Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 4786c15f4977c7cee98fde3ebdee213dba23848b / . / docs / UsersManual.html
blob: 1fcd59e26e31b63ee3d0af8956401bc47d3d158e [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>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]56<li><a href="#target_features">Target-Specific Features and Limitations</a>
57 <ul>
58 <li><a href="#target_arch">CPU Architectures Features and Limitations</a>
59 <ul>
60 <li><a href="#target_arch_x86">X86</a></li>
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]61 <li><a href="#target_arch_arm">ARM</a></li>
62 <li><a href="#target_arch_other">Other platforms</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]63 </ul>
64 </li>
65 <li><a href="#target_os">Operating System Features and Limitations</a>
66 <ul>
67 <li><a href="#target_os_darwin">Darwin (Mac OS/X)</a></li>
68 <li>Linux, etc.</li>
69 </ul>
70
71 </li>
72 </ul>
73</li>
74</ul>
75
76
77<!-- ======================================================================= -->
78<h2 id="intro">Introduction</h2>
79<!-- ======================================================================= -->
80
81<p>The Clang Compiler is an open-source compiler for the C family of programming
82languages, aiming to be the best in class implementation of these languages.
83Clang builds on the LLVM optimizer and code generator, allowing it to provide
84high-quality optimization and code generation support for many targets. For
85more general information, please see the <a href="http://clang.llvm.org">Clang
86Web Site</a> or the <a href="http://llvm.org">LLVM Web Site</a>.</p>
87
88<p>This document describes important notes about using Clang as a compiler for
89an end-user, documenting the supported features, command line options, etc. If
90you are interested in using Clang to build a tool that processes code, please
91see <a href="InternalsManual.html">the Clang Internals Manual</a>. If you are
92interested in the <a href="http://clang.llvm.org/StaticAnalysis.html">Clang
93Static Analyzer</a>, please see its web page.</p>
94
95<p>Clang is designed to support the C family of programming languages, which
96includes <a href="#c">C</a>, <a href="#objc">Objective-C</a>, <a
97href="#cxx">C++</a>, and <a href="#objcxx">Objective-C++</a> as well as many
98dialects of those. For language-specific information, please see the
99corresponding language specific section:</p>
100
101<ul>
102<li><a href="#c">C Language</a>: K&amp;R C, ANSI C89, ISO C90, ISO C94
103 (C89+AMD1), ISO C99 (+TC1, TC2, TC3). </li>
104<li><a href="#objc">Objective-C Language</a>: ObjC 1, ObjC 2, ObjC 2.1, plus
105 variants depending on base language.</li>
106<li><a href="#cxx">C++ Language Features</a></li>
107<li><a href="#objcxx">Objective C++ Language</a></li>
108</ul>
109
110<p>In addition to these base languages and their dialects, Clang supports a
111broad variety of language extensions, which are documented in the corresponding
112language section. These extensions are provided to be compatible with the GCC,
113Microsoft, and other popular compilers as well as to improve functionality
114through Clang-specific features. The Clang driver and language features are
115intentionally designed to be as compatible with the GNU GCC compiler as
116reasonably possible, easing migration from GCC to Clang. In most cases, code
117"just works".</p>
118
119<p>In addition to language specific features, Clang has a variety of features
120that depend on what CPU architecture or operating system is being compiled for.
Douglas Gregorcd5a5052009-11-09 15:15:41 +0000[diff] [blame]121Please see the <a href="#target_features">Target-Specific Features and
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]122Limitations</a> section for more details.</p>
123
124<p>The rest of the introduction introduces some basic <a
125href="#terminology">compiler terminology</a> that is used throughout this manual
126and contains a basic <a href="#basicusage">introduction to using Clang</a>
127as a command line compiler.</p>
128
129<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
130<h3 id="terminology">Terminology</h3>
131<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
132
133<p>Front end, parser, backend, preprocessor, undefined behavior, diagnostic,
134 optimizer</p>
135
136<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
137<h3 id="basicusage">Basic Usage</h3>
138<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
139
140<p>Intro to how to use a C compiler for newbies.</p>
141<p>
142compile + link
143
144compile then link
145
146debug info
147
148enabling optimizations
149
150picking a language to use, defaults to C99 by default. Autosenses based on
151extension.
152
153using a makefile
154</p>
155
156
157<!-- ======================================================================= -->
158<h2 id="commandline">Command Line Options</h2>
159<!-- ======================================================================= -->
160
161<p>
162This section is generally an index into other sections. It does not go into
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]163depth on the ones that are covered by other sections. However, the first part
164introduces the language selection and other high level options like -c, -g, etc.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]165</p>
166
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]167
168<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
169<h3 id="cl_diagnostics">Options to Control Error and Warning Messages</h3>
170<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
171
172<p><b>-Werror</b>: Turn warnings into errors.</p>
173<p><b>-Werror=foo</b>: Turn warning "foo" into an error.</p>
174<p><b>-Wno-error=foo</b>: Turn warning "foo" into an warning even if -Werror is
175 specified.</p>
176<p><b>-Wfoo</b>: Enable warning foo</p>
177<p><b>-Wno-foo</b>: Disable warning foo</p>
178<p><b>-w</b>: Disable all warnings.</p>
179<p><b>-pedantic</b>: Warn on language extensions.</p>
180<p><b>-pedantic-errors</b>: Error on language extensions.</p>
181<p><b>-Wsystem-headers</b>: Enable warnings from system headers.</p>
182
Chris Lattner0f0c9632010-04-07 20:49:23 +0000[diff] [blame]183<p><b>-ferror-limit=123</b>: Stop emitting diagnostics after 123 errors have
184 been produced. The default is 20, and the error limit can be disabled with
185 -ferror-limit=0.</p>
186
Douglas Gregor575cf372010-04-20 07:18:24 +0000[diff] [blame]187<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>
188
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]189<!-- ================================================= -->
190<h4 id="cl_diag_formatting">Formatting of Diagnostics</h4>
191<!-- ================================================= -->
192
193<p>Clang aims to produce beautiful diagnostics by default, particularly for new
194users that first come to Clang. However, different people have different
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]195preferences, and sometimes Clang is driven by another program that wants to
196parse simple and consistent output, not a person. For these cases, Clang
197provides a wide range of options to control the exact output format of the
198diagnostics that it generates.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]199
200<dl>
201
202<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
203<dt id="opt_fshow-column"><b>-f[no-]show-column</b>: Print column number in
204diagnostic.</dt>
205<dd>This option, which defaults to on, controls whether or not Clang prints the
206column number of a diagnostic. For example, when this is enabled, Clang will
207print something like:</p>
208
209<pre>
210 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
211 #endif bad
212 ^
213 //
214</pre>
215
216<p>When this is disabled, Clang will print "test.c:28: warning..." with no
217column number.</p>
218</dd>
219
220<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
221<dt id="opt_fshow-source-location"><b>-f[no-]show-source-location</b>: Print
222source file/line/column information in diagnostic.</dt>
223<dd>This option, which defaults to on, controls whether or not Clang prints the
224filename, line number and column number of a diagnostic. For example,
225when this is enabled, Clang will print something like:</p>
226
227<pre>
228 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
229 #endif bad
230 ^
231 //
232</pre>
233
234<p>When this is disabled, Clang will not print the "test.c:28:8: " part.</p>
235</dd>
236
237<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
238<dt id="opt_fcaret-diagnostics"><b>-f[no-]caret-diagnostics</b>: Print source
239line and ranges from source code in diagnostic.</dt>
240<dd>This option, which defaults to on, controls whether or not Clang prints the
241source line, source ranges, and caret when emitting a diagnostic. For example,
242when this is enabled, Clang will print something like:</p>
243
244<pre>
245 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
246 #endif bad
247 ^
248 //
249</pre>
250
251<p>When this is disabled, Clang will just print:</p>
252
253<pre>
Douglas Gregore8d44dd2010-07-09 16:31:58 +0000[diff] [blame]254 <b><font color="black">test.c:28:8: <font color="magenta">warning</font>: extra tokens at end of #endif directive [-Wextra-tokens]</font></b>
255 #endif bad
256 <font color="green">^</font>
257 <font color="green">//</font>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]258</pre>
259
260</dd>
261
262<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Douglas Gregore8d44dd2010-07-09 16:31:58 +0000[diff] [blame]263<dt id="opt_fcolor_diagnostics"><b>-f[no-]color-diagnostics</b>: </dt>
264<dd>This option, which defaults to on when a color-capable terminal is
265 detected, controls whether or not Clang prints diagnostics in color.
266 When this option is enabled, Clang will use colors to highlight
267 specific parts of the diagnostic, e.g.,
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>'
276<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]277<dt id="opt_fdiagnostics-show-option"><b>-f[no-]diagnostics-show-option</b>:
278Enable <tt>[-Woption]</tt> information in diagnostic line.</dt>
279<dd>This option, which defaults to on,
280controls whether or not Clang prints the associated <A
281href="#cl_diag_warning_groups">warning group</a> option name when outputting
282a warning diagnostic. For example, in this output:</p>
283
284<pre>
285 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
286 #endif bad
287 ^
288 //
289</pre>
290
291<p>Passing <b>-fno-diagnostics-show-option</b> will prevent Clang from printing
292the [<a href="#opt_Wextra-tokens">-Wextra-tokens</a>] information in the
293diagnostic. This information tells you the flag needed to enable or disable the
294diagnostic, either from the command line or through <a
295href="#pragma_GCC_diagnostic">#pragma GCC diagnostic</a>.</dd>
296
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]297<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
298<dt id="opt_fdiagnostics-show-category"><b>-fdiagnostics-show-category=none/id/name</b>:
299Enable printing category information in diagnostic line.</dt>
300<dd>This option, which defaults to "none",
301controls whether or not Clang prints the category associated with a diagnostic
302when emitting it. Each diagnostic may or many not have an associated category,
303if it has one, it is listed in the diagnostic categorization field of the
304diagnostic line (in the []'s).</p>
305
306<p>For example, a format string warning will produce these three renditions
307based on the setting of this option:</p>
308
309<pre>
310 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat]
311 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,1</b>]
312 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,Format String</b>]
313</pre>
314
315<p>This category can be used by clients that want to group diagnostics by
316category, so it should be a high level category. We want dozens of these, not
317hundreds or thousands of them.</p>
318</dd>
319
320
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]321
322<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
323<dt id="opt_fdiagnostics-fixit-info"><b>-f[no-]diagnostics-fixit-info</b>:
324Enable "FixIt" information in the diagnostics output.</dt>
325<dd>This option, which defaults to on, controls whether or not Clang prints the
326information on how to fix a specific diagnostic underneath it when it knows.
327For example, in this output:</p>
328
329<pre>
330 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
331 #endif bad
332 ^
333 //
334</pre>
335
336<p>Passing <b>-fno-diagnostics-fixit-info</b> will prevent Clang from printing
337the "//" line at the end of the message. This information is useful for users
338who may not understand what is wrong, but can be confusing for machine
339parsing.</p>
340</dd>
341
342<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]343<dt id="opt_fdiagnostics-print-source-range-info">
344<b>-f[no-]diagnostics-print-source-range-info</b>:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]345Print machine parsable information about source ranges.</dt>
346<dd>This option, which defaults to off, controls whether or not Clang prints
347information about source ranges in a machine parsable format after the
348file/line/column number information. The information is a simple sequence of
349brace enclosed ranges, where each range lists the start and end line/column
350locations. For example, in this output:</p>
351
352<pre>
353exprs.c:47:15:{47:8-47:14}{47:17-47:24}: error: invalid operands to binary expression ('int *' and '_Complex float')
354 P = (P-42) + Gamma*4;
355 ~~~~~~ ^ ~~~~~~~
356</pre>
357
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]358<p>The {}'s are generated by -fdiagnostics-print-source-range-info.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]359</dd>
360
Douglas Gregor4786c152010-08-19 20:24:43 +0000[diff] [blame^]361<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
362<dt id="opt_fdiagnostics-parseable-fixits">
363<b>-fdiagnostics-parseable-fixits</b>:
364Print Fix-Its in a machine parseable form.</dt>
365<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>
366
367<pre>
368 fix-it: "t.cpp":{7:25-7:29}: "Gamma"
369</pre>
370
371<p>The range printed is a half-open range, so in this example the characters at column 25 up to but not including column 29 on line 7 in t.cpp should be replaced with the string "Gamma". Either the range or the replacement string may be empty (representing strict insertions and strict erasures, respectively). Both the file name and the insertion string escape '\', tabs (as "\n"), newlines (as "\n"), double quotes(as "\n") and non-printable characters (as octal "\xxx").</p>
372</dd>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]373
374</dl>
375
376
377
378
379<!-- ===================================================== -->
380<h4 id="cl_diag_warning_groups">Individual Warning Groups</h4>
381<!-- ===================================================== -->
382
383<p>TODO: Generate this from tblgen. Define one anchor per warning group.</p>
384
385
386<dl>
387
388
389<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
390<dt id="opt_Wextra-tokens"><b>-Wextra-tokens</b>: Warn about excess tokens at
391 the end of a preprocessor directive.</dt>
392<dd>This option, which defaults to on, enables warnings about extra tokens at
393the end of preprocessor directives. For example:</p>
394
395<pre>
396 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
397 #endif bad
398 ^
399</pre>
400
401<p>These extra tokens are not strictly conforming, and are usually best handled
402by commenting them out.</p>
403
404<p>This option is also enabled by <a href="">-Wfoo</a>, <a href="">-Wbar</a>,
405 and <a href="">-Wbaz</a>.</p>
406</dd>
407
Jeffrey Yasskin21d07e42010-06-05 01:39:57 +0000[diff] [blame]408<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
409<dt id="opt_Wambiguous-member-template"><b>-Wambiguous-member-template</b>:
410Warn about unqualified uses of a member template whose name resolves
411to another template at the location of the use.</dt>
412<dd>This option, which defaults to on, enables a warning in the
413following code:</p>
414
415<pre>
416template&lt;typename T> struct set{};
417template&lt;typename T> struct trait { typedef const T& type; };
418struct Value {
419 template&lt;typename T> void set(typename trait&lt;T>::type value) {}
420};
421void foo() {
422 Value v;
423 v.set&lt;double>(3.2);
424}
425</pre>
426
427<p>C++ [basic.lookup.classref] requires this to be an error, but,
428because it's hard to work around, Clang downgrades it to a warning as
429an extension.</p>
430</dd>
431
Jeffrey Yasskin57d12fd2010-06-07 15:58:05 +0000[diff] [blame]432<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
433<dt id="opt_Wbind-to-temporary-copy"><b>-Wbind-to-temporary-copy</b>: Warn about
434an unusable copy constructor when binding a reference to a temporary.</dt>
435<dd>This option, which defaults to on, enables warnings about binding a
436reference to a temporary when the temporary doesn't have a usable copy
437constructor. For example:</p>
438
439<pre>
440 struct NonCopyable {
441 NonCopyable();
442 private:
443 NonCopyable(const NonCopyable&);
444 };
445 void foo(const NonCopyable&);
446 void bar() {
447 foo(NonCopyable()); // Disallowed in C++98; allowed in C++0x.
448 }
449</pre>
450<pre>
451 struct NonCopyable2 {
452 NonCopyable2();
453 NonCopyable2(NonCopyable2&);
454 };
455 void foo(const NonCopyable2&);
456 void bar() {
457 foo(NonCopyable2()); // Disallowed in C++98; allowed in C++0x.
458 }
459</pre>
460
461<p>Note that if <tt>NonCopyable2::NonCopyable2()</tt> has a default
462argument whose instantiation produces a compile error, that error will
463still be a hard error in C++98 mode even if this warning is turned
464off.</p>
465
466</dd>
467
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]468</dl>
469
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]470<!-- ======================================================================= -->
471<h2 id="general_features">Language and Target-Independent Features</h2>
472<!-- ======================================================================= -->
473
474
475<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
476<h3 id="diagnostics">Controlling Errors and Warnings</h3>
477<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
478
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]479<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]480emit errors and warning messages, and how they are displayed to the console.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]481
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]482<h4 id="diagnostics_display">Controlling How Clang Displays Diagnostics</h4>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]483
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]484<p>When Clang emits a diagnostic, it includes rich information in the output,
485and gives you fine-grain control over which information is printed. Clang has
486the ability to print this information, and these are the options that control
487it:</p>
488
489<p>
490<ol>
491<li>A file/line/column indicator that shows exactly where the diagnostic occurs
492 in your code [<a href="#opt_fshow-column">-fshow-column</a>, <a
493 href="#opt_fshow-source-location">-fshow-source-location</a>].</li>
494<li>A categorization of the diagnostic as a note, warning, error, or fatal
495 error.</li>
496<li>A text string that describes what the problem is.</li>
497<li>An option that indicates how to control the diagnostic (for diagnostics that
498 support it) [<a
499 href="#opt_fdiagnostics-show-option">-fdiagnostics-show-option</a>].</li>
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]500<li>A <a href="#diagnostics_categories">high-level category</a> for the
501 diagnostic for clients that want to group diagnostics by class (for
502 diagnostics that support it) [<a
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]503 href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]504<li>The line of source code that the issue occurs on, along with a caret and
505 ranges that indicate the important locations [<a
506 href="opt_fcaret-diagnostics">-fcaret-diagnostics</a>].</li>
507<li>"FixIt" information, which is a concise explanation of how to fix the
508 problem (when Clang is certain it knows) [<a
509 href="opt_fdiagnostics-fixit-info">-fdiagnostics-fixit-info</a>].</li>
510<li>A machine-parsable representation of the ranges involved (off by
511 default) [<a
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]512 href="opt_fdiagnostics-print-source-range-info">-fdiagnostics-print-source-range-info</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]513</ol></p>
514
515<p>For more information please see <a href="#cl_diag_formatting">Formatting of
516Diagnostics</a>.</p>
517
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]518
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]519<h4 id="diagnostics_mappings">Diagnostic Mappings</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]520
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]521<p>All diagnostics are mapped into one of these 5 classes:</p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]522
523<p>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]524<ul>
525<li>Ignored</li>
526<li>Note</li>
527<li>Warning</li>
528<li>Error</li>
529<li>Fatal</li>
530</ul></p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]531
Chris Lattner3f145382010-05-24 21:35:18 +0000[diff] [blame]532<h4 id="diagnostics_categories">Diagnostic Categories</h4>
533
534<p>Though not shown by default, diagnostics may each be associated with a
535 high-level category. This category is intended to make it possible to triage
536 builds that produce a large number of errors or warnings in a grouped way.
537</p>
538
539<p>Categories are not shown by default, but they can be turned on with the
540<a href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a> option.
541When set to "<tt>name</tt>", the category is printed textually in the diagnostic
542output. When it is set to "<tt>id</tt>", a category number is printed. The
543mapping of category names to category id's can be obtained by running '<tt>clang
544 --print-diagnostic-categories</tt>'.
545</p>
546
547<h4 id="diagnostics_commandline">Controlling Diagnostics via Command Line
548 Flags</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]549
550<p>-W flags, -pedantic, etc</p>
551
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]552<h4 id="diagnostics_pragmas">Controlling Diagnostics via Pragmas</h4>
553
554<p>Clang can also control what diagnostics are enabled through the use of
555pragmas in the source code. This is useful for turning off specific warnings
556in a section of source code. Clang supports GCC's pragma for compatibility
557with existing source code, as well as several extensions. </p>
558
559<p>The pragma may control any warning that can be used from the command line.
560Warnings may be set to ignored, warning, error, or fatal. The following
561example code will tell Clang or GCC to ignore the -Wall warnings:</p>
562
563<pre>
564#pragma GCC diagnostic ignored "-Wall"
565</pre>
566
567<p>In addition to all of the functionality of provided by GCC's pragma, Clang
568also allows you to push and pop the current warning state. This is particularly
569useful when writing a header file that will be compiled by other people, because
570you don't know what warning flags they build with.</p>
571
572<p>In the below example
573-Wmultichar is ignored for only a single line of code, after which the
574diagnostics return to whatever state had previously existed.</p>
575
576<pre>
577#pragma clang diagnostic push
578#pragma clang diagnostic ignored "-Wmultichar"
579
580char b = 'df'; // no warning.
581
582#pragma clang diagnostic pop
583</pre>
584
585<p>The push and pop pragmas will save and restore the full diagnostic state of
586the compiler, regardless of how it was set. That means that it is possible to
587use push and pop around GCC compatible diagnostics and Clang will push and pop
588them appropriately, while GCC will ignore the pushes and pops as unknown
589pragmas. It should be noted that while Clang supports the GCC pragma, Clang and
590GCC do not support the exact same set of warnings, so even when using GCC
591compatible #pragmas there is no guarantee that they will have identical behaviour
592on both compilers. </p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]593
594<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
595<h3 id="precompiledheaders">Precompiled Headers</h3>
596<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
597
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]598<p><a href="http://en.wikipedia.org/wiki/Precompiled_header">Precompiled
599headers</a> are a general approach employed by many compilers to reduce
600compilation time. The underlying motivation of the approach is that it is
601common for the same (and often large) header files to be included by
602multiple source files. Consequently, compile times can often be greatly improved
603by caching some of the (redundant) work done by a compiler to process headers.
604Precompiled header files, which represent one of many ways to implement
605this optimization, are literally files that represent an on-disk cache that
606contains the vital information necessary to reduce some of the work
607needed to process a corresponding header file. While details of precompiled
608headers vary between compilers, precompiled headers have been shown to be a
609highly effective at speeding up program compilation on systems with very large
610system headers (e.g., Mac OS/X).</p>
611
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]612<h4>Generating a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]613
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]614<p>To generate a PCH file using Clang, one invokes Clang with
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]615the <b><tt>-x <i>&lt;language&gt;</i>-header</tt></b> option. This mirrors the
616interface in GCC for generating PCH files:</p>
617
618<pre>
619 $ gcc -x c-header test.h -o test.h.gch
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]620 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]621</pre>
622
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]623<h4>Using a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]624
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]625<p>A PCH file can then be used as a prefix header when a
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]626<b><tt>-include</tt></b> option is passed to <tt>clang</tt>:</p>
627
628<pre>
629 $ clang -include test.h test.c -o test
630</pre>
631
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]632<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]633is 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]634will be processed from the PCH file. Otherwise, Clang falls back to
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]635directly processing the content of <tt>test.h</tt>. This mirrors the behavior of
636GCC.</p>
637
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]638<p><b>NOTE:</b> Clang does <em>not</em> automatically use PCH files
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]639for headers that are directly included within a source file. For example:</p>
640
641<pre>
Chris Lattnere42ec542009-06-13 20:35:58 +0000[diff] [blame]642 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]643 $ cat test.c
644 #include "test.h"
645 $ clang test.c -o test
646</pre>
647
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]648<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]649<tt>test.h</tt> since <tt>test.h</tt> was included directly in the source file
650and not specified on the command line using <tt>-include</tt>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]651
Douglas Gregore650c8c2009-07-07 00:12:59 +0000[diff] [blame]652<h4>Relocatable PCH Files</h4>
653<p>It is sometimes necessary to build a precompiled header from headers that
654are not yet in their final, installed locations. For example, one might build a
655precompiled header within the build tree that is then meant to be installed
656alongside the headers. Clang permits the creation of "relocatable" precompiled
657headers, which are built with a given path (into the build directory) and can
658later be used from an installed location.</p>
659
660<p>To build a relocatable precompiled header, place your headers into a
661subdirectory whose structure mimics the installed location. For example, if you
662want to build a precompiled header for the header <code>mylib.h</code> that
663will be installed into <code>/usr/include</code>, create a subdirectory
664<code>build/usr/include</code> and place the header <code>mylib.h</code> into
665that subdirectory. If <code>mylib.h</code> depends on other headers, then
666they can be stored within <code>build/usr/include</code> in a way that mimics
667the installed location.</p>
668
669<p>Building a relocatable precompiled header requires two additional arguments.
670First, pass the <code>--relocatable-pch</code> flag to indicate that the
671resulting PCH file should be relocatable. Second, pass
672<code>-isysroot /path/to/build</code>, which makes all includes for your
673library relative to the build directory. For example:</p>
674
675<pre>
676 # clang -x c-header --relocatable-pch -isysroot /path/to/build /path/to/build/mylib.h mylib.h.pch
677</pre>
678
679<p>When loading the relocatable PCH file, the various headers used in the PCH
680file are found from the system header root. For example, <code>mylib.h</code>
681can be found in <code>/usr/include/mylib.h</code>. If the headers are installed
682in some other system root, the <code>-isysroot</code> option can be used provide
683a different system root from which the headers will be based. For example,
684<code>-isysroot /Developer/SDKs/MacOSX10.4u.sdk</code> will look for
685<code>mylib.h</code> in
686<code>/Developer/SDKs/MacOSX10.4u.sdk/usr/include/mylib.h</code>.</p>
687
688<p>Relocatable precompiled headers are intended to be used in a limited number
689of cases where the compilation environment is tightly controlled and the
690precompiled header cannot be generated after headers have been installed.
691Relocatable precompiled headers also have some performance impact, because
692the difference in location between the header locations at PCH build time vs.
693at the time of PCH use requires one of the PCH optimizations,
694<code>stat()</code> caching, to be disabled. However, this change is only
695likely to affect PCH files that reference a large number of headers.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]696
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]697<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
698<h3 id="codegen">Controlling Code Generation</h3>
699<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
700
701<p>Clang provides a number of ways to control code generation. The options are listed below.</p>
702
703<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
704<dt id="opt_fcatch-undefined-behavior"><b>-fcatch-undefined-behavior</b>: Turn
705on runtime code generation to check for undefined behavior.</dt>
706
707<dd>This option, which defaults to off, controls whether or not Clang
708adds runtime checks for undefined runtime behavior. If the check fails,
709<tt>__builtin_trap()</tt> is used to indicate failure.
710The checks are:
711<p>
Mike Stump88b2a172009-12-16 03:25:12 +0000[diff] [blame]712<li>Subscripting where the static type of one operand is variable
713 which is decayed from an array type and the other operand is
714 greater than the size of the array or less than zero.</li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]715<li>Shift operators where the amount shifted is greater or equal to the
716 promoted bit-width of the left-hand-side or less than zero.</li>
Mike Stump8f6a3ed2009-12-16 03:18:14 +0000[diff] [blame]717<li>If control flow reaches __builtin_unreachable.
718<li>When llvm implements more __builtin_object_size support, reads and
719 writes for objects that __builtin_object_size indicates we aren't
720 accessing valid memory. Bit-fields and vectors are not yet checked.
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]721</p>
722</dd>
723
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]724<dt id="opt_fno-assume-sane-operator-new"><b>-fno-assume-sane-operator-new</b>:
725Don't assume that the C++'s new operator is sane.</dt>
Nuno Lopesb23f20d2009-12-17 10:15:49 +0000[diff] [blame]726<dd>This option tells the compiler to do not assume that C++'s global new
727operator will always return a pointer that do not
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]728alias any other pointer when the function returns.</dd>
729
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]730<!-- ======================================================================= -->
731<h2 id="c">C Language Features</h2>
732<!-- ======================================================================= -->
733
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]734<p>The support for standard C in clang is feature-complete except for the C99
735floating-point pragmas.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]736
737<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]738<h3 id="c_ext">Extensions supported by clang</h3>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]739<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
740
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]741<p>See <a href="LanguageExtensions.html">clang language extensions</a>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]742
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]743<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
744<h3 id="c_modes">Differences between various standard modes</h3>
745<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]746
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]747<p>clang supports the -std option, which changes what language mode clang uses.
748The supported modes for C are c89, gnu89, c94, c99, gnu99 and various aliases
749for those modes. If no -std option is specified, clang defaults to gnu99 mode.
750</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]751
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]752<p>Differences between all c* and gnu* modes:</p>
753<ul>
754<li>c* modes define "__STRICT_ANSI__".</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]755<li>Target-specific defines not prefixed by underscores, like "linux", are
756defined in gnu* modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]757<li>Trigraphs default to being off in gnu* modes; they can be enabled by the
758-trigraphs option.</li>
759<li>The parser recognizes "asm" and "typeof" as keywords in gnu* modes; the
760variants "__asm__" and "__typeof__" are recognized in all modes.</li>
Eli Friedmanb0ac5452009-05-16 23:17:30 +0000[diff] [blame]761<li>The Apple "blocks" extension is recognized by default in gnu* modes
762on some platforms; it can be enabled in any mode with the "-fblocks"
763option.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]764</ul>
765
766<p>Differences between *89 and *99 modes:</p>
767<ul>
768<li>The *99 modes default to implementing "inline" as specified in C99, while
769the *89 modes implement the GNU version. This can be overridden for individual
770functions with the __gnu_inline__ attribute.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]771<li>Digraphs are not recognized in c89 mode.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]772<li>The scope of names defined inside a "for", "if", "switch", "while", or "do"
773statement is different. (example: "if ((struct x {int x;}*)0) {}".)</li>
774<li>__STDC_VERSION__ is not defined in *89 modes.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]775<li>"inline" is not recognized as a keyword in c89 mode.</li>
776<li>"restrict" is not recognized as a keyword in *89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]777<li>Commas are allowed in integer constant expressions in *99 modes.</li>
778<li>Arrays which are not lvalues are not implicitly promoted to pointers in
779*89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]780<li>Some warnings are different.</li>
781</ul>
782
783<p>c94 mode is identical to c89 mode except that digraphs are enabled in
784c94 mode (FIXME: And __STDC_VERSION__ should be defined!).</p>
785
786<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
787<h3 id="c_unimpl_gcc">GCC extensions not implemented yet</h3>
788<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
789
790<p>clang tries to be compatible with gcc as much as possible, but some gcc
791extensions are not implemented yet:</p>
792
793<ul>
794<li>clang does not support __label__
795(<a href="http://llvm.org/bugs/show_bug.cgi?id=3429">bug 3429</a>). This is
796a relatively small feature, so it is likely to be implemented relatively
797soon.</li>
798
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]799<li>clang does not support #pragma weak
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]800(<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]801the uses described in the bug, this is likely to be implemented at some
802point, at least partially.</li>
803
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]804<li>clang does not support code generation for local variables pinned to
805registers (<a href="http://llvm.org/bugs/show_bug.cgi?id=3933">bug 3933</a>).
806This is a relatively small feature, so it is likely to be implemented
807relatively soon.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]808
809<li>clang does not support decimal floating point types (_Decimal32 and
810friends) or fixed-point types (_Fract and friends); nobody has expressed
811interest in these features yet, so it's hard to say when they will be
812implemented.</li>
813
814<li>clang does not support nested functions; this is a complex feature which
815is infrequently used, so it is unlikely to be implemented anytime soon.</li>
816
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]817<li>clang does not support global register variables, this is unlikely
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]818to be implemented soon because it requires additional LLVM backend support.
819</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]820
821<li>clang does not support static initialization of flexible array
822members. This appears to be a rarely used extension, but could be
823implemented pending user demand.</li>
824
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]825<li>clang does not support __builtin_va_arg_pack/__builtin_va_arg_pack_len.
826This is used rarely, but in some potentially interesting places, like the
827glibc headers, so it may be implemented pending user demand. Note that
828because clang pretends to be like GCC 4.2, and this extension was introduced
829in 4.3, the glibc headers will not try to use this extension with clang at
830the moment.</li>
831
832<li>clang does not support the gcc extension for forward-declaring function
833parameters; this has not showed up in any real-world code yet, though, so it
834might never be implemented.</li>
835
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]836</ul>
837
838<p>This is not a complete list; if you find an unsupported extension
839missing from this list, please send an e-mail to cfe-dev. This list
840currently excludes C++; see <a href="#cxx">C++ Language Features</a>.
841Also, this list does not include bugs in mostly-implemented features; please
842see 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">
843bug tracker</a> for known existing bugs (FIXME: Is there a section for
844bug-reporting guidelines somewhere?).</p>
845
846<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
847<h3 id="c_unsupp_gcc">Intentionally unsupported GCC extensions</h3>
848<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
849
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]850<ul>
851
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]852<li>clang does not support the gcc extension that allows variable-length arrays
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]853in structures. This is for a few of reasons: one, it is tricky
854to implement, two, the extension is completely undocumented, and three, the
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]855extension appears to be rarely used.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]856
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]857<li>clang does not support duplicate definitions of a function where one is
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]858inline. This complicates clients of the AST which normally can expect there is
859at most one definition for each function. Source code using this feature should
860be changed to define the inline and out-of-line definitions in separate
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]861translation units.</li>
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]862
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]863<li>clang does not have an equivalent to gcc's "fold"; this means that
864clang doesn't accept some constructs gcc might accept in contexts where a
865constant expression is required, like "x-x" where x is a variable, or calls
866to C library functions like strlen.</li>
867
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]868<li>clang does not support multiple alternative constraints in inline asm; this
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]869is an extremely obscure feature which would be complicated to implement
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]870correctly.</li>
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]871
872<li>clang does not support __builtin_apply and friends; this extension is
873extremely obscure and difficult to implement reliably.</li>
874
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]875</ul>
876
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]877<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
878<h3 id="c_ms">Microsoft extensions</h3>
879<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
880
881<p>clang has some experimental support for extensions from
882Microsoft Visual C++; to enable it, use the -fms-extensions command-line
Eli Friedmana1821b52009-06-08 06:21:03 +0000[diff] [blame]883option. This is the default for Windows targets. Note that the
884support is incomplete; enabling Microsoft extensions will silently drop
885certain constructs (including __declspec and Microsoft-style asm statements).
886</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]887
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]888<li>clang does not support the Microsoft extension where anonymous
Daniel Dunbar869e0542009-05-13 00:23:53 +0000[diff] [blame]889record members can be declared using user defined typedefs.</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]890
Daniel Dunbar9375ed12009-05-13 21:40:49 +0000[diff] [blame]891<li>clang supports the Microsoft "#pragma pack" feature for
892controlling record layout. GCC also contains support for this feature,
893however where MSVC and GCC are incompatible clang follows the MSVC
894definition.</li>
895
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]896<!-- ======================================================================= -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]897<h2 id="target_features">Target-Specific Features and Limitations</h2>
898<!-- ======================================================================= -->
899
900
901<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
902<h3 id="target_arch">CPU Architectures Features and Limitations</h3>
903<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
904
905<!-- ======================== -->
906<h4 id="target_arch_x86">X86</h4>
907<!-- ======================== -->
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]908<p>The support for X86 (both 32-bit and 64-bit) is considered stable
909on Darwin (Mac OS/X), Linux, FreeBSD, and Dragonfly BSD: it has been tested to
910correctly compile large C and Objective-C codebases. (FIXME: Anything specific
911we want to say here? Possibly mention some LLVM x86 limitations?)
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]912
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]913<!-- ======================== -->
914<h4 id="target_arch_arm">ARM</h4>
915<!-- ======================== -->
916ARM support is mostly feature-complete, but still experimental; it hasn't
917undergone significant testing.
918
919<!-- ======================== -->
920<h4 id="target_arch_other">Other platforms</h4>
921<!-- ======================== -->
922clang currently contains some support for PPC and Sparc; however, significant
923pieces of code generation are still missing, and they haven't undergone
924significant testing.
925
926<p>clang contains some support for the embedded PIC16 processor
927(FIXME: I haven't been keeping track of this; what should this say?).
928
929<p>clang contains limited support for the MSP430 embedded processor, but both
930the clang support and the LLVM backend support are highly experimental.
931
932<p>Other platforms are completely unsupported at the moment. Adding the
933minimal support needed for parsing and semantic analysis on a new platform
934is quite easy; see lib/Basic/Targets.cpp in the clang source tree. This level
935of support is also sufficient for conversion to LLVM IR for simple programs.
936Proper support for conversion to LLVM IR requires adding code to
937lib/CodeGen/CGCall.cpp at the moment; this is likely to change soon, though.
938Generating assembly requires a suitable LLVM backend.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]939
940<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
941<h3 id="target_os">Operating System Features and Limitations</h3>
942<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
943
944<!-- ======================================= -->
945<h4 id="target_os_darwin">Darwin (Mac OS/X)</h4>
946<!-- ======================================= -->
947
948<p>No __thread support, 64-bit ObjC support requires SL tools.</p>
949
950</div>
951</body>
952</html>