even try to use clang to compile C++ code unless you pass the <tt>-ccc-clang-cxx</tt> option to the driver. To turn on Clang's C++ support, please pass that flag. Clang compilers built from the Subversion trunk enable C++ support by default, and do not require the <tt>-ccc-clang-cxx</tt> flag.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]905
Rafael Espindola9b2fc952010-03-17 04:31:53 +0000[diff] [blame]906<p>Clang strives to strictly conform to the C++ standard. That means
907it will reject invalid C++ code that another compiler may accept. If
908Clang reports errors in your code, please check
909the <a href="http://clang.llvm.org/cxx_compatibility.html">C++
910Compatibility</a> page to see whether they are C++-conformance bugs
911and how you can fix them.</p>
912
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]913<!-- ======================================================================= -->
914<h2 id="objcxx">Objective C++ Language Features</h2>
915<!-- ======================================================================= -->
916
917<p>At this point, Clang C++ support is not generally useful (and therefore,
918neither is Objective-C++). Please see the <a href="#cxx">C++ section</a> for
919more information.</p>
920
921<!-- ======================================================================= -->
922<h2 id="target_features">Target-Specific Features and Limitations</h2>
923<!-- ======================================================================= -->
924
925
926<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
927<h3 id="target_arch">CPU Architectures Features and Limitations</h3>
928<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
929
930<!-- ======================== -->
931<h4 id="target_arch_x86">X86</h4>
932<!-- ======================== -->
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]933<p>The support for X86 (both 32-bit and 64-bit) is considered stable
934on Darwin (Mac OS/X), Linux, FreeBSD, and Dragonfly BSD: it has been tested to
935correctly compile large C and Objective-C codebases. (FIXME: Anything specific
936we want to say here? Possibly mention some LLVM x86 limitations?)
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]937
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]938<!-- ======================== -->
939<h4 id="target_arch_arm">ARM</h4>
940<!-- ======================== -->
941ARM support is mostly feature-complete, but still experimental; it hasn't
942undergone significant testing.
943
944<!-- ======================== -->
945<h4 id="target_arch_other">Other platforms</h4>
946<!-- ======================== -->
947clang currently contains some support for PPC and Sparc; however, significant
948pieces of code generation are still missing, and they haven't undergone
949significant testing.
950
951<p>clang contains some support for the embedded PIC16 processor
952(FIXME: I haven't been keeping track of this; what should this say?).
953
954<p>clang contains limited support for the MSP430 embedded processor, but both
955the clang support and the LLVM backend support are highly experimental.
956
957<p>Other platforms are completely unsupported at the moment. Adding the
958minimal support needed for parsing and semantic analysis on a new platform
959is quite easy; see lib/Basic/Targets.cpp in the clang source tree. This level
960of support is also sufficient for conversion to LLVM IR for simple programs.
961Proper support for conversion to LLVM IR requires adding code to
962lib/CodeGen/CGCall.cpp at the moment; this is likely to change soon, though.
963Generating assembly requires a suitable LLVM backend.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]964
965<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
966<h3 id="target_os">Operating System Features and Limitations</h3>
967<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
968
969<!-- ======================================= -->
970<h4 id="target_os_darwin">Darwin (Mac OS/X)</h4>
971<!-- ======================================= -->
972
973<p>No __thread support, 64-bit ObjC support requires SL tools.</p>
974
975</div>
976</body>
977</html>