Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 4dbc9d5907bc065ba0af52ef56ae97b63a4f49a5 / . / docs / CodingStandards.html
blob: 90795aa0ba43c8b3bfec26a9ddbd3b393ba6f936 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <link rel="stylesheet" href="llvm.css" type="text/css">
Misha Brukman4dbc9d52008-12-10 23:07:02 +0000[diff] [blame^]6 <title>LLVM Coding Standards</title>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7</head>
8<body>
9
10<div class="doc_title">
Misha Brukman4dbc9d52008-12-10 23:07:02 +0000[diff] [blame^]11 LLVM Coding Standards
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]12</div>
13
14<ol>
15 <li><a href="#introduction">Introduction</a></li>
16 <li><a href="#mechanicalissues">Mechanical Source Issues</a>
17 <ol>
18 <li><a href="#sourceformating">Source Code Formatting</a>
19 <ol>
20 <li><a href="#scf_commenting">Commenting</a></li>
21 <li><a href="#scf_commentformat">Comment Formatting</a></li>
22 <li><a href="#scf_includes"><tt>#include</tt> Style</a></li>
23 <li><a href="#scf_codewidth">Source Code Width</a></li>
24 <li><a href="#scf_spacestabs">Use Spaces Instead of Tabs</a></li>
25 <li><a href="#scf_indentation">Indent Code Consistently</a></li>
26 </ol></li>
27 <li><a href="#compilerissues">Compiler Issues</a>
28 <ol>
29 <li><a href="#ci_warningerrors">Treat Compiler Warnings Like
30 Errors</a></li>
31 <li><a href="#ci_portable_code">Write Portable Code</a></li>
32 <li><a href="#ci_class_struct">Use of class/struct Keywords</a></li>
33 </ol></li>
34 </ol></li>
35 <li><a href="#styleissues">Style Issues</a>
36 <ol>
37 <li><a href="#macro">The High Level Issues</a>
38 <ol>
39 <li><a href="#hl_module">A Public Header File <b>is</b> a
40 Module</a></li>
41 <li><a href="#hl_dontinclude">#include as Little as Possible</a></li>
42 <li><a href="#hl_privateheaders">Keep "internal" Headers
43 Private</a></li>
44 <li><a href="#ll_iostream"><tt>#include &lt;iostream&gt;</tt> is
45 <em>forbidden</em></a></li>
46 </ol></li>
47 <li><a href="#micro">The Low Level Issues</a>
48 <ol>
49 <li><a href="#ll_assert">Assert Liberally</a></li>
50 <li><a href="#ll_ns_std">Do not use 'using namespace std'</a></li>
51 <li><a href="#ll_virtual_anch">Provide a virtual method anchor for
52 classes in headers</a></li>
53 <li><a href="#ll_preincrement">Prefer Preincrement</a></li>
54 <li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
55 </ol></li>
56 </ol></li>
57 <li><a href="#seealso">See Also</a></li>
58</ol>
59
60<div class="doc_author">
61 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
62 <a href="mailto:void@nondot.org">Bill Wendling</a></p>
63</div>
64
65
66<!-- *********************************************************************** -->
67<div class="doc_section">
68 <a name="introduction">Introduction</a>
69</div>
70<!-- *********************************************************************** -->
71
72<div class="doc_text">
73
74<p>This document attempts to describe a few coding standards that are being used
75in the LLVM source tree. Although no coding standards should be regarded as
76absolute requirements to be followed in all instances, coding standards can be
77useful.</p>
78
79<p>This document intentionally does not prescribe fixed standards for religious
80issues such as brace placement and space usage. For issues like this, follow
81the golden rule:</p>
82
83<blockquote>
84
85<p><b><a name="goldenrule">If you are adding a significant body of source to a
86project, feel free to use whatever style you are most comfortable with. If you
87are extending, enhancing, or bug fixing already implemented code, use the style
88that is already being used so that the source is uniform and easy to
89follow.</a></b></p>
90
91</blockquote>
92
93<p>The ultimate goal of these guidelines is the increase readability and
94maintainability of our common source base. If you have suggestions for topics to
95be included, please mail them to <a
96href="mailto:sabre@nondot.org">Chris</a>.</p>
97
98</div>
99
100<!-- *********************************************************************** -->
101<div class="doc_section">
102 <a name="mechanicalissues">Mechanical Source Issues</a>
103</div>
104<!-- *********************************************************************** -->
105
106<!-- ======================================================================= -->
107<div class="doc_subsection">
108 <a name="sourceformating">Source Code Formatting</a>
109</div>
110
111<!-- _______________________________________________________________________ -->
112<div class="doc_subsubsection">
113 <a name="scf_commenting">Commenting</a>
114</div>
115
116<div class="doc_text">
117
118<p>Comments are one critical part of readability and maintainability. Everyone
119knows they should comment, so should you. Although we all should probably
120comment our code more than we do, there are a few very critical places that
121documentation is very useful:</p>
122
123<b>File Headers</b>
124
125<p>Every source file should have a header on it that describes the basic
126purpose of the file. If a file does not have a header, it should not be
127checked into Subversion. Most source trees will probably have a standard
128file header format. The standard format for the LLVM source tree looks like
129this:</p>
130
131<div class="doc_code">
132<pre>
133//===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
134//
135// The LLVM Compiler Infrastructure
136//
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]137// This file is distributed under the University of Illinois Open Source
138// License. See LICENSE.TXT for details.
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]139//
140//===----------------------------------------------------------------------===//
141//
142// This file contains the declaration of the Instruction class, which is the
143// base class for all of the VM instructions.
144//
145//===----------------------------------------------------------------------===//
146</pre>
147</div>
148
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]149<p>A few things to note about this particular format: The "<tt>-*- C++
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]150-*-</tt>" string on the first line is there to tell Emacs that the source file
151is a C++ file, not a C file (Emacs assumes .h files are C files by default).
152Note that this tag is not necessary in .cpp files. The name of the file is also
153on the first line, along with a very short description of the purpose of the
154file. This is important when printing out code and flipping though lots of
155pages.</p>
156
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]157<p>The next section in the file is a concise note that defines the license
158that the file is released under. This makes it perfectly clear what terms the
159source code can be distributed under and should not be modified in any way.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]160
161<p>The main body of the description does not have to be very long in most cases.
162Here it's only two lines. If an algorithm is being implemented or something
163tricky is going on, a reference to the paper where it is published should be
164included, as well as any notes or "gotchas" in the code to watch out for.</p>
165
166<b>Class overviews</b>
167
168<p>Classes are one fundamental part of a good object oriented design. As such,
169a class definition should have a comment block that explains what the class is
170used for... if it's not obvious. If it's so completely obvious your grandma
171could figure it out, it's probably safe to leave it out. Naming classes
172something sane goes a long ways towards avoiding writing documentation.</p>
173
174
175<b>Method information</b>
176
177<p>Methods defined in a class (as well as any global functions) should also be
178documented properly. A quick note about what it does any a description of the
179borderline behaviour is all that is necessary here (unless something
180particularly tricky or insideous is going on). The hope is that people can
181figure out how to use your interfaces without reading the code itself... that is
182the goal metric.</p>
183
184<p>Good things to talk about here are what happens when something unexpected
185happens: does the method return null? Abort? Format your hard disk?</p>
186
187</div>
188
189<!-- _______________________________________________________________________ -->
190<div class="doc_subsubsection">
191 <a name="scf_commentformat">Comment Formatting</a>
192</div>
193
194<div class="doc_text">
195
196<p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
197require less typing, don't have nesting problems, etc. There are a few cases
198when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
199
200<ol>
201 <li>When writing a C code: Obviously if you are writing C code, use C style
202 comments.</li>
203 <li>When writing a header file that may be <tt>#include</tt>d by a C source
204 file.</li>
205 <li>When writing a source file that is used by a tool that only accepts C
206 style comments.</li>
207</ol>
208
209<p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
210These nest properly and are better behaved in general than C style comments.</p>
211
212</div>
213
214<!-- _______________________________________________________________________ -->
215<div class="doc_subsubsection">
216 <a name="scf_includes"><tt>#include</tt> Style</a>
217</div>
218
219<div class="doc_text">
220
221<p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
222include guards if working on a header file), the <a
223href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
224file should be listed. We prefer these <tt>#include</tt>s to be listed in this
225order:</p>
226
227<ol>
228 <li><a href="#mmheader">Main Module header</a></li>
229 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
230 <li><tt>llvm/*</tt></li>
231 <li><tt>llvm/Analysis/*</tt></li>
232 <li><tt>llvm/Assembly/*</tt></li>
233 <li><tt>llvm/Bytecode/*</tt></li>
234 <li><tt>llvm/CodeGen/*</tt></li>
235 <li>...</li>
236 <li><tt>Support/*</tt></li>
237 <li><tt>Config/*</tt></li>
238 <li>System <tt>#includes</tt></li>
239</ol>
240
Nick Lewycky61819f72008-11-29 20:13:25 +0000[diff] [blame]241<p>... and each category should be sorted by name.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]242
243<p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
244which implement an interface defined by a .h file. This <tt>#include</tt>
245should always be included <b>first</b> regardless of where it lives on the file
246system. By including a header file first in the .cpp files that implement the
247interfaces, we ensure that the header does not have any hidden dependencies
248which are not explicitly #included in the header, but should be. It is also a
249form of documentation in the .cpp file to indicate where the interfaces it
250implements are defined.</p>
251
252</div>
253
254<!-- _______________________________________________________________________ -->
255<div class="doc_subsubsection">
256 <a name="scf_codewidth">Source Code Width</a>
257</div>
258
259<div class="doc_text">
260
261<p>Write your code to fit within 80 columns of text. This helps those of us who
262like to print out code and look at your code in an xterm without resizing
263it.</p>
264
Chris Lattnerb5daf702008-07-08 05:12:37 +0000[diff] [blame]265<p>The longer answer is that there must be some limit to the width of the code
266in order to reasonably allow developers to have multiple files side-by-side in
267windows on a modest display. If you are going to pick a width limit, it is
268somewhat arbitrary but you might as well pick something standard. Going with
26990 columns (for example) instead of 80 columns wouldn't add any significant
270value and would be detrimental to printing out code. Also many other projects
271have standardized on 80 columns, so some people have already configured their
272editors for it (vs something else, like 90 columns).</p>
273
274<p>This is one of many contentious issues in coding standards, but is not up
275for debate.</p>
276
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]277</div>
278
279<!-- _______________________________________________________________________ -->
280<div class="doc_subsubsection">
281 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
282</div>
283
284<div class="doc_text">
285
286<p>In all cases, prefer spaces to tabs in source files. People have different
287prefered indentation levels, and different styles of indentation that they
288like... this is fine. What isn't is that different editors/viewers expand tabs
289out to different tab stops. This can cause your code to look completely
290unreadable, and it is not worth dealing with.</p>
291
292<p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
293style of existing code if your are modifying and extending it. If you like four
294spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
295with two spaces of indentation. Also, do not reindent a whole source file: it
296makes for incredible diffs that are absolutely worthless.</p>
297
298</div>
299
300<!-- _______________________________________________________________________ -->
301<div class="doc_subsubsection">
302 <a name="scf_indentation">Indent Code Consistently</a>
303</div>
304
305<div class="doc_text">
306
307<p>Okay, your first year of programming you were told that indentation is
308important. If you didn't believe and internalize this then, now is the time.
309Just do it.</p>
310
311</div>
312
313
314<!-- ======================================================================= -->
315<div class="doc_subsection">
316 <a name="compilerissues">Compiler Issues</a>
317</div>
318
319
320<!-- _______________________________________________________________________ -->
321<div class="doc_subsubsection">
322 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
323</div>
324
325<div class="doc_text">
326
327<p>If your code has compiler warnings in it, something is wrong: you aren't
328casting values correctly, your have "questionable" constructs in your code, or
329you are doing something legitimately wrong. Compiler warnings can cover up
330legitimate errors in output and make dealing with a translation unit
331difficult.</p>
332
333<p>It is not possible to prevent all warnings from all compilers, nor is it
334desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
335a good thorough set of warnings, and stick to them. At least in the case of
336<tt>gcc</tt>, it is possible to work around any spurious errors by changing the
337syntax of the code slightly. For example, an warning that annoys me occurs when
338I write code like this:</p>
339
340<div class="doc_code">
341<pre>
342if (V = getValue()) {
343 ...
344}
345</pre>
346</div>
347
348<p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
349operator, and that I probably mistyped it. In most cases, I haven't, and I
350really don't want the spurious errors. To fix this particular problem, I
351rewrite the code like this:</p>
352
353<div class="doc_code">
354<pre>
355if ((V = getValue())) {
356 ...
357}
358</pre>
359</div>
360
361<p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
362be fixed by massaging the code appropriately.</p>
363
364<p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
365-Winline -W -Wwrite-strings -Wno-unused</tt></p>
366
367</div>
368
369<!-- _______________________________________________________________________ -->
370<div class="doc_subsubsection">
371 <a name="ci_portable_code">Write Portable Code</a>
372</div>
373
374<div class="doc_text">
375
376<p>In almost all cases, it is possible and within reason to write completely
377portable code. If there are cases where it isn't possible to write portable
378code, isolate it behind a well defined (and well documented) interface.</p>
379
380<p>In practice, this means that you shouldn't assume much about the host
381compiler, including its support for "high tech" features like partial
382specialization of templates. In fact, Visual C++ 6 could be an important target
383for our work in the future, and we don't want to have to rewrite all of our code
384to support it.</p>
385
386</div>
387
388<!-- _______________________________________________________________________ -->
389<div class="doc_subsubsection">
390<a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
391</div>
392<div class="doc_text">
393
394<p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
395interchangeably. The only difference is when they are used to declare a class:
396<tt>class</tt> makes all members private by default while <tt>struct</tt> makes
397all members public by default.</p>
398
399<p>Unfortunately, not all compilers follow the rules and some will generate
400different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
401declare the symbol. This can lead to problems at link time.</p>
402
403<p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
404<b>all</b> members are public, in which case <tt>struct</tt> is allowed.</p>
405
406</div>
407
408<!-- *********************************************************************** -->
409<div class="doc_section">
410 <a name="styleissues">Style Issues</a>
411</div>
412<!-- *********************************************************************** -->
413
414
415<!-- ======================================================================= -->
416<div class="doc_subsection">
417 <a name="macro">The High Level Issues</a>
418</div>
419
420
421<!-- _______________________________________________________________________ -->
422<div class="doc_subsubsection">
423 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
424</div>
425
426<div class="doc_text">
427
428<p>C++ doesn't do too well in the modularity department. There is no real
429encapsulation or data hiding (unless you use expensive protocol classes), but it
430is what we have to work with. When you write a public header file (in the LLVM
431source tree, they live in the top level "include" directory), you are defining a
432module of functionality.</p>
433
434<p>Ideally, modules should be completely independent of each other, and their
435header files should only include the absolute minimum number of headers
436possible. A module is not just a class, a function, or a namespace: <a
437href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
438of these</a> that defines an interface. This interface may be several
439functions, classes or data structures, but the important issue is how they work
440together.</p>
441
442<p>In general, a module should be implemented with one or more <tt>.cpp</tt>
443files. Each of these <tt>.cpp</tt> files should include the header that defines
444their interface first. This ensure that all of the dependences of the module
445header have been properly added to the module header itself, and are not
446implicit. System headers should be included after user headers for a
447translation unit.</p>
448
449</div>
450
451<!-- _______________________________________________________________________ -->
452<div class="doc_subsubsection">
453 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
454</div>
455
456<div class="doc_text">
457
458<p><tt>#include</tt> hurts compile time performance. Don't do it unless you
459have to, especially in header files.</p>
460
461<p>But wait, sometimes you need to have the definition of a class to use it, or
462to inherit from it. In these cases go ahead and <tt>#include</tt> that header
463file. Be aware however that there are many cases where you don't need to have
464the full definition of a class. If you are using a pointer or reference to a
465class, you don't need the header file. If you are simply returning a class
466instance from a prototyped function or method, you don't need it. In fact, for
467most cases, you simply don't need the definition of a class... and not
468<tt>#include</tt>'ing speeds up compilation.</p>
469
470<p>It is easy to try to go too overboard on this recommendation, however. You
471<b>must</b> include all of the header files that you are using -- you can
472include them either directly
473or indirectly (through another header file). To make sure that you don't
474accidently forget to include a header file in your module header, make sure to
475include your module header <b>first</b> in the implementation file (as mentioned
476above). This way there won't be any hidden dependencies that you'll find out
477about later...</p>
478
479</div>
480
481<!-- _______________________________________________________________________ -->
482<div class="doc_subsubsection">
483 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
484</div>
485
486<div class="doc_text">
487
488<p>Many modules have a complex implementation that causes them to use more than
489one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
490internal communication interface (helper classes, extra functions, etc) in the
491public module header file. Don't do this.</p>
492
493<p>If you really need to do something like this, put a private header file in
494the same directory as the source files, and include it locally. This ensures
495that your private interface remains private and undisturbed by outsiders.</p>
496
497<p>Note however, that it's okay to put extra implementation methods a public
498class itself... just make them private (or protected), and all is well.</p>
499
500</div>
501
502<!-- _______________________________________________________________________ -->
503<div class="doc_subsubsection">
504 <a name="ll_iostream"><tt>#include &lt;iostream&gt;</tt> is forbidden</a>
505</div>
506
507<div class="doc_text">
508
509<p>The use of <tt>#include &lt;iostream&gt;</tt> in library files is
510hereby <b><em>forbidden</em></b>. The primary reason for doing this is to
511support clients using LLVM libraries as part of larger systems. In particular,
512we statically link LLVM into some dynamic libraries. Even if LLVM isn't used,
513the static c'tors are run whenever an application start up that uses the dynamic
514library. There are two problems with this:</p>
515
516<ol>
517 <li>The time to run the static c'tors impacts startup time of
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]518 applications&mdash;a critical time for GUI apps.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]519 <li>The static c'tors cause the app to pull many extra pages of memory off the
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]520 disk: both the code for the static c'tors in each <tt>.o</tt> file and the
521 small amount of data that gets touched. In addition, touched/dirty pages
522 put more pressure on the VM system on low-memory machines.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]523</ol>
524
Matthijs Kooijmanc4c0af22008-07-30 12:14:10 +0000[diff] [blame]525<p>Note that using the other stream headers (<tt>&lt;sstream&gt;</tt> for
526example) is allowed normally, it is just <tt>&lt;iostream&gt;</tt> that is
527causing problems.</p>
528
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]529<div align="center">
Bill Wendling496e2b92007-09-22 10:03:25 +0000[diff] [blame]530<table>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]531 <tbody>
532 <tr>
533 <th>Old Way</th>
534 <th>New Way</th>
535 </tr>
536 <tr>
537 <td align="left"><pre>#include &lt;iostream&gt;</pre></td>
538 <td align="left"><pre>#include "llvm/Support/Streams.h"</pre></td>
539 </tr>
540 <tr>
541 <td align="left"><pre>DEBUG(std::cerr &lt;&lt; ...);
542DEBUG(dump(std::cerr));</pre></td>
543 <td align="left"><pre>DOUT &lt;&lt; ...;
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]544DEBUG(dump(DOUT));</pre></td>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]545 </tr>
546 <tr>
547 <td align="left"><pre>std::cerr &lt;&lt; "Hello world\n";</pre></td>
548 <td align="left"><pre>llvm::cerr &lt;&lt; "Hello world\n";</pre></td>
549 </tr>
550 <tr>
551 <td align="left"><pre>std::cout &lt;&lt; "Hello world\n";</pre></td>
552 <td align="left"><pre>llvm::cout &lt;&lt; "Hello world\n";</pre></td>
553 </tr>
554 <tr>
555 <td align="left"><pre>std::cin &gt;&gt; Var;</pre></td>
556 <td align="left"><pre>llvm::cin &gt;&gt; Var;</pre></td>
557 </tr>
558 <tr>
559 <td align="left"><pre>std::ostream</pre></td>
560 <td align="left"><pre>llvm::OStream</pre></td>
561 </tr>
562 <tr>
563 <td align="left"><pre>std::istream</pre></td>
564 <td align="left"><pre>llvm::IStream</pre></td>
565 </tr>
566 <tr>
567 <td align="left"><pre>std::stringstream</pre></td>
568 <td align="left"><pre>llvm::StringStream</pre></td>
569 </tr>
570 <tr>
Bill Wendling496e2b92007-09-22 10:03:25 +0000[diff] [blame]571 <td align="left"><pre>void print(std::ostream &amp;Out);
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]572// ...
573print(std::cerr);</pre></td>
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]574 <td align="left"><pre>void print(llvm::OStream Out);<sup>1</sup>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]575// ...
576print(llvm::cerr);</pre>
577
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]578</td> </tbody> </table>
579</div>
580
Bill Wendling496e2b92007-09-22 10:03:25 +0000[diff] [blame]581<div class="doc_text">
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]582<p><sup>1</sup><tt>llvm::OStream</tt> is a light-weight class so it should never
583be passed by reference. This is important because in some configurations,
584<tt>DOUT</tt> is an rvalue.</p>
Bill Wendling496e2b92007-09-22 10:03:25 +0000[diff] [blame]585</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]586
587</div>
588
589
590<!-- ======================================================================= -->
591<div class="doc_subsection">
592 <a name="micro">The Low Level Issues</a>
593</div>
594
595
596<!-- _______________________________________________________________________ -->
597<div class="doc_subsubsection">
598 <a name="ll_assert">Assert Liberally</a>
599</div>
600
601<div class="doc_text">
602
603<p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
604preconditions and assumptions, you never know when a bug (not neccesarily even
605yours) might be caught early by an assertion, which reduces debugging time
606dramatically. The "<tt>&lt;cassert&gt;</tt>" header file is probably already
607included by the header files you are using, so it doesn't cost anything to use
608it.</p>
609
610<p>To further assist with debugging, make sure to put some kind of error message
611in the assertion statement (which is printed if the assertion is tripped). This
612helps the poor debugging make sense of why an assertion is being made and
613enforced, and hopefully what to do about it. Here is one complete example:</p>
614
615<div class="doc_code">
616<pre>
617inline Value *getOperand(unsigned i) {
618 assert(i &lt; Operands.size() &amp;&amp; "getOperand() out of range!");
619 return Operands[i];
620}
621</pre>
622</div>
623
624<p>Here are some examples:</p>
625
626<div class="doc_code">
627<pre>
628assert(Ty-&gt;isPointerType() &amp;&amp; "Can't allocate a non pointer type!");
629
630assert((Opcode == Shl || Opcode == Shr) &amp;&amp; "ShiftInst Opcode invalid!");
631
632assert(idx &lt; getNumSuccessors() &amp;&amp; "Successor # out of range!");
633
634assert(V1.getType() == V2.getType() &amp;&amp; "Constant types must be identical!");
635
636assert(isa&lt;PHINode&gt;(Succ-&gt;front()) &amp;&amp; "Only works on PHId BBs!");
637</pre>
638</div>
639
640<p>You get the idea...</p>
641
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]642<p>Please be aware when adding assert statements that not all compilers are aware of
643the semantics of the assert. In some places, asserts are used to indicate a piece of
644code that should not be reached. These are typically of the form:</p>
645
646<div class="doc_code">
647<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]648assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]649</pre>
650</div>
651
652<p>When used in a function that returns a value, they should be followed with a return
653statement and a comment indicating that this line is never reached. This will prevent
654a compiler which is unable to deduce that the assert statement never returns from
655generating a warning.</p>
656
657<div class="doc_code">
658<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]659assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]660// Not reached
661return 0;
662</pre>
663</div>
664
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]665</div>
666
667<!-- _______________________________________________________________________ -->
668<div class="doc_subsubsection">
669 <a name="ll_ns_std">Do not use '<tt>using namespace std</tt>'</a>
670</div>
671
672<div class="doc_text">
673<p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
674namespace with an "<tt>std::</tt>" prefix, rather than rely on
675"<tt>using namespace std;</tt>".</p>
676
677<p> In header files, adding a '<tt>using namespace XXX</tt>' directive pollutes
678the namespace of any source file that includes the header. This is clearly a
679bad thing.</p>
680
681<p>In implementation files (e.g. .cpp files), the rule is more of a stylistic
682rule, but is still important. Basically, using explicit namespace prefixes
683makes the code <b>clearer</b>, because it is immediately obvious what facilities
684are being used and where they are coming from, and <b>more portable</b>, because
685namespace clashes cannot occur between LLVM code and other namespaces. The
686portability rule is important because different standard library implementations
687expose different symbols (potentially ones they shouldn't), and future revisions
688to the C++ standard will add more symbols to the <tt>std</tt> namespace. As
689such, we never use '<tt>using namespace std;</tt>' in LLVM.</p>
690
691<p>The exception to the general rule (i.e. it's not an exception for
692the <tt>std</tt> namespace) is for implementation files. For example, all of
693the code in the LLVM project implements code that lives in the 'llvm' namespace.
694As such, it is ok, and actually clearer, for the .cpp files to have a '<tt>using
695namespace llvm</tt>' directive at their top, after the <tt>#include</tt>s. The
696general form of this rule is that any .cpp file that implements code in any
697namespace may use that namespace (and its parents'), but should not use any
698others.</p>
699
700</div>
701
702<!-- _______________________________________________________________________ -->
703<div class="doc_subsubsection">
704 <a name="ll_virtual_anch">Provide a virtual method anchor for classes
705 in headers</a>
706</div>
707
708<div class="doc_text">
709
710<p>If a class is defined in a header file and has a v-table (either it has
711virtual methods or it derives from classes with virtual methods), it must
712always have at least one out-of-line virtual method in the class. Without
713this, the compiler will copy the vtable and RTTI into every .o file that
714#includes the header, bloating .o file sizes and increasing link times.
715</p>
716
717</div>
718
719
720<!-- _______________________________________________________________________ -->
721<div class="doc_subsubsection">
722 <a name="ll_preincrement">Prefer Preincrement</a>
723</div>
724
725<div class="doc_text">
726
727<p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
728postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
729preincrementation whenever possible.</p>
730
731<p>The semantics of postincrement include making a copy of the value being
732incremented, returning it, and then preincrementing the "work value". For
733primitive types, this isn't a big deal... but for iterators, it can be a huge
734issue (for example, some iterators contains stack and set objects in them...
735copying an iterator could invoke the copy ctor's of these as well). In general,
736get in the habit of always using preincrement, and you won't have a problem.</p>
737
738</div>
739
740<!-- _______________________________________________________________________ -->
741<div class="doc_subsubsection">
742 <a name="ll_avoidendl">Avoid <tt>std::endl</tt></a>
743</div>
744
745<div class="doc_text">
746
747<p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
748to the output stream specified. In addition to doing this, however, it also
749flushes the output stream. In other words, these are equivalent:</p>
750
751<div class="doc_code">
752<pre>
753std::cout &lt;&lt; std::endl;
754std::cout &lt;&lt; '\n' &lt;&lt; std::flush;
755</pre>
756</div>
757
758<p>Most of the time, you probably have no reason to flush the output stream, so
759it's better to use a literal <tt>'\n'</tt>.</p>
760
761</div>
762
763
764<!-- *********************************************************************** -->
765<div class="doc_section">
766 <a name="seealso">See Also</a>
767</div>
768<!-- *********************************************************************** -->
769
770<div class="doc_text">
771
772<p>A lot of these comments and recommendations have been culled for other
773sources. Two particularly important books for our work are:</p>
774
775<ol>
776
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]777<li><a href="http://www.amazon.com/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0321334876">Effective
778C++</a> by Scott Meyers. Also
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]779interesting and useful are "More Effective C++" and "Effective STL" by the same
780author.</li>
781
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]782<li>Large-Scale C++ Software Design by John Lakos</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]783
784</ol>
785
786<p>If you get some free time, and you haven't read them: do so, you might learn
787something.</p>
788
789</div>
790
791<!-- *********************************************************************** -->
792
793<hr>
794<address>
795 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
796 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
797 <a href="http://validator.w3.org/check/referer"><img
798 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
799
800 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
801 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
802 Last modified: $Date$
803</address>
804
805</body>
806</html>