Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 48386693bb5b4b2f71acaa3f2d38ffad6705912b / . / docs / CodingStandards.html
blob: 79c816ef65f3fbfc3add9b5f70bbc3f35d34fab2 [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>
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]53 <li><a href="#ll_end">Don't evaluate end() every time through a
54 loop</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]55 <li><a href="#ll_preincrement">Prefer Preincrement</a></li>
56 <li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
57 </ol></li>
58 </ol></li>
59 <li><a href="#seealso">See Also</a></li>
60</ol>
61
62<div class="doc_author">
63 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
64 <a href="mailto:void@nondot.org">Bill Wendling</a></p>
65</div>
66
67
68<!-- *********************************************************************** -->
69<div class="doc_section">
70 <a name="introduction">Introduction</a>
71</div>
72<!-- *********************************************************************** -->
73
74<div class="doc_text">
75
76<p>This document attempts to describe a few coding standards that are being used
77in the LLVM source tree. Although no coding standards should be regarded as
78absolute requirements to be followed in all instances, coding standards can be
79useful.</p>
80
81<p>This document intentionally does not prescribe fixed standards for religious
82issues such as brace placement and space usage. For issues like this, follow
83the golden rule:</p>
84
85<blockquote>
86
87<p><b><a name="goldenrule">If you are adding a significant body of source to a
88project, feel free to use whatever style you are most comfortable with. If you
89are extending, enhancing, or bug fixing already implemented code, use the style
90that is already being used so that the source is uniform and easy to
91follow.</a></b></p>
92
93</blockquote>
94
95<p>The ultimate goal of these guidelines is the increase readability and
96maintainability of our common source base. If you have suggestions for topics to
97be included, please mail them to <a
98href="mailto:sabre@nondot.org">Chris</a>.</p>
99
100</div>
101
102<!-- *********************************************************************** -->
103<div class="doc_section">
104 <a name="mechanicalissues">Mechanical Source Issues</a>
105</div>
106<!-- *********************************************************************** -->
107
108<!-- ======================================================================= -->
109<div class="doc_subsection">
110 <a name="sourceformating">Source Code Formatting</a>
111</div>
112
113<!-- _______________________________________________________________________ -->
114<div class="doc_subsubsection">
115 <a name="scf_commenting">Commenting</a>
116</div>
117
118<div class="doc_text">
119
120<p>Comments are one critical part of readability and maintainability. Everyone
Chris Lattner48386692009-07-12 00:10:24 +0000[diff] [blame^]121knows they should comment, so should you. When writing comments, write them as
122English prose, which means they should use proper capitalization, punctuation,
123etc. Although we all should probably
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]124comment our code more than we do, there are a few very critical places that
125documentation is very useful:</p>
126
127<b>File Headers</b>
128
129<p>Every source file should have a header on it that describes the basic
130purpose of the file. If a file does not have a header, it should not be
131checked into Subversion. Most source trees will probably have a standard
132file header format. The standard format for the LLVM source tree looks like
133this:</p>
134
135<div class="doc_code">
136<pre>
137//===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
Misha Brukman0fd5f3e2009-01-02 16:58:42 +0000[diff] [blame]138//
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]139// The LLVM Compiler Infrastructure
140//
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]141// This file is distributed under the University of Illinois Open Source
142// License. See LICENSE.TXT for details.
Misha Brukman0fd5f3e2009-01-02 16:58:42 +0000[diff] [blame]143//
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]144//===----------------------------------------------------------------------===//
145//
146// This file contains the declaration of the Instruction class, which is the
147// base class for all of the VM instructions.
148//
149//===----------------------------------------------------------------------===//
150</pre>
151</div>
152
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]153<p>A few things to note about this particular format: The "<tt>-*- C++
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]154-*-</tt>" string on the first line is there to tell Emacs that the source file
155is a C++ file, not a C file (Emacs assumes .h files are C files by default).
156Note that this tag is not necessary in .cpp files. The name of the file is also
157on the first line, along with a very short description of the purpose of the
158file. This is important when printing out code and flipping though lots of
159pages.</p>
160
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]161<p>The next section in the file is a concise note that defines the license
162that the file is released under. This makes it perfectly clear what terms the
163source code can be distributed under and should not be modified in any way.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]164
165<p>The main body of the description does not have to be very long in most cases.
166Here it's only two lines. If an algorithm is being implemented or something
167tricky is going on, a reference to the paper where it is published should be
168included, as well as any notes or "gotchas" in the code to watch out for.</p>
169
170<b>Class overviews</b>
171
172<p>Classes are one fundamental part of a good object oriented design. As such,
173a class definition should have a comment block that explains what the class is
174used for... if it's not obvious. If it's so completely obvious your grandma
175could figure it out, it's probably safe to leave it out. Naming classes
176something sane goes a long ways towards avoiding writing documentation.</p>
177
178
179<b>Method information</b>
180
181<p>Methods defined in a class (as well as any global functions) should also be
182documented properly. A quick note about what it does any a description of the
183borderline behaviour is all that is necessary here (unless something
184particularly tricky or insideous is going on). The hope is that people can
185figure out how to use your interfaces without reading the code itself... that is
186the goal metric.</p>
187
188<p>Good things to talk about here are what happens when something unexpected
189happens: does the method return null? Abort? Format your hard disk?</p>
190
191</div>
192
193<!-- _______________________________________________________________________ -->
194<div class="doc_subsubsection">
195 <a name="scf_commentformat">Comment Formatting</a>
196</div>
197
198<div class="doc_text">
199
200<p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
201require less typing, don't have nesting problems, etc. There are a few cases
202when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
203
204<ol>
205 <li>When writing a C code: Obviously if you are writing C code, use C style
206 comments.</li>
207 <li>When writing a header file that may be <tt>#include</tt>d by a C source
208 file.</li>
209 <li>When writing a source file that is used by a tool that only accepts C
210 style comments.</li>
211</ol>
212
213<p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
214These nest properly and are better behaved in general than C style comments.</p>
215
216</div>
217
218<!-- _______________________________________________________________________ -->
219<div class="doc_subsubsection">
220 <a name="scf_includes"><tt>#include</tt> Style</a>
221</div>
222
223<div class="doc_text">
224
225<p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
226include guards if working on a header file), the <a
227href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
228file should be listed. We prefer these <tt>#include</tt>s to be listed in this
229order:</p>
230
231<ol>
232 <li><a href="#mmheader">Main Module header</a></li>
233 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
234 <li><tt>llvm/*</tt></li>
235 <li><tt>llvm/Analysis/*</tt></li>
236 <li><tt>llvm/Assembly/*</tt></li>
237 <li><tt>llvm/Bytecode/*</tt></li>
238 <li><tt>llvm/CodeGen/*</tt></li>
239 <li>...</li>
240 <li><tt>Support/*</tt></li>
241 <li><tt>Config/*</tt></li>
242 <li>System <tt>#includes</tt></li>
243</ol>
244
Nick Lewycky61819f72008-11-29 20:13:25 +0000[diff] [blame]245<p>... and each category should be sorted by name.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]246
247<p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
248which implement an interface defined by a .h file. This <tt>#include</tt>
249should always be included <b>first</b> regardless of where it lives on the file
250system. By including a header file first in the .cpp files that implement the
251interfaces, we ensure that the header does not have any hidden dependencies
252which are not explicitly #included in the header, but should be. It is also a
253form of documentation in the .cpp file to indicate where the interfaces it
254implements are defined.</p>
255
256</div>
257
258<!-- _______________________________________________________________________ -->
259<div class="doc_subsubsection">
260 <a name="scf_codewidth">Source Code Width</a>
261</div>
262
263<div class="doc_text">
264
265<p>Write your code to fit within 80 columns of text. This helps those of us who
266like to print out code and look at your code in an xterm without resizing
267it.</p>
268
Chris Lattnerb5daf702008-07-08 05:12:37 +0000[diff] [blame]269<p>The longer answer is that there must be some limit to the width of the code
270in order to reasonably allow developers to have multiple files side-by-side in
271windows on a modest display. If you are going to pick a width limit, it is
272somewhat arbitrary but you might as well pick something standard. Going with
27390 columns (for example) instead of 80 columns wouldn't add any significant
274value and would be detrimental to printing out code. Also many other projects
275have standardized on 80 columns, so some people have already configured their
276editors for it (vs something else, like 90 columns).</p>
277
278<p>This is one of many contentious issues in coding standards, but is not up
279for debate.</p>
280
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]281</div>
282
283<!-- _______________________________________________________________________ -->
284<div class="doc_subsubsection">
285 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
286</div>
287
288<div class="doc_text">
289
290<p>In all cases, prefer spaces to tabs in source files. People have different
291prefered indentation levels, and different styles of indentation that they
292like... this is fine. What isn't is that different editors/viewers expand tabs
293out to different tab stops. This can cause your code to look completely
294unreadable, and it is not worth dealing with.</p>
295
296<p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
297style of existing code if your are modifying and extending it. If you like four
298spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
299with two spaces of indentation. Also, do not reindent a whole source file: it
300makes for incredible diffs that are absolutely worthless.</p>
301
302</div>
303
304<!-- _______________________________________________________________________ -->
305<div class="doc_subsubsection">
306 <a name="scf_indentation">Indent Code Consistently</a>
307</div>
308
309<div class="doc_text">
310
311<p>Okay, your first year of programming you were told that indentation is
312important. If you didn't believe and internalize this then, now is the time.
313Just do it.</p>
314
315</div>
316
317
318<!-- ======================================================================= -->
319<div class="doc_subsection">
320 <a name="compilerissues">Compiler Issues</a>
321</div>
322
323
324<!-- _______________________________________________________________________ -->
325<div class="doc_subsubsection">
326 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
327</div>
328
329<div class="doc_text">
330
331<p>If your code has compiler warnings in it, something is wrong: you aren't
332casting values correctly, your have "questionable" constructs in your code, or
333you are doing something legitimately wrong. Compiler warnings can cover up
334legitimate errors in output and make dealing with a translation unit
335difficult.</p>
336
337<p>It is not possible to prevent all warnings from all compilers, nor is it
338desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
339a good thorough set of warnings, and stick to them. At least in the case of
340<tt>gcc</tt>, it is possible to work around any spurious errors by changing the
341syntax of the code slightly. For example, an warning that annoys me occurs when
342I write code like this:</p>
343
344<div class="doc_code">
345<pre>
346if (V = getValue()) {
347 ...
348}
349</pre>
350</div>
351
352<p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
353operator, and that I probably mistyped it. In most cases, I haven't, and I
354really don't want the spurious errors. To fix this particular problem, I
355rewrite the code like this:</p>
356
357<div class="doc_code">
358<pre>
359if ((V = getValue())) {
360 ...
361}
362</pre>
363</div>
364
365<p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
366be fixed by massaging the code appropriately.</p>
367
368<p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
369-Winline -W -Wwrite-strings -Wno-unused</tt></p>
370
371</div>
372
373<!-- _______________________________________________________________________ -->
374<div class="doc_subsubsection">
375 <a name="ci_portable_code">Write Portable Code</a>
376</div>
377
378<div class="doc_text">
379
380<p>In almost all cases, it is possible and within reason to write completely
381portable code. If there are cases where it isn't possible to write portable
382code, isolate it behind a well defined (and well documented) interface.</p>
383
384<p>In practice, this means that you shouldn't assume much about the host
385compiler, including its support for "high tech" features like partial
Chris Lattnerafec0252009-03-23 04:52:53 +0000[diff] [blame]386specialization of templates. If these features are used, they should only be
387an implementation detail of a library which has a simple exposed API.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]388
389</div>
390
391<!-- _______________________________________________________________________ -->
392<div class="doc_subsubsection">
393<a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
394</div>
395<div class="doc_text">
396
397<p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
398interchangeably. The only difference is when they are used to declare a class:
399<tt>class</tt> makes all members private by default while <tt>struct</tt> makes
400all members public by default.</p>
401
402<p>Unfortunately, not all compilers follow the rules and some will generate
403different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
404declare the symbol. This can lead to problems at link time.</p>
405
406<p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
407<b>all</b> members are public, in which case <tt>struct</tt> is allowed.</p>
408
409</div>
410
411<!-- *********************************************************************** -->
412<div class="doc_section">
413 <a name="styleissues">Style Issues</a>
414</div>
415<!-- *********************************************************************** -->
416
417
418<!-- ======================================================================= -->
419<div class="doc_subsection">
420 <a name="macro">The High Level Issues</a>
421</div>
422
423
424<!-- _______________________________________________________________________ -->
425<div class="doc_subsubsection">
426 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
427</div>
428
429<div class="doc_text">
430
431<p>C++ doesn't do too well in the modularity department. There is no real
432encapsulation or data hiding (unless you use expensive protocol classes), but it
433is what we have to work with. When you write a public header file (in the LLVM
434source tree, they live in the top level "include" directory), you are defining a
435module of functionality.</p>
436
437<p>Ideally, modules should be completely independent of each other, and their
438header files should only include the absolute minimum number of headers
439possible. A module is not just a class, a function, or a namespace: <a
440href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
441of these</a> that defines an interface. This interface may be several
442functions, classes or data structures, but the important issue is how they work
443together.</p>
444
445<p>In general, a module should be implemented with one or more <tt>.cpp</tt>
446files. Each of these <tt>.cpp</tt> files should include the header that defines
447their interface first. This ensure that all of the dependences of the module
448header have been properly added to the module header itself, and are not
449implicit. System headers should be included after user headers for a
450translation unit.</p>
451
452</div>
453
454<!-- _______________________________________________________________________ -->
455<div class="doc_subsubsection">
456 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
457</div>
458
459<div class="doc_text">
460
461<p><tt>#include</tt> hurts compile time performance. Don't do it unless you
462have to, especially in header files.</p>
463
464<p>But wait, sometimes you need to have the definition of a class to use it, or
465to inherit from it. In these cases go ahead and <tt>#include</tt> that header
466file. Be aware however that there are many cases where you don't need to have
467the full definition of a class. If you are using a pointer or reference to a
468class, you don't need the header file. If you are simply returning a class
469instance from a prototyped function or method, you don't need it. In fact, for
470most cases, you simply don't need the definition of a class... and not
471<tt>#include</tt>'ing speeds up compilation.</p>
472
473<p>It is easy to try to go too overboard on this recommendation, however. You
474<b>must</b> include all of the header files that you are using -- you can
475include them either directly
476or indirectly (through another header file). To make sure that you don't
477accidently forget to include a header file in your module header, make sure to
478include your module header <b>first</b> in the implementation file (as mentioned
479above). This way there won't be any hidden dependencies that you'll find out
480about later...</p>
481
482</div>
483
484<!-- _______________________________________________________________________ -->
485<div class="doc_subsubsection">
486 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
487</div>
488
489<div class="doc_text">
490
491<p>Many modules have a complex implementation that causes them to use more than
492one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
493internal communication interface (helper classes, extra functions, etc) in the
494public module header file. Don't do this.</p>
495
496<p>If you really need to do something like this, put a private header file in
497the same directory as the source files, and include it locally. This ensures
498that your private interface remains private and undisturbed by outsiders.</p>
499
500<p>Note however, that it's okay to put extra implementation methods a public
501class itself... just make them private (or protected), and all is well.</p>
502
503</div>
504
505<!-- _______________________________________________________________________ -->
506<div class="doc_subsubsection">
507 <a name="ll_iostream"><tt>#include &lt;iostream&gt;</tt> is forbidden</a>
508</div>
509
510<div class="doc_text">
511
512<p>The use of <tt>#include &lt;iostream&gt;</tt> in library files is
513hereby <b><em>forbidden</em></b>. The primary reason for doing this is to
514support clients using LLVM libraries as part of larger systems. In particular,
515we statically link LLVM into some dynamic libraries. Even if LLVM isn't used,
516the static c'tors are run whenever an application start up that uses the dynamic
517library. There are two problems with this:</p>
518
519<ol>
520 <li>The time to run the static c'tors impacts startup time of
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]521 applications&mdash;a critical time for GUI apps.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]522 <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]523 disk: both the code for the static c'tors in each <tt>.o</tt> file and the
524 small amount of data that gets touched. In addition, touched/dirty pages
525 put more pressure on the VM system on low-memory machines.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]526</ol>
527
Matthijs Kooijmanc4c0af22008-07-30 12:14:10 +0000[diff] [blame]528<p>Note that using the other stream headers (<tt>&lt;sstream&gt;</tt> for
529example) is allowed normally, it is just <tt>&lt;iostream&gt;</tt> that is
530causing problems.</p>
531
Chris Lattnerafec0252009-03-23 04:52:53 +0000[diff] [blame]532<p>The preferred replacement for stream functionality is the
Chris Lattner85187f12009-03-23 04:53:34 +0000[diff] [blame]533<tt>llvm::raw_ostream</tt> class (for writing to output streams of various
534sorts) and the <tt>llvm::MemoryBuffer</tt> API (for reading in files).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]535
536</div>
537
538
539<!-- ======================================================================= -->
540<div class="doc_subsection">
541 <a name="micro">The Low Level Issues</a>
542</div>
543
544
545<!-- _______________________________________________________________________ -->
546<div class="doc_subsubsection">
547 <a name="ll_assert">Assert Liberally</a>
548</div>
549
550<div class="doc_text">
551
552<p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
553preconditions and assumptions, you never know when a bug (not neccesarily even
554yours) might be caught early by an assertion, which reduces debugging time
555dramatically. The "<tt>&lt;cassert&gt;</tt>" header file is probably already
556included by the header files you are using, so it doesn't cost anything to use
557it.</p>
558
559<p>To further assist with debugging, make sure to put some kind of error message
560in the assertion statement (which is printed if the assertion is tripped). This
561helps the poor debugging make sense of why an assertion is being made and
562enforced, and hopefully what to do about it. Here is one complete example:</p>
563
564<div class="doc_code">
565<pre>
566inline Value *getOperand(unsigned i) {
567 assert(i &lt; Operands.size() &amp;&amp; "getOperand() out of range!");
568 return Operands[i];
569}
570</pre>
571</div>
572
573<p>Here are some examples:</p>
574
575<div class="doc_code">
576<pre>
577assert(Ty-&gt;isPointerType() &amp;&amp; "Can't allocate a non pointer type!");
578
579assert((Opcode == Shl || Opcode == Shr) &amp;&amp; "ShiftInst Opcode invalid!");
580
581assert(idx &lt; getNumSuccessors() &amp;&amp; "Successor # out of range!");
582
583assert(V1.getType() == V2.getType() &amp;&amp; "Constant types must be identical!");
584
585assert(isa&lt;PHINode&gt;(Succ-&gt;front()) &amp;&amp; "Only works on PHId BBs!");
586</pre>
587</div>
588
589<p>You get the idea...</p>
590
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]591<p>Please be aware when adding assert statements that not all compilers are aware of
592the semantics of the assert. In some places, asserts are used to indicate a piece of
593code that should not be reached. These are typically of the form:</p>
594
595<div class="doc_code">
596<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]597assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]598</pre>
599</div>
600
601<p>When used in a function that returns a value, they should be followed with a return
602statement and a comment indicating that this line is never reached. This will prevent
603a compiler which is unable to deduce that the assert statement never returns from
604generating a warning.</p>
605
606<div class="doc_code">
607<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]608assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]609// Not reached
610return 0;
611</pre>
612</div>
613
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]614</div>
615
616<!-- _______________________________________________________________________ -->
617<div class="doc_subsubsection">
618 <a name="ll_ns_std">Do not use '<tt>using namespace std</tt>'</a>
619</div>
620
621<div class="doc_text">
622<p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
623namespace with an "<tt>std::</tt>" prefix, rather than rely on
624"<tt>using namespace std;</tt>".</p>
625
626<p> In header files, adding a '<tt>using namespace XXX</tt>' directive pollutes
Misha Brukmanbcc53732008-12-11 19:37:04 +0000[diff] [blame]627the namespace of any source file that <tt>#include</tt>s the header. This is
628clearly a bad thing.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]629
630<p>In implementation files (e.g. .cpp files), the rule is more of a stylistic
631rule, but is still important. Basically, using explicit namespace prefixes
632makes the code <b>clearer</b>, because it is immediately obvious what facilities
633are being used and where they are coming from, and <b>more portable</b>, because
634namespace clashes cannot occur between LLVM code and other namespaces. The
635portability rule is important because different standard library implementations
636expose different symbols (potentially ones they shouldn't), and future revisions
637to the C++ standard will add more symbols to the <tt>std</tt> namespace. As
638such, we never use '<tt>using namespace std;</tt>' in LLVM.</p>
639
640<p>The exception to the general rule (i.e. it's not an exception for
641the <tt>std</tt> namespace) is for implementation files. For example, all of
642the code in the LLVM project implements code that lives in the 'llvm' namespace.
643As such, it is ok, and actually clearer, for the .cpp files to have a '<tt>using
644namespace llvm</tt>' directive at their top, after the <tt>#include</tt>s. The
645general form of this rule is that any .cpp file that implements code in any
646namespace may use that namespace (and its parents'), but should not use any
647others.</p>
648
649</div>
650
651<!-- _______________________________________________________________________ -->
652<div class="doc_subsubsection">
653 <a name="ll_virtual_anch">Provide a virtual method anchor for classes
654 in headers</a>
655</div>
656
657<div class="doc_text">
658
659<p>If a class is defined in a header file and has a v-table (either it has
660virtual methods or it derives from classes with virtual methods), it must
661always have at least one out-of-line virtual method in the class. Without
Misha Brukmanbcc53732008-12-11 19:37:04 +0000[diff] [blame]662this, the compiler will copy the vtable and RTTI into every <tt>.o</tt> file
663that <tt>#include</tt>s the header, bloating <tt>.o</tt> file sizes and
664increasing link times.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]665
666</div>
667
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]668<!-- _______________________________________________________________________ -->
669<div class="doc_subsubsection">
670 <a name="ll_end">Don't evaluate end() every time through a loop</a>
671</div>
672
673<div class="doc_text">
674
675<p>Because C++ doesn't have a standard "foreach" loop (though it can be emulated
676with macros and may be coming in C++'0x) we end up writing a lot of loops that
677manually iterate from begin to end on a variety of containers or through other
678data structures. One common mistake is to write a loop in this style:</p>
679
680<div class="doc_code">
681<pre>
682 BasicBlock *BB = ...
683 for (BasicBlock::iterator I = BB->begin(); I != <b>BB->end()</b>; ++I)
684 ... use I ...
685</pre>
686</div>
687
688<p>The problem with this construct is that it evaluates "<tt>BB->end()</tt>"
689every time through the loop. Instead of writing the loop like this, we strongly
690prefer loops to be written so that they evaluate it once before the loop starts.
691A convenient way to do this is like so:</p>
692
693<div class="doc_code">
694<pre>
695 BasicBlock *BB = ...
696 for (BasicBlock::iterator I = BB->begin(), E = <b>BB->end()</b>; I != E; ++I)
697 ... use I ...
698</pre>
699</div>
700
701<p>The observant may quickly point out that these two loops may have different
702semantics: if the container (a basic block in this case) is being mutated, then
703"<tt>BB->end()</tt>" may change its value every time through the loop and the
704second loop may not in fact be correct. If you actually do depend on this
Chris Lattnerdc43b342009-06-30 06:20:03 +0000[diff] [blame]705behavior, please write the loop in the first form and add a comment indicating
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]706that you did it intentionally.</p>
707
708<p>Why do we prefer the second form (when correct)? Writing the loop in the
709first form has two problems: First it may be less efficient than evaluating it
710at the start of the loop. In this case, the cost is probably minor: a few extra
711loads every time through the loop. However, if the base expression is more
712complex, then the cost can rise quickly. I've seen loops where the end
713expression was actually something like: "<tt>SomeMap[x]->end()</tt>" and map
Chris Lattner088affa2009-06-30 06:27:54 +0000[diff] [blame]714lookups really aren't cheap. By writing it in the second form consistently, you
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]715eliminate the issue entirely and don't even have to think about it.</p>
716
Chris Lattnerdc43b342009-06-30 06:20:03 +0000[diff] [blame]717<p>The second (even bigger) issue is that writing the loop in the first form
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]718hints to the reader that the loop is mutating the container (a fact that a
719comment would handily confirm!). If you write the loop in the second form, it
720is immediately obvious without even looking at the body of the loop that the
721container isn't being modified, which makes it easier to read the code and
722understand what it does.</p>
723
724<p>While the second form of the loop is a few extra keystrokes, we do strongly
725prefer it.</p>
726
727</div>
728
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]729
730<!-- _______________________________________________________________________ -->
731<div class="doc_subsubsection">
732 <a name="ll_preincrement">Prefer Preincrement</a>
733</div>
734
735<div class="doc_text">
736
737<p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
738postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
739preincrementation whenever possible.</p>
740
741<p>The semantics of postincrement include making a copy of the value being
742incremented, returning it, and then preincrementing the "work value". For
743primitive types, this isn't a big deal... but for iterators, it can be a huge
744issue (for example, some iterators contains stack and set objects in them...
745copying an iterator could invoke the copy ctor's of these as well). In general,
746get in the habit of always using preincrement, and you won't have a problem.</p>
747
748</div>
749
750<!-- _______________________________________________________________________ -->
751<div class="doc_subsubsection">
752 <a name="ll_avoidendl">Avoid <tt>std::endl</tt></a>
753</div>
754
755<div class="doc_text">
756
757<p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
758to the output stream specified. In addition to doing this, however, it also
759flushes the output stream. In other words, these are equivalent:</p>
760
761<div class="doc_code">
762<pre>
763std::cout &lt;&lt; std::endl;
764std::cout &lt;&lt; '\n' &lt;&lt; std::flush;
765</pre>
766</div>
767
768<p>Most of the time, you probably have no reason to flush the output stream, so
769it's better to use a literal <tt>'\n'</tt>.</p>
770
771</div>
772
773
774<!-- *********************************************************************** -->
775<div class="doc_section">
776 <a name="seealso">See Also</a>
777</div>
778<!-- *********************************************************************** -->
779
780<div class="doc_text">
781
782<p>A lot of these comments and recommendations have been culled for other
783sources. Two particularly important books for our work are:</p>
784
785<ol>
786
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]787<li><a href="http://www.amazon.com/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0321334876">Effective
788C++</a> by Scott Meyers. Also
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]789interesting and useful are "More Effective C++" and "Effective STL" by the same
790author.</li>
791
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]792<li>Large-Scale C++ Software Design by John Lakos</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]793
794</ol>
795
796<p>If you get some free time, and you haven't read them: do so, you might learn
797something.</p>
798
799</div>
800
801<!-- *********************************************************************** -->
802
803<hr>
804<address>
805 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]806 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]807 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]808 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]809
810 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
811 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
812 Last modified: $Date$
813</address>
814
815</body>
816</html>