Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / clang / 204ce17e0cfd9bbe229627e1e5a20c3f2f587c8c / . / docs / UsersManual.html
blob: 4ba00e0b9af6f9729ef25aa381eed5ba0ecb06e8 [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>
39 <li><a href="#diagnostics_commandline">Controlling Diagnostics via Command Line Flags</a></li>
40 <li><a href="#diagnostics_pragmas">Controlling Diagnostics via Pragmas</a></li>
41 </ul>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]42 <li><a href="#precompiledheaders">Precompiled Headers</a></li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]43 <li><a href="#codegen">Controlling Code Generation</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]44 </ul>
45</li>
46<li><a href="#c">C Language Features</a>
47 <ul>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]48 <li><a href="#c_ext">Extensions supported by clang</a></li>
49 <li><a href="#c_modes">Differences between various standard modes</a></li>
50 <li><a href="#c_unimpl_gcc">GCC extensions not implemented yet</a></li>
51 <li><a href="#c_unsupp_gcc">Intentionally unsupported GCC extensions</a></li>
52 <li><a href="#c_ms">Microsoft extensions</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]53 </ul>
54</li>
55<li><a href="#objc">Objective-C Language Features</a>
56 <ul>
57 <li><a href="#objc_incompatibilities">Intentional Incompatibilities with
58 GCC</a></li>
59 </ul>
60</li>
61<li><a href="#cxx">C++ Language Features</a>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]62</li>
63<li><a href="#objcxx">Objective C++ Language Features</a>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]64</li>
65<li><a href="#target_features">Target-Specific Features and Limitations</a>
66 <ul>
67 <li><a href="#target_arch">CPU Architectures Features and Limitations</a>
68 <ul>
69 <li><a href="#target_arch_x86">X86</a></li>
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]70 <li><a href="#target_arch_arm">ARM</a></li>
71 <li><a href="#target_arch_other">Other platforms</a></li>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]72 </ul>
73 </li>
74 <li><a href="#target_os">Operating System Features and Limitations</a>
75 <ul>
76 <li><a href="#target_os_darwin">Darwin (Mac OS/X)</a></li>
77 <li>Linux, etc.</li>
78 </ul>
79
80 </li>
81 </ul>
82</li>
83</ul>
84
85
86<!-- ======================================================================= -->
87<h2 id="intro">Introduction</h2>
88<!-- ======================================================================= -->
89
90<p>The Clang Compiler is an open-source compiler for the C family of programming
91languages, aiming to be the best in class implementation of these languages.
92Clang builds on the LLVM optimizer and code generator, allowing it to provide
93high-quality optimization and code generation support for many targets. For
94more general information, please see the <a href="http://clang.llvm.org">Clang
95Web Site</a> or the <a href="http://llvm.org">LLVM Web Site</a>.</p>
96
97<p>This document describes important notes about using Clang as a compiler for
98an end-user, documenting the supported features, command line options, etc. If
99you are interested in using Clang to build a tool that processes code, please
100see <a href="InternalsManual.html">the Clang Internals Manual</a>. If you are
101interested in the <a href="http://clang.llvm.org/StaticAnalysis.html">Clang
102Static Analyzer</a>, please see its web page.</p>
103
104<p>Clang is designed to support the C family of programming languages, which
105includes <a href="#c">C</a>, <a href="#objc">Objective-C</a>, <a
106href="#cxx">C++</a>, and <a href="#objcxx">Objective-C++</a> as well as many
107dialects of those. For language-specific information, please see the
108corresponding language specific section:</p>
109
110<ul>
111<li><a href="#c">C Language</a>: K&amp;R C, ANSI C89, ISO C90, ISO C94
112 (C89+AMD1), ISO C99 (+TC1, TC2, TC3). </li>
113<li><a href="#objc">Objective-C Language</a>: ObjC 1, ObjC 2, ObjC 2.1, plus
114 variants depending on base language.</li>
115<li><a href="#cxx">C++ Language Features</a></li>
116<li><a href="#objcxx">Objective C++ Language</a></li>
117</ul>
118
119<p>In addition to these base languages and their dialects, Clang supports a
120broad variety of language extensions, which are documented in the corresponding
121language section. These extensions are provided to be compatible with the GCC,
122Microsoft, and other popular compilers as well as to improve functionality
123through Clang-specific features. The Clang driver and language features are
124intentionally designed to be as compatible with the GNU GCC compiler as
125reasonably possible, easing migration from GCC to Clang. In most cases, code
126"just works".</p>
127
128<p>In addition to language specific features, Clang has a variety of features
129that depend on what CPU architecture or operating system is being compiled for.
Douglas Gregorcd5a5052009-11-09 15:15:41 +0000[diff] [blame]130Please see the <a href="#target_features">Target-Specific Features and
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]131Limitations</a> section for more details.</p>
132
133<p>The rest of the introduction introduces some basic <a
134href="#terminology">compiler terminology</a> that is used throughout this manual
135and contains a basic <a href="#basicusage">introduction to using Clang</a>
136as a command line compiler.</p>
137
138<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
139<h3 id="terminology">Terminology</h3>
140<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
141
142<p>Front end, parser, backend, preprocessor, undefined behavior, diagnostic,
143 optimizer</p>
144
145<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
146<h3 id="basicusage">Basic Usage</h3>
147<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
148
149<p>Intro to how to use a C compiler for newbies.</p>
150<p>
151compile + link
152
153compile then link
154
155debug info
156
157enabling optimizations
158
159picking a language to use, defaults to C99 by default. Autosenses based on
160extension.
161
162using a makefile
163</p>
164
165
166<!-- ======================================================================= -->
167<h2 id="commandline">Command Line Options</h2>
168<!-- ======================================================================= -->
169
170<p>
171This section is generally an index into other sections. It does not go into
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]172depth on the ones that are covered by other sections. However, the first part
173introduces the language selection and other high level options like -c, -g, etc.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]174</p>
175
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]176
177<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
178<h3 id="cl_diagnostics">Options to Control Error and Warning Messages</h3>
179<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
180
181<p><b>-Werror</b>: Turn warnings into errors.</p>
182<p><b>-Werror=foo</b>: Turn warning "foo" into an error.</p>
183<p><b>-Wno-error=foo</b>: Turn warning "foo" into an warning even if -Werror is
184 specified.</p>
185<p><b>-Wfoo</b>: Enable warning foo</p>
186<p><b>-Wno-foo</b>: Disable warning foo</p>
187<p><b>-w</b>: Disable all warnings.</p>
188<p><b>-pedantic</b>: Warn on language extensions.</p>
189<p><b>-pedantic-errors</b>: Error on language extensions.</p>
190<p><b>-Wsystem-headers</b>: Enable warnings from system headers.</p>
191
Chris Lattner0f0c9632010-04-07 20:49:23 +0000[diff] [blame]192<p><b>-ferror-limit=123</b>: Stop emitting diagnostics after 123 errors have
193 been produced. The default is 20, and the error limit can be disabled with
194 -ferror-limit=0.</p>
195
Douglas Gregor575cf372010-04-20 07:18:24 +0000[diff] [blame]196<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>
197
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]198<!-- ================================================= -->
199<h4 id="cl_diag_formatting">Formatting of Diagnostics</h4>
200<!-- ================================================= -->
201
202<p>Clang aims to produce beautiful diagnostics by default, particularly for new
203users that first come to Clang. However, different people have different
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]204preferences, and sometimes Clang is driven by another program that wants to
205parse simple and consistent output, not a person. For these cases, Clang
206provides a wide range of options to control the exact output format of the
207diagnostics that it generates.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]208
209<dl>
210
211<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
212<dt id="opt_fshow-column"><b>-f[no-]show-column</b>: Print column number in
213diagnostic.</dt>
214<dd>This option, which defaults to on, controls whether or not Clang prints the
215column number of a diagnostic. For example, when this is enabled, Clang will
216print something like:</p>
217
218<pre>
219 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
220 #endif bad
221 ^
222 //
223</pre>
224
225<p>When this is disabled, Clang will print "test.c:28: warning..." with no
226column number.</p>
227</dd>
228
229<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
230<dt id="opt_fshow-source-location"><b>-f[no-]show-source-location</b>: Print
231source file/line/column information in diagnostic.</dt>
232<dd>This option, which defaults to on, controls whether or not Clang prints the
233filename, line number and column number of a diagnostic. For example,
234when this is enabled, Clang will print something like:</p>
235
236<pre>
237 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
238 #endif bad
239 ^
240 //
241</pre>
242
243<p>When this is disabled, Clang will not print the "test.c:28:8: " part.</p>
244</dd>
245
246<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
247<dt id="opt_fcaret-diagnostics"><b>-f[no-]caret-diagnostics</b>: Print source
248line and ranges from source code in diagnostic.</dt>
249<dd>This option, which defaults to on, controls whether or not Clang prints the
250source line, source ranges, and caret when emitting a diagnostic. For example,
251when this is enabled, Clang will print something like:</p>
252
253<pre>
254 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
255 #endif bad
256 ^
257 //
258</pre>
259
260<p>When this is disabled, Clang will just print:</p>
261
262<pre>
263 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
264</pre>
265
266</dd>
267
268<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
269<dt id="opt_fdiagnostics-show-option"><b>-f[no-]diagnostics-show-option</b>:
270Enable <tt>[-Woption]</tt> information in diagnostic line.</dt>
271<dd>This option, which defaults to on,
272controls whether or not Clang prints the associated <A
273href="#cl_diag_warning_groups">warning group</a> option name when outputting
274a warning diagnostic. For example, in this output:</p>
275
276<pre>
277 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
278 #endif bad
279 ^
280 //
281</pre>
282
283<p>Passing <b>-fno-diagnostics-show-option</b> will prevent Clang from printing
284the [<a href="#opt_Wextra-tokens">-Wextra-tokens</a>] information in the
285diagnostic. This information tells you the flag needed to enable or disable the
286diagnostic, either from the command line or through <a
287href="#pragma_GCC_diagnostic">#pragma GCC diagnostic</a>.</dd>
288
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]289<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
290<dt id="opt_fdiagnostics-show-category"><b>-fdiagnostics-show-category=none/id/name</b>:
291Enable printing category information in diagnostic line.</dt>
292<dd>This option, which defaults to "none",
293controls whether or not Clang prints the category associated with a diagnostic
294when emitting it. Each diagnostic may or many not have an associated category,
295if it has one, it is listed in the diagnostic categorization field of the
296diagnostic line (in the []'s).</p>
297
298<p>For example, a format string warning will produce these three renditions
299based on the setting of this option:</p>
300
301<pre>
302 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat]
303 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,1</b>]
304 t.c:3:11: warning: conversion specifies type 'char *' but the argument has type 'int' [-Wformat<b>,Format String</b>]
305</pre>
306
307<p>This category can be used by clients that want to group diagnostics by
308category, so it should be a high level category. We want dozens of these, not
309hundreds or thousands of them.</p>
310</dd>
311
312
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]313
314<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
315<dt id="opt_fdiagnostics-fixit-info"><b>-f[no-]diagnostics-fixit-info</b>:
316Enable "FixIt" information in the diagnostics output.</dt>
317<dd>This option, which defaults to on, controls whether or not Clang prints the
318information on how to fix a specific diagnostic underneath it when it knows.
319For example, in this output:</p>
320
321<pre>
322 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
323 #endif bad
324 ^
325 //
326</pre>
327
328<p>Passing <b>-fno-diagnostics-fixit-info</b> will prevent Clang from printing
329the "//" line at the end of the message. This information is useful for users
330who may not understand what is wrong, but can be confusing for machine
331parsing.</p>
332</dd>
333
334<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]335<dt id="opt_fdiagnostics-print-source-range-info">
336<b>-f[no-]diagnostics-print-source-range-info</b>:
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]337Print machine parsable information about source ranges.</dt>
338<dd>This option, which defaults to off, controls whether or not Clang prints
339information about source ranges in a machine parsable format after the
340file/line/column number information. The information is a simple sequence of
341brace enclosed ranges, where each range lists the start and end line/column
342locations. For example, in this output:</p>
343
344<pre>
345exprs.c:47:15:{47:8-47:14}{47:17-47:24}: error: invalid operands to binary expression ('int *' and '_Complex float')
346 P = (P-42) + Gamma*4;
347 ~~~~~~ ^ ~~~~~~~
348</pre>
349
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]350<p>The {}'s are generated by -fdiagnostics-print-source-range-info.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]351</dd>
352
353
354</dl>
355
356
357
358
359<!-- ===================================================== -->
360<h4 id="cl_diag_warning_groups">Individual Warning Groups</h4>
361<!-- ===================================================== -->
362
363<p>TODO: Generate this from tblgen. Define one anchor per warning group.</p>
364
365
366<dl>
367
368
369<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
370<dt id="opt_Wextra-tokens"><b>-Wextra-tokens</b>: Warn about excess tokens at
371 the end of a preprocessor directive.</dt>
372<dd>This option, which defaults to on, enables warnings about extra tokens at
373the end of preprocessor directives. For example:</p>
374
375<pre>
376 test.c:28:8: warning: extra tokens at end of #endif directive [-Wextra-tokens]
377 #endif bad
378 ^
379</pre>
380
381<p>These extra tokens are not strictly conforming, and are usually best handled
382by commenting them out.</p>
383
384<p>This option is also enabled by <a href="">-Wfoo</a>, <a href="">-Wbar</a>,
385 and <a href="">-Wbaz</a>.</p>
386</dd>
387
388</dl>
389
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]390<!-- ======================================================================= -->
391<h2 id="general_features">Language and Target-Independent Features</h2>
392<!-- ======================================================================= -->
393
394
395<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
396<h3 id="diagnostics">Controlling Errors and Warnings</h3>
397<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
398
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]399<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]400emit errors and warning messages, and how they are displayed to the console.</p>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]401
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]402<h4 id="diagnostics_display">Controlling How Clang Displays Diagnostics</h4>
Chris Lattner65a795b2009-04-20 06:00:23 +0000[diff] [blame]403
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]404<p>When Clang emits a diagnostic, it includes rich information in the output,
405and gives you fine-grain control over which information is printed. Clang has
406the ability to print this information, and these are the options that control
407it:</p>
408
409<p>
410<ol>
411<li>A file/line/column indicator that shows exactly where the diagnostic occurs
412 in your code [<a href="#opt_fshow-column">-fshow-column</a>, <a
413 href="#opt_fshow-source-location">-fshow-source-location</a>].</li>
414<li>A categorization of the diagnostic as a note, warning, error, or fatal
415 error.</li>
416<li>A text string that describes what the problem is.</li>
417<li>An option that indicates how to control the diagnostic (for diagnostics that
418 support it) [<a
419 href="#opt_fdiagnostics-show-option">-fdiagnostics-show-option</a>].</li>
Chris Lattner28a43a42010-05-05 01:35:28 +0000[diff] [blame]420<li>A high-level category for the diagnostic for clients that want to group
421 diagnostics by class (for diagnostics that
422 support it) [<a
423 href="#opt_fdiagnostics-show-category">-fdiagnostics-show-category</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]424<li>The line of source code that the issue occurs on, along with a caret and
425 ranges that indicate the important locations [<a
426 href="opt_fcaret-diagnostics">-fcaret-diagnostics</a>].</li>
427<li>"FixIt" information, which is a concise explanation of how to fix the
428 problem (when Clang is certain it knows) [<a
429 href="opt_fdiagnostics-fixit-info">-fdiagnostics-fixit-info</a>].</li>
430<li>A machine-parsable representation of the ranges involved (off by
431 default) [<a
Chris Lattner2a9cc232009-04-21 05:35:32 +0000[diff] [blame]432 href="opt_fdiagnostics-print-source-range-info">-fdiagnostics-print-source-range-info</a>].</li>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]433</ol></p>
434
435<p>For more information please see <a href="#cl_diag_formatting">Formatting of
436Diagnostics</a>.</p>
437
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]438<h4 id="diagnostics_mappings">Diagnostic Mappings</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]439
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]440<p>All diagnostics are mapped into one of these 5 classes:</p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]441
442<p>
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]443<ul>
444<li>Ignored</li>
445<li>Note</li>
446<li>Warning</li>
447<li>Error</li>
448<li>Fatal</li>
449</ul></p>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]450
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]451<h4 id="diagnostics_commandline">Controlling Diagnostics via Command Line Flags</h4>
Chris Lattner8217f4e2009-04-20 06:26:18 +0000[diff] [blame]452
453<p>-W flags, -pedantic, etc</p>
454
Chris Lattner3401cf82009-07-12 21:22:10 +0000[diff] [blame]455<h4 id="diagnostics_pragmas">Controlling Diagnostics via Pragmas</h4>
456
457<p>Clang can also control what diagnostics are enabled through the use of
458pragmas in the source code. This is useful for turning off specific warnings
459in a section of source code. Clang supports GCC's pragma for compatibility
460with existing source code, as well as several extensions. </p>
461
462<p>The pragma may control any warning that can be used from the command line.
463Warnings may be set to ignored, warning, error, or fatal. The following
464example code will tell Clang or GCC to ignore the -Wall warnings:</p>
465
466<pre>
467#pragma GCC diagnostic ignored "-Wall"
468</pre>
469
470<p>In addition to all of the functionality of provided by GCC's pragma, Clang
471also allows you to push and pop the current warning state. This is particularly
472useful when writing a header file that will be compiled by other people, because
473you don't know what warning flags they build with.</p>
474
475<p>In the below example
476-Wmultichar is ignored for only a single line of code, after which the
477diagnostics return to whatever state had previously existed.</p>
478
479<pre>
480#pragma clang diagnostic push
481#pragma clang diagnostic ignored "-Wmultichar"
482
483char b = 'df'; // no warning.
484
485#pragma clang diagnostic pop
486</pre>
487
488<p>The push and pop pragmas will save and restore the full diagnostic state of
489the compiler, regardless of how it was set. That means that it is possible to
490use push and pop around GCC compatible diagnostics and Clang will push and pop
491them appropriately, while GCC will ignore the pushes and pops as unknown
492pragmas. It should be noted that while Clang supports the GCC pragma, Clang and
493GCC do not support the exact same set of warnings, so even when using GCC
494compatible #pragmas there is no guarantee that they will have identical behaviour
495on both compilers. </p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]496
497<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
498<h3 id="precompiledheaders">Precompiled Headers</h3>
499<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
500
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]501<p><a href="http://en.wikipedia.org/wiki/Precompiled_header">Precompiled
502headers</a> are a general approach employed by many compilers to reduce
503compilation time. The underlying motivation of the approach is that it is
504common for the same (and often large) header files to be included by
505multiple source files. Consequently, compile times can often be greatly improved
506by caching some of the (redundant) work done by a compiler to process headers.
507Precompiled header files, which represent one of many ways to implement
508this optimization, are literally files that represent an on-disk cache that
509contains the vital information necessary to reduce some of the work
510needed to process a corresponding header file. While details of precompiled
511headers vary between compilers, precompiled headers have been shown to be a
512highly effective at speeding up program compilation on systems with very large
513system headers (e.g., Mac OS/X).</p>
514
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]515<h4>Generating a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]516
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]517<p>To generate a PCH file using Clang, one invokes Clang with
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]518the <b><tt>-x <i>&lt;language&gt;</i>-header</tt></b> option. This mirrors the
519interface in GCC for generating PCH files:</p>
520
521<pre>
522 $ gcc -x c-header test.h -o test.h.gch
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]523 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]524</pre>
525
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]526<h4>Using a PCH File</h4>
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]527
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]528<p>A PCH file can then be used as a prefix header when a
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]529<b><tt>-include</tt></b> option is passed to <tt>clang</tt>:</p>
530
531<pre>
532 $ clang -include test.h test.c -o test
533</pre>
534
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]535<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]536is 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]537will be processed from the PCH file. Otherwise, Clang falls back to
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]538directly processing the content of <tt>test.h</tt>. This mirrors the behavior of
539GCC.</p>
540
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]541<p><b>NOTE:</b> Clang does <em>not</em> automatically use PCH files
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]542for headers that are directly included within a source file. For example:</p>
543
544<pre>
Chris Lattnere42ec542009-06-13 20:35:58 +0000[diff] [blame]545 $ clang -x c-header test.h -o test.h.pch
Chris Lattner5c3074f2009-04-20 04:37:38 +0000[diff] [blame]546 $ cat test.c
547 #include "test.h"
548 $ clang test.c -o test
549</pre>
550
Douglas Gregorf4d59532009-06-03 22:37:00 +0000[diff] [blame]551<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]552<tt>test.h</tt> since <tt>test.h</tt> was included directly in the source file
553and not specified on the command line using <tt>-include</tt>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]554
Douglas Gregore650c8c2009-07-07 00:12:59 +0000[diff] [blame]555<h4>Relocatable PCH Files</h4>
556<p>It is sometimes necessary to build a precompiled header from headers that
557are not yet in their final, installed locations. For example, one might build a
558precompiled header within the build tree that is then meant to be installed
559alongside the headers. Clang permits the creation of "relocatable" precompiled
560headers, which are built with a given path (into the build directory) and can
561later be used from an installed location.</p>
562
563<p>To build a relocatable precompiled header, place your headers into a
564subdirectory whose structure mimics the installed location. For example, if you
565want to build a precompiled header for the header <code>mylib.h</code> that
566will be installed into <code>/usr/include</code>, create a subdirectory
567<code>build/usr/include</code> and place the header <code>mylib.h</code> into
568that subdirectory. If <code>mylib.h</code> depends on other headers, then
569they can be stored within <code>build/usr/include</code> in a way that mimics
570the installed location.</p>
571
572<p>Building a relocatable precompiled header requires two additional arguments.
573First, pass the <code>--relocatable-pch</code> flag to indicate that the
574resulting PCH file should be relocatable. Second, pass
575<code>-isysroot /path/to/build</code>, which makes all includes for your
576library relative to the build directory. For example:</p>
577
578<pre>
579 # clang -x c-header --relocatable-pch -isysroot /path/to/build /path/to/build/mylib.h mylib.h.pch
580</pre>
581
582<p>When loading the relocatable PCH file, the various headers used in the PCH
583file are found from the system header root. For example, <code>mylib.h</code>
584can be found in <code>/usr/include/mylib.h</code>. If the headers are installed
585in some other system root, the <code>-isysroot</code> option can be used provide
586a different system root from which the headers will be based. For example,
587<code>-isysroot /Developer/SDKs/MacOSX10.4u.sdk</code> will look for
588<code>mylib.h</code> in
589<code>/Developer/SDKs/MacOSX10.4u.sdk/usr/include/mylib.h</code>.</p>
590
591<p>Relocatable precompiled headers are intended to be used in a limited number
592of cases where the compilation environment is tightly controlled and the
593precompiled header cannot be generated after headers have been installed.
594Relocatable precompiled headers also have some performance impact, because
595the difference in location between the header locations at PCH build time vs.
596at the time of PCH use requires one of the PCH optimizations,
597<code>stat()</code> caching, to be disabled. However, this change is only
598likely to affect PCH files that reference a large number of headers.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]599
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]600<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
601<h3 id="codegen">Controlling Code Generation</h3>
602<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
603
604<p>Clang provides a number of ways to control code generation. The options are listed below.</p>
605
606<!-- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -->
607<dt id="opt_fcatch-undefined-behavior"><b>-fcatch-undefined-behavior</b>: Turn
608on runtime code generation to check for undefined behavior.</dt>
609
610<dd>This option, which defaults to off, controls whether or not Clang
611adds runtime checks for undefined runtime behavior. If the check fails,
612<tt>__builtin_trap()</tt> is used to indicate failure.
613The checks are:
614<p>
Mike Stump88b2a172009-12-16 03:25:12 +0000[diff] [blame]615<li>Subscripting where the static type of one operand is variable
616 which is decayed from an array type and the other operand is
617 greater than the size of the array or less than zero.</li>
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]618<li>Shift operators where the amount shifted is greater or equal to the
619 promoted bit-width of the left-hand-side or less than zero.</li>
Mike Stump8f6a3ed2009-12-16 03:18:14 +0000[diff] [blame]620<li>If control flow reaches __builtin_unreachable.
621<li>When llvm implements more __builtin_object_size support, reads and
622 writes for objects that __builtin_object_size indicates we aren't
623 accessing valid memory. Bit-fields and vectors are not yet checked.
Mike Stump53664ca2009-12-14 23:53:10 +0000[diff] [blame]624</p>
625</dd>
626
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]627<dt id="opt_fno-assume-sane-operator-new"><b>-fno-assume-sane-operator-new</b>:
628Don't assume that the C++'s new operator is sane.</dt>
Nuno Lopesb23f20d2009-12-17 10:15:49 +0000[diff] [blame]629<dd>This option tells the compiler to do not assume that C++'s global new
630operator will always return a pointer that do not
Nuno Lopesaa526242009-12-17 10:00:52 +0000[diff] [blame]631alias any other pointer when the function returns.</dd>
632
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]633<!-- ======================================================================= -->
634<h2 id="c">C Language Features</h2>
635<!-- ======================================================================= -->
636
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]637<p>The support for standard C in clang is feature-complete except for the C99
638floating-point pragmas.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]639
640<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]641<h3 id="c_ext">Extensions supported by clang</h3>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]642<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
643
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]644<p>See <a href="LanguageExtensions.html">clang language extensions</a>.</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]645
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]646<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
647<h3 id="c_modes">Differences between various standard modes</h3>
648<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]649
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]650<p>clang supports the -std option, which changes what language mode clang uses.
651The supported modes for C are c89, gnu89, c94, c99, gnu99 and various aliases
652for those modes. If no -std option is specified, clang defaults to gnu99 mode.
653</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]654
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]655<p>Differences between all c* and gnu* modes:</p>
656<ul>
657<li>c* modes define "__STRICT_ANSI__".</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]658<li>Target-specific defines not prefixed by underscores, like "linux", are
659defined in gnu* modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]660<li>Trigraphs default to being off in gnu* modes; they can be enabled by the
661-trigraphs option.</li>
662<li>The parser recognizes "asm" and "typeof" as keywords in gnu* modes; the
663variants "__asm__" and "__typeof__" are recognized in all modes.</li>
Eli Friedmanb0ac5452009-05-16 23:17:30 +0000[diff] [blame]664<li>The Apple "blocks" extension is recognized by default in gnu* modes
665on some platforms; it can be enabled in any mode with the "-fblocks"
666option.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]667</ul>
668
669<p>Differences between *89 and *99 modes:</p>
670<ul>
671<li>The *99 modes default to implementing "inline" as specified in C99, while
672the *89 modes implement the GNU version. This can be overridden for individual
673functions with the __gnu_inline__ attribute.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]674<li>Digraphs are not recognized in c89 mode.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]675<li>The scope of names defined inside a "for", "if", "switch", "while", or "do"
676statement is different. (example: "if ((struct x {int x;}*)0) {}".)</li>
677<li>__STDC_VERSION__ is not defined in *89 modes.</li>
Eli Friedman26fa0ed2009-05-27 23:02:57 +0000[diff] [blame]678<li>"inline" is not recognized as a keyword in c89 mode.</li>
679<li>"restrict" is not recognized as a keyword in *89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]680<li>Commas are allowed in integer constant expressions in *99 modes.</li>
681<li>Arrays which are not lvalues are not implicitly promoted to pointers in
682*89 modes.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]683<li>Some warnings are different.</li>
684</ul>
685
686<p>c94 mode is identical to c89 mode except that digraphs are enabled in
687c94 mode (FIXME: And __STDC_VERSION__ should be defined!).</p>
688
689<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
690<h3 id="c_unimpl_gcc">GCC extensions not implemented yet</h3>
691<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
692
693<p>clang tries to be compatible with gcc as much as possible, but some gcc
694extensions are not implemented yet:</p>
695
696<ul>
697<li>clang does not support __label__
698(<a href="http://llvm.org/bugs/show_bug.cgi?id=3429">bug 3429</a>). This is
699a relatively small feature, so it is likely to be implemented relatively
700soon.</li>
701
702<li>clang does not support attributes on function pointers
703(<a href="http://llvm.org/bugs/show_bug.cgi?id=2461">bug 2461</a>). This is
704a relatively important feature, so it is likely to be implemented relatively
705soon.</li>
706
707<li>clang does not support #pragma weak
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]708(<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]709the uses described in the bug, this is likely to be implemented at some
710point, at least partially.</li>
711
712<li>clang does not support #pragma align
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]713(<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]714relatively small feature, so it is likely to be implemented relatively
715soon.</li>
716
Eli Friedman4da92552009-06-02 08:21:31 +0000[diff] [blame]717<li>clang does not support code generation for local variables pinned to
718registers (<a href="http://llvm.org/bugs/show_bug.cgi?id=3933">bug 3933</a>).
719This is a relatively small feature, so it is likely to be implemented
720relatively soon.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]721
722<li>clang does not support decimal floating point types (_Decimal32 and
723friends) or fixed-point types (_Fract and friends); nobody has expressed
724interest in these features yet, so it's hard to say when they will be
725implemented.</li>
726
727<li>clang does not support nested functions; this is a complex feature which
728is infrequently used, so it is unlikely to be implemented anytime soon.</li>
729
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]730<li>clang does not support global register variables, this is unlikely
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]731to be implemented soon because it requires additional LLVM backend support.
732</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]733
734<li>clang does not support static initialization of flexible array
735members. This appears to be a rarely used extension, but could be
736implemented pending user demand.</li>
737
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]738<li>clang does not support __builtin_va_arg_pack/__builtin_va_arg_pack_len.
739This is used rarely, but in some potentially interesting places, like the
740glibc headers, so it may be implemented pending user demand. Note that
741because clang pretends to be like GCC 4.2, and this extension was introduced
742in 4.3, the glibc headers will not try to use this extension with clang at
743the moment.</li>
744
745<li>clang does not support the gcc extension for forward-declaring function
746parameters; this has not showed up in any real-world code yet, though, so it
747might never be implemented.</li>
748
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]749</ul>
750
751<p>This is not a complete list; if you find an unsupported extension
752missing from this list, please send an e-mail to cfe-dev. This list
753currently excludes C++; see <a href="#cxx">C++ Language Features</a>.
754Also, this list does not include bugs in mostly-implemented features; please
755see 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">
756bug tracker</a> for known existing bugs (FIXME: Is there a section for
757bug-reporting guidelines somewhere?).</p>
758
759<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
760<h3 id="c_unsupp_gcc">Intentionally unsupported GCC extensions</h3>
761<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
762
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]763<ul>
764
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]765<li>clang does not support the gcc extension that allows variable-length arrays
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]766in structures. This is for a few of reasons: one, it is tricky
767to implement, two, the extension is completely undocumented, and three, the
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]768extension appears to be rarely used.</li>
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]769
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]770<li>clang does not support duplicate definitions of a function where one is
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]771inline. This complicates clients of the AST which normally can expect there is
772at most one definition for each function. Source code using this feature should
773be changed to define the inline and out-of-line definitions in separate
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]774translation units.</li>
Daniel Dunbarc5f928c2009-06-04 18:37:52 +0000[diff] [blame]775
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]776<li>clang does not have an equivalent to gcc's "fold"; this means that
777clang doesn't accept some constructs gcc might accept in contexts where a
778constant expression is required, like "x-x" where x is a variable, or calls
779to C library functions like strlen.</li>
780
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]781<li>clang does not support multiple alternative constraints in inline asm; this
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]782is an extremely obscure feature which would be complicated to implement
Eli Friedman2f3fca02009-06-12 20:08:48 +0000[diff] [blame]783correctly.</li>
Eli Friedmanca130bd2009-06-12 20:04:25 +0000[diff] [blame]784
785<li>clang does not support __builtin_apply and friends; this extension is
786extremely obscure and difficult to implement reliably.</li>
787
Eli Friedman0b326002009-06-12 20:11:05 +0000[diff] [blame]788</ul>
789
Eli Friedman660a5ac2009-04-28 18:48:34 +0000[diff] [blame]790<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
791<h3 id="c_ms">Microsoft extensions</h3>
792<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
793
794<p>clang has some experimental support for extensions from
795Microsoft Visual C++; to enable it, use the -fms-extensions command-line
Eli Friedmana1821b52009-06-08 06:21:03 +0000[diff] [blame]796option. This is the default for Windows targets. Note that the
797support is incomplete; enabling Microsoft extensions will silently drop
798certain constructs (including __declspec and Microsoft-style asm statements).
799</p>
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]800
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]801<li>clang does not support the Microsoft extension where anonymous
Daniel Dunbar869e0542009-05-13 00:23:53 +0000[diff] [blame]802record members can be declared using user defined typedefs.</li>
Daniel Dunbar05fa6292009-05-12 23:12:07 +0000[diff] [blame]803
Daniel Dunbar9375ed12009-05-13 21:40:49 +0000[diff] [blame]804<li>clang supports the Microsoft "#pragma pack" feature for
805controlling record layout. GCC also contains support for this feature,
806however where MSVC and GCC are incompatible clang follows the MSVC
807definition.</li>
808
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]809<!-- ======================================================================= -->
810<h2 id="objc">Objective-C Language Features</h2>
811<!-- ======================================================================= -->
812
813
814<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
815<h3 id="objc_incompatibilities">Intentional Incompatibilities with GCC</h3>
816<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
817
818<p>No cast of super, no lvalue casts.</p>
819
820
821
822<!-- ======================================================================= -->
823<h2 id="cxx">C++ Language Features</h2>
824<!-- ======================================================================= -->
825
Douglas Gregor57f1a002010-02-05 23:51:14 +0000[diff] [blame]826<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]827is under active development and is progressing rapidly. Please see the <a
828href="http://clang.llvm.org/cxx_status.html">C++ Status</a> page for details or
829ask on the mailing list about how you can help.</p>
830
Douglas Gregor57f1a002010-02-05 23:51:14 +0000[diff] [blame]831<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]832
Rafael Espindola9b2fc952010-03-17 04:31:53 +0000[diff] [blame]833<p>Clang strives to strictly conform to the C++ standard. That means
834it will reject invalid C++ code that another compiler may accept. If
835Clang reports errors in your code, please check
836the <a href="http://clang.llvm.org/cxx_compatibility.html">C++
837Compatibility</a> page to see whether they are C++-conformance bugs
838and how you can fix them.</p>
839
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]840<!-- ======================================================================= -->
841<h2 id="objcxx">Objective C++ Language Features</h2>
842<!-- ======================================================================= -->
843
844<p>At this point, Clang C++ support is not generally useful (and therefore,
845neither is Objective-C++). Please see the <a href="#cxx">C++ section</a> for
846more information.</p>
847
848<!-- ======================================================================= -->
849<h2 id="target_features">Target-Specific Features and Limitations</h2>
850<!-- ======================================================================= -->
851
852
853<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
854<h3 id="target_arch">CPU Architectures Features and Limitations</h3>
855<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
856
857<!-- ======================== -->
858<h4 id="target_arch_x86">X86</h4>
859<!-- ======================== -->
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]860<p>The support for X86 (both 32-bit and 64-bit) is considered stable
861on Darwin (Mac OS/X), Linux, FreeBSD, and Dragonfly BSD: it has been tested to
862correctly compile large C and Objective-C codebases. (FIXME: Anything specific
863we want to say here? Possibly mention some LLVM x86 limitations?)
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]864
Eli Friedman3b658d32009-06-08 05:12:39 +0000[diff] [blame]865<!-- ======================== -->
866<h4 id="target_arch_arm">ARM</h4>
867<!-- ======================== -->
868ARM support is mostly feature-complete, but still experimental; it hasn't
869undergone significant testing.
870
871<!-- ======================== -->
872<h4 id="target_arch_other">Other platforms</h4>
873<!-- ======================== -->
874clang currently contains some support for PPC and Sparc; however, significant
875pieces of code generation are still missing, and they haven't undergone
876significant testing.
877
878<p>clang contains some support for the embedded PIC16 processor
879(FIXME: I haven't been keeping track of this; what should this say?).
880
881<p>clang contains limited support for the MSP430 embedded processor, but both
882the clang support and the LLVM backend support are highly experimental.
883
884<p>Other platforms are completely unsupported at the moment. Adding the
885minimal support needed for parsing and semantic analysis on a new platform
886is quite easy; see lib/Basic/Targets.cpp in the clang source tree. This level
887of support is also sufficient for conversion to LLVM IR for simple programs.
888Proper support for conversion to LLVM IR requires adding code to
889lib/CodeGen/CGCall.cpp at the moment; this is likely to change soon, though.
890Generating assembly requires a suitable LLVM backend.
Chris Lattnercf17d9d2009-04-20 04:23:09 +0000[diff] [blame]891
892<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
893<h3 id="target_os">Operating System Features and Limitations</h3>
894<!-- = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = -->
895
896<!-- ======================================= -->
897<h4 id="target_os_darwin">Darwin (Mac OS/X)</h4>
898<!-- ======================================= -->
899
900<p>No __thread support, 64-bit ObjC support requires SL tools.</p>
901
902</div>
903</body>
904</html>