Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / 2ea8ee7c76b8d8754d81072e691caf25d23289e3 / . / docs / CodingStandards.html
blob: 7791e8dd872b2993ac7881e0cb6f5b26023ffb69 [file] [log] [blame]
Misha Brukman75f258e2003-10-24 17:57:33 +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 Brukmanc1449e62008-12-10 23:07:02 +0000[diff] [blame]6 <title>LLVM Coding Standards</title>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]7</head>
8<body>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]9
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]10<div class="doc_title">
Misha Brukmanc1449e62008-12-10 23:07:02 +0000[diff] [blame]11 LLVM Coding Standards
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]12</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]13
14<ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]15 <li><a href="#introduction">Introduction</a></li>
Chris Lattner7ae36bb2001-07-23 20:40:41 +0000[diff] [blame]16 <li><a href="#mechanicalissues">Mechanical Source Issues</a>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]17 <ol>
18 <li><a href="#sourceformating">Source Code Formatting</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]19 <ol>
20 <li><a href="#scf_commenting">Commenting</a></li>
21 <li><a href="#scf_commentformat">Comment Formatting</a></li>
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]22 <li><a href="#scf_includes"><tt>#include</tt> Style</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]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>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]27 <li><a href="#compilerissues">Compiler Issues</a>
28 <ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]29 <li><a href="#ci_warningerrors">Treat Compiler Warnings Like
30 Errors</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]31 <li><a href="#ci_portable_code">Write Portable Code</a></li>
Chris Lattner558f5d22010-11-30 19:20:40 +0000[diff] [blame]32 <li><a href="#ci_rtti_exceptions">Do not use RTTI or Exceptions</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]33 <li><a href="#ci_class_struct">Use of <tt>class</tt>/<tt>struct</tt> Keywords</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]34 </ol></li>
35 </ol></li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]36 <li><a href="#styleissues">Style Issues</a>
37 <ol>
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]38 <li><a href="#macro">The High-Level Issues</a>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]39 <ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]40 <li><a href="#hl_module">A Public Header File <b>is</b> a
41 Module</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]42 <li><a href="#hl_dontinclude"><tt>#include</tt> as Little as Possible</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]43 <li><a href="#hl_privateheaders">Keep "internal" Headers
44 Private</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]45 <li><a href="#hl_earlyexit">Use Early Exits and <tt>continue</tt> to Simplify
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]46 Code</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]47 <li><a href="#hl_else_after_return">Don't use <tt>else</tt> after a
48 <tt>return</tt></a></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]49 <li><a href="#hl_predicateloops">Turn Predicate Loops into Predicate
50 Functions</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]51 </ol></li>
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]52 <li><a href="#micro">The Low-Level Issues</a>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]53 <ol>
Zhanyong Wan7fcd4dc2010-12-02 05:10:07 +0000[diff] [blame]54 <li><a href="#ll_naming">Name Types, Functions, Variables, and Enumerators Properly</a></li>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]55 <li><a href="#ll_assert">Assert Liberally</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]56 <li><a href="#ll_ns_std">Do not use '<tt>using namespace std</tt>'</a></li>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]57 <li><a href="#ll_virtual_anch">Provide a virtual method anchor for
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]58 classes in headers</a></li>
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]59 <li><a href="#ll_end">Don't evaluate <tt>end()</tt> every time through a
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]60 loop</a></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]61 <li><a href="#ll_iostream"><tt>#include &lt;iostream&gt;</tt> is
62 <em>forbidden</em></a></li>
Daniel Dunbare8530a32009-07-24 23:04:51 +0000[diff] [blame]63 <li><a href="#ll_raw_ostream">Use <tt>raw_ostream</tt></a</li>
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]64 <li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]65 </ol></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]66
67 <li><a href="#nano">Microscopic Details</a>
68 <ol>
69 <li><a href="#micro_spaceparen">Spaces Before Parentheses</a></li>
70 <li><a href="#micro_preincrement">Prefer Preincrement</a></li>
71 <li><a href="#micro_namespaceindent">Namespace Indentation</a></li>
72 <li><a href="#micro_anonns">Anonymous Namespaces</a></li>
73 </ol></li>
74
75
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]76 </ol></li>
77 <li><a href="#seealso">See Also</a></li>
78</ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]79
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]80<div class="doc_author">
Chris Lattner219bd292009-07-22 05:43:01 +0000[diff] [blame]81 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a></p>
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]82</div>
83
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]84
85<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]86<div class="doc_section">
87 <a name="introduction">Introduction</a>
88</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]89<!-- *********************************************************************** -->
90
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]91<div class="doc_text">
92
93<p>This document attempts to describe a few coding standards that are being used
94in the LLVM source tree. Although no coding standards should be regarded as
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]95absolute requirements to be followed in all instances, coding standards can be
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]96useful.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]97
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]98<p>This document intentionally does not prescribe fixed standards for religious
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]99issues such as brace placement and space usage. For issues like this, follow
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]100the golden rule:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]101
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]102<blockquote>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]103
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]104<p><b><a name="goldenrule">If you are adding a significant body of source to a
105project, feel free to use whatever style you are most comfortable with. If you
106are extending, enhancing, or bug fixing already implemented code, use the style
107that is already being used so that the source is uniform and easy to
108follow.</a></b></p>
109
110</blockquote>
111
112<p>The ultimate goal of these guidelines is the increase readability and
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]113maintainability of our common source base. If you have suggestions for topics to
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]114be included, please mail them to <a
115href="mailto:sabre@nondot.org">Chris</a>.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]116
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]117</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]118
119<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]120<div class="doc_section">
121 <a name="mechanicalissues">Mechanical Source Issues</a>
122</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]123<!-- *********************************************************************** -->
124
125<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]126<div class="doc_subsection">
127 <a name="sourceformating">Source Code Formatting</a>
128</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]129
130<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]131<div class="doc_subsubsection">
132 <a name="scf_commenting">Commenting</a>
133</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]134
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]135<div class="doc_text">
136
137<p>Comments are one critical part of readability and maintainability. Everyone
Chris Lattnerb927ca82009-07-12 00:10:24 +0000[diff] [blame]138knows they should comment, so should you. When writing comments, write them as
139English prose, which means they should use proper capitalization, punctuation,
140etc. Although we all should probably
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]141comment our code more than we do, there are a few very critical places that
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]142documentation is very useful:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]143
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]144<b>File Headers</b>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]145
Reid Spencer669ed452007-07-09 08:04:31 +0000[diff] [blame]146<p>Every source file should have a header on it that describes the basic
147purpose of the file. If a file does not have a header, it should not be
148checked into Subversion. Most source trees will probably have a standard
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]149file header format. The standard format for the LLVM source tree looks like
150this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]151
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]152<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]153<pre>
Chris Lattnerc49fcd12003-10-13 14:58:11 +0000[diff] [blame]154//===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
Misha Brukman98117cd2009-01-02 16:58:42 +0000[diff] [blame]155//
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]156// The LLVM Compiler Infrastructure
157//
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]158// This file is distributed under the University of Illinois Open Source
159// License. See LICENSE.TXT for details.
Misha Brukman98117cd2009-01-02 16:58:42 +0000[diff] [blame]160//
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]161//===----------------------------------------------------------------------===//
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]162//
163// This file contains the declaration of the Instruction class, which is the
164// base class for all of the VM instructions.
165//
166//===----------------------------------------------------------------------===//
167</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]168</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]169
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]170<p>A few things to note about this particular format: The "<tt>-*- C++
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]171-*-</tt>" string on the first line is there to tell Emacs that the source file
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]172is a C++ file, not a C file (Emacs assumes <tt>.h</tt> files are C files by default).
173Note that this tag is not necessary in <tt>.cpp</tt> files. The name of the file is also
Misha Brukman69c58892004-07-28 22:37:57 +0000[diff] [blame]174on the first line, along with a very short description of the purpose of the
175file. This is important when printing out code and flipping though lots of
176pages.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]177
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]178<p>The next section in the file is a concise note that defines the license
179that the file is released under. This makes it perfectly clear what terms the
180source code can be distributed under and should not be modified in any way.</p>
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]181
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]182<p>The main body of the description does not have to be very long in most cases.
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]183Here it's only two lines. If an algorithm is being implemented or something
184tricky is going on, a reference to the paper where it is published should be
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]185included, as well as any notes or "gotchas" in the code to watch out for.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]186
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]187<b>Class overviews</b>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]188
Jim Laskeyf55914a2006-07-31 20:18:49 +0000[diff] [blame]189<p>Classes are one fundamental part of a good object oriented design. As such,
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]190a class definition should have a comment block that explains what the class is
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]191used for... if it's not obvious. If it's so completely obvious your grandma
192could figure it out, it's probably safe to leave it out. Naming classes
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]193something sane goes a long ways towards avoiding writing documentation.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]194
195
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]196<b>Method information</b>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]197
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]198<p>Methods defined in a class (as well as any global functions) should also be
Dan Gohman621a4152010-02-26 20:18:32 +0000[diff] [blame]199documented properly. A quick note about what it does and a description of the
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]200borderline behaviour is all that is necessary here (unless something
Dan Gohman621a4152010-02-26 20:18:32 +0000[diff] [blame]201particularly tricky or insidious is going on). The hope is that people can
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]202figure out how to use your interfaces without reading the code itself... that is
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]203the goal metric.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]204
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]205<p>Good things to talk about here are what happens when something unexpected
206happens: does the method return null? Abort? Format your hard disk?</p>
207
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]208</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]209
210<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]211<div class="doc_subsubsection">
212 <a name="scf_commentformat">Comment Formatting</a>
213</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]214
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]215<div class="doc_text">
216
217<p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]218require less typing, don't have nesting problems, etc. There are a few cases
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]219when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]220
221<ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]222 <li>When writing a C code: Obviously if you are writing C code, use C style
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]223 comments.</li>
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]224 <li>When writing a header file that may be <tt>#include</tt>d by a C source
225 file.</li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]226 <li>When writing a source file that is used by a tool that only accepts C
227 style comments.</li>
228</ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]229
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]230<p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
231These nest properly and are better behaved in general than C style comments.</p>
232
233</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]234
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]235<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]236<div class="doc_subsubsection">
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]237 <a name="scf_includes"><tt>#include</tt> Style</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]238</div>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]239
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]240<div class="doc_text">
241
242<p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]243include guards if working on a header file), the <a
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]244href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
245file should be listed. We prefer these <tt>#include</tt>s to be listed in this
246order:</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]247
248<ol>
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]249 <li><a href="#mmheader">Main Module Header</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]250 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]251 <li><tt>llvm/*</tt></li>
252 <li><tt>llvm/Analysis/*</tt></li>
253 <li><tt>llvm/Assembly/*</tt></li>
Duncan Sands18d52f22010-09-29 20:09:55 +0000[diff] [blame]254 <li><tt>llvm/Bitcode/*</tt></li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]255 <li><tt>llvm/CodeGen/*</tt></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]256 <li>...</li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]257 <li><tt>Support/*</tt></li>
258 <li><tt>Config/*</tt></li>
259 <li>System <tt>#includes</tt></li>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]260</ol>
261
Nick Lewycky822293a2008-11-29 20:13:25 +0000[diff] [blame]262<p>... and each category should be sorted by name.</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]263
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]264<p><a name="mmheader">The "Main Module Header"</a> file applies to <tt>.cpp</tt> files
265which implement an interface defined by a <tt>.h</tt> file. This <tt>#include</tt>
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]266should always be included <b>first</b> regardless of where it lives on the file
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]267system. By including a header file first in the <tt>.cpp</tt> files that implement the
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]268interfaces, we ensure that the header does not have any hidden dependencies
269which are not explicitly #included in the header, but should be. It is also a
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]270form of documentation in the <tt>.cpp</tt> file to indicate where the interfaces it
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]271implements are defined.</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]272
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]273</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]274
275<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]276<div class="doc_subsubsection">
277 <a name="scf_codewidth">Source Code Width</a>
278</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]279
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]280<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]281
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]282<p>Write your code to fit within 80 columns of text. This helps those of us who
283like to print out code and look at your code in an xterm without resizing
284it.</p>
285
Chris Lattner5822e9e2008-07-08 05:12:37 +0000[diff] [blame]286<p>The longer answer is that there must be some limit to the width of the code
287in order to reasonably allow developers to have multiple files side-by-side in
288windows on a modest display. If you are going to pick a width limit, it is
289somewhat arbitrary but you might as well pick something standard. Going with
29090 columns (for example) instead of 80 columns wouldn't add any significant
291value and would be detrimental to printing out code. Also many other projects
292have standardized on 80 columns, so some people have already configured their
293editors for it (vs something else, like 90 columns).</p>
294
295<p>This is one of many contentious issues in coding standards, but is not up
296for debate.</p>
297
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]298</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]299
300<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]301<div class="doc_subsubsection">
302 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
303</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]304
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]305<div class="doc_text">
306
307<p>In all cases, prefer spaces to tabs in source files. People have different
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]308preferred indentation levels, and different styles of indentation that they
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]309like... this is fine. What isn't is that different editors/viewers expand tabs
310out to different tab stops. This can cause your code to look completely
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]311unreadable, and it is not worth dealing with.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]312
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]313<p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]314style of existing code if your are modifying and extending it. If you like four
315spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
316with two spaces of indentation. Also, do not reindent a whole source file: it
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]317makes for incredible diffs that are absolutely worthless.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]318
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]319</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]320
321<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]322<div class="doc_subsubsection">
323 <a name="scf_indentation">Indent Code Consistently</a>
324</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]325
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]326<div class="doc_text">
327
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]328<p>Okay, in your first year of programming you were told that indentation is
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]329important. If you didn't believe and internalize this then, now is the time.
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]330Just do it.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]331
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]332</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]333
334
335<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]336<div class="doc_subsection">
337 <a name="compilerissues">Compiler Issues</a>
338</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]339
340
341<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]342<div class="doc_subsubsection">
343 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
344</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]345
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]346<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]347
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]348<p>If your code has compiler warnings in it, something is wrong: you aren't
349casting values correctly, your have "questionable" constructs in your code, or
350you are doing something legitimately wrong. Compiler warnings can cover up
351legitimate errors in output and make dealing with a translation unit
352difficult.</p>
353
354<p>It is not possible to prevent all warnings from all compilers, nor is it
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]355desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
356a good thorough set of warnings, and stick to them. At least in the case of
357<tt>gcc</tt>, it is possible to work around any spurious errors by changing the
358syntax of the code slightly. For example, an warning that annoys me occurs when
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]359I write code like this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]360
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]361<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]362<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]363if (V = getValue()) {
364 ...
365}
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]366</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]367</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]368
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]369<p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
370operator, and that I probably mistyped it. In most cases, I haven't, and I
371really don't want the spurious errors. To fix this particular problem, I
372rewrite the code like this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]373
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]374<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]375<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]376if ((V = getValue())) {
377 ...
378}
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]379</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]380</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]381
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]382<p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
383be fixed by massaging the code appropriately.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]384
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]385<p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
386-Winline -W -Wwrite-strings -Wno-unused</tt></p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]387
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]388</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]389
390<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]391<div class="doc_subsubsection">
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]392 <a name="ci_portable_code">Write Portable Code</a>
393</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]394
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]395<div class="doc_text">
396
397<p>In almost all cases, it is possible and within reason to write completely
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]398portable code. If there are cases where it isn't possible to write portable
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]399code, isolate it behind a well defined (and well documented) interface.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]400
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]401<p>In practice, this means that you shouldn't assume much about the host
Chris Lattner221b2392010-11-16 22:19:06 +0000[diff] [blame]402compiler, and Visual Studio tends to be the lowest common denominator.
403If advanced features are used, they should only be an implementation detail of
404a library which has a simple exposed API, and preferably be buried in
405libSystem.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]406
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]407</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]408
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]409<!-- _______________________________________________________________________ -->
410<div class="doc_subsubsection">
Chris Lattner558f5d22010-11-30 19:20:40 +0000[diff] [blame]411<a name="ci_rtti_exceptions">Do not use RTTI or Exceptions</a>
412</div>
413<div class="doc_text">
414
415<p>LLVM does not use RTTI (e.g. dynamic_cast&lt;&gt;) or exceptions, in an
416effort to reduce code and executable size. These two language features violate
417the general C++ principle of "you only pay for what you use", causing executable
418bloat even if exceptions are never used in a code base, or if RTTI is never used
419for a class. Because of this, we turn them off globally in the code.
420</p>
421
422<p>
423That said, LLVM does make extensive use of a hand-rolled form of RTTI that use
424templates like <a href="ProgrammersManual.html#isa">isa&lt;&gt;, cast&lt;&gt;,
425and dyn_cast&lt;&gt;</a>. This form of RTTI is opt-in and can be added to any
426class. It is also substantially more efficient than dynamic_cast&lt;&gt;.
427</p>
428
429</div>
430
431<!-- _______________________________________________________________________ -->
432<div class="doc_subsubsection">
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]433<a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
434</div>
435<div class="doc_text">
Misha Brukmanedcc78a2004-10-26 15:45:13 +0000[diff] [blame]436
437<p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
438interchangeably. The only difference is when they are used to declare a class:
439<tt>class</tt> makes all members private by default while <tt>struct</tt> makes
440all members public by default.</p>
441
442<p>Unfortunately, not all compilers follow the rules and some will generate
443different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
444declare the symbol. This can lead to problems at link time.</p>
445
446<p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
Chris Lattner57310202009-08-11 17:38:15 +0000[diff] [blame]447<b>all</b> members are public and the type is a C++ "POD" type, in which case
448<tt>struct</tt> is allowed.</p>
Misha Brukmanedcc78a2004-10-26 15:45:13 +0000[diff] [blame]449
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]450</div>
451
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]452<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]453<div class="doc_section">
454 <a name="styleissues">Style Issues</a>
455</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]456<!-- *********************************************************************** -->
457
458
459<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]460<div class="doc_subsection">
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]461 <a name="macro">The High-Level Issues</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]462</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]463<!-- ======================================================================= -->
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]464
465
466<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]467<div class="doc_subsubsection">
468 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
469</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]470
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]471<div class="doc_text">
472
473<p>C++ doesn't do too well in the modularity department. There is no real
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]474encapsulation or data hiding (unless you use expensive protocol classes), but it
475is what we have to work with. When you write a public header file (in the LLVM
476source tree, they live in the top level "include" directory), you are defining a
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]477module of functionality.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]478
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]479<p>Ideally, modules should be completely independent of each other, and their
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]480header files should only include the absolute minimum number of headers
481possible. A module is not just a class, a function, or a namespace: <a
482href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
483of these</a> that defines an interface. This interface may be several
484functions, classes or data structures, but the important issue is how they work
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]485together.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]486
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]487<p>In general, a module should be implemented with one or more <tt>.cpp</tt>
488files. Each of these <tt>.cpp</tt> files should include the header that defines
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]489their interface first. This ensures that all of the dependences of the module
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]490header have been properly added to the module header itself, and are not
491implicit. System headers should be included after user headers for a
492translation unit.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]493
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]494</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]495
496<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]497<div class="doc_subsubsection">
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]498 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]499</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]500
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]501<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]502
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]503<p><tt>#include</tt> hurts compile time performance. Don't do it unless you
504have to, especially in header files.</p>
505
506<p>But wait, sometimes you need to have the definition of a class to use it, or
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]507to inherit from it. In these cases go ahead and <tt>#include</tt> that header
508file. Be aware however that there are many cases where you don't need to have
509the full definition of a class. If you are using a pointer or reference to a
510class, you don't need the header file. If you are simply returning a class
511instance from a prototyped function or method, you don't need it. In fact, for
512most cases, you simply don't need the definition of a class... and not
513<tt>#include</tt>'ing speeds up compilation.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]514
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]515<p>It is easy to try to go too overboard on this recommendation, however. You
Chris Lattner5da59b12007-02-10 18:35:31 +0000[diff] [blame]516<b>must</b> include all of the header files that you are using -- you can
517include them either directly
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]518or indirectly (through another header file). To make sure that you don't
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]519accidentally forget to include a header file in your module header, make sure to
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]520include your module header <b>first</b> in the implementation file (as mentioned
521above). This way there won't be any hidden dependencies that you'll find out
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]522about later...</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]523
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]524</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]525
526<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]527<div class="doc_subsubsection">
528 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
529</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]530
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]531<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]532
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]533<p>Many modules have a complex implementation that causes them to use more than
534one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
535internal communication interface (helper classes, extra functions, etc) in the
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]536public module header file. Don't do this.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]537
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]538<p>If you really need to do something like this, put a private header file in
539the same directory as the source files, and include it locally. This ensures
540that your private interface remains private and undisturbed by outsiders.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]541
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]542<p>Note however, that it's okay to put extra implementation methods a public
543class itself... just make them private (or protected), and all is well.</p>
544
545</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]546
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]547<!-- _______________________________________________________________________ -->
548<div class="doc_subsubsection">
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]549 <a name="hl_earlyexit">Use Early Exits and <tt>continue</tt> to Simplify Code</a>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]550</div>
551
552<div class="doc_text">
553
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]554<p>When reading code, keep in mind how much state and how many previous
555decisions have to be remembered by the reader to understand a block of code.
556Aim to reduce indentation where possible when it doesn't make it more difficult
557to understand the code. One great way to do this is by making use of early
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]558exits and the <tt>continue</tt> keyword in long loops. As an example of using an early
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]559exit from a function, consider this "bad" code:</p>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]560
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]561<div class="doc_code">
562<pre>
563Value *DoSomething(Instruction *I) {
564 if (!isa&lt;TerminatorInst&gt;(I) &amp;&amp;
565 I-&gt;hasOneUse() &amp;&amp; SomeOtherThing(I)) {
566 ... some long code ....
567 }
568
569 return 0;
570}
571</pre>
572</div>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]573
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]574<p>This code has several problems if the body of the 'if' is large. When you're
575looking at the top of the function, it isn't immediately clear that this
576<em>only</em> does interesting things with non-terminator instructions, and only
577applies to things with the other predicates. Second, it is relatively difficult
578to describe (in comments) why these predicates are important because the if
579statement makes it difficult to lay out the comments. Third, when you're deep
580within the body of the code, it is indented an extra level. Finally, when
581reading the top of the function, it isn't clear what the result is if the
582predicate isn't true, you have to read to the end of the function to know that
583it returns null.</p>
Matthijs Kooijmaned58a972008-07-30 12:14:10 +0000[diff] [blame]584
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]585<p>It is much preferred to format the code like this:</p>
586
587<div class="doc_code">
588<pre>
589Value *DoSomething(Instruction *I) {
590 // Terminators never need 'something' done to them because, ...
591 if (isa&lt;TerminatorInst&gt;(I))
592 return 0;
593
594 // We conservatively avoid transforming instructions with multiple uses
595 // because goats like cheese.
596 if (!I-&gt;hasOneUse())
597 return 0;
598
599 // This is really just here for example.
600 if (!SomeOtherThing(I))
601 return 0;
602
603 ... some long code ....
604}
605</pre>
606</div>
607
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]608<p>This fixes these problems. A similar problem frequently happens in <tt>for</tt>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]609loops. A silly example is something like this:</p>
610
611<div class="doc_code">
612<pre>
613 for (BasicBlock::iterator II = BB-&gt;begin(), E = BB-&gt;end(); II != E; ++II) {
614 if (BinaryOperator *BO = dyn_cast&lt;BinaryOperator&gt;(II)) {
615 Value *LHS = BO-&gt;getOperand(0);
616 Value *RHS = BO-&gt;getOperand(1);
617 if (LHS != RHS) {
618 ...
619 }
620 }
621 }
622</pre>
623</div>
624
625<p>When you have very very small loops, this sort of structure is fine, but if
626it exceeds more than 10-15 lines, it becomes difficult for people to read and
627understand at a glance.
628The problem with this sort of code is that it gets very nested very quickly,
629meaning that the reader of the code has to keep a lot of context in their brain
630to remember what is going immediately on in the loop, because they don't know
631if/when the if conditions will have elses etc. It is strongly preferred to
632structure the loop like this:</p>
633
634<div class="doc_code">
635<pre>
636 for (BasicBlock::iterator II = BB-&gt;begin(), E = BB-&gt;end(); II != E; ++II) {
637 BinaryOperator *BO = dyn_cast&lt;BinaryOperator&gt;(II);
638 if (!BO) continue;
639
640 Value *LHS = BO-&gt;getOperand(0);
641 Value *RHS = BO-&gt;getOperand(1);
642 if (LHS == RHS) continue;
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]643 ...
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]644 }
645</pre>
646</div>
647
648<p>This has all the benefits of using early exits from functions: it reduces
649nesting of the loop, it makes it easier to describe why the conditions are true,
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]650and it makes it obvious to the reader that there is no <tt>else</tt> coming up that
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]651they have to push context into their brain for. If a loop is large, this can
652be a big understandability win.</p>
653
654</div>
655
Chris Lattner9eb7e0a2009-07-28 22:54:04 +0000[diff] [blame]656<!-- _______________________________________________________________________ -->
657<div class="doc_subsubsection">
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]658 <a name="hl_else_after_return">Don't use <tt>else</tt> after a <tt>return</tt></a>
Chris Lattner9eb7e0a2009-07-28 22:54:04 +0000[diff] [blame]659</div>
660
661<div class="doc_text">
662
663<p>For similar reasons above (reduction of indentation and easier reading),
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]664 please do not use <tt>else</tt> or '<tt>else if</tt>' after something that interrupts
665 control flow like <tt>return</tt>, <tt>break</tt>, <tt>continue</tt>, <tt>goto</tt>, etc. For example, this is
Chris Lattner9eb7e0a2009-07-28 22:54:04 +0000[diff] [blame]666 "bad":</p>
667
668<div class="doc_code">
669<pre>
670 case 'J': {
671 if (Signed) {
672 Type = Context.getsigjmp_bufType();
673 if (Type.isNull()) {
674 Error = ASTContext::GE_Missing_sigjmp_buf;
675 return QualType();
676 } else {
677 break;
678 }
679 } else {
680 Type = Context.getjmp_bufType();
681 if (Type.isNull()) {
682 Error = ASTContext::GE_Missing_jmp_buf;
683 return QualType();
684 } else {
685 break;
686 }
687 }
688 }
689 }
690</pre>
691</div>
692
693<p>It is better to write this something like:</p>
694
695<div class="doc_code">
696<pre>
697 case 'J':
698 if (Signed) {
699 Type = Context.getsigjmp_bufType();
700 if (Type.isNull()) {
701 Error = ASTContext::GE_Missing_sigjmp_buf;
702 return QualType();
703 }
704 } else {
705 Type = Context.getjmp_bufType();
706 if (Type.isNull()) {
707 Error = ASTContext::GE_Missing_jmp_buf;
708 return QualType();
709 }
710 }
711 break;
712</pre>
713</div>
714
715<p>Or better yet (in this case), as:</p>
716
717<div class="doc_code">
718<pre>
719 case 'J':
720 if (Signed)
721 Type = Context.getsigjmp_bufType();
722 else
723 Type = Context.getjmp_bufType();
724
725 if (Type.isNull()) {
726 Error = Signed ? ASTContext::GE_Missing_sigjmp_buf :
727 ASTContext::GE_Missing_jmp_buf;
728 return QualType();
729 }
730 break;
731</pre>
732</div>
733
734<p>The idea is to reduce indentation and the amount of code you have to keep
735 track of when reading the code.</p>
736
737</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]738
739<!-- _______________________________________________________________________ -->
740<div class="doc_subsubsection">
741 <a name="hl_predicateloops">Turn Predicate Loops into Predicate Functions</a>
742</div>
743
744<div class="doc_text">
745
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]746<p>It is very common to write small loops that just compute a boolean
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]747 value. There are a number of ways that people commonly write these, but an
748 example of this sort of thing is:</p>
749
750<div class="doc_code">
751<pre>
752 <b>bool FoundFoo = false;</b>
753 for (unsigned i = 0, e = BarList.size(); i != e; ++i)
754 if (BarList[i]-&gt;isFoo()) {
755 <b>FoundFoo = true;</b>
756 break;
757 }
758
759 <b>if (FoundFoo) {</b>
760 ...
761 }
762</pre>
763</div>
764
765<p>This sort of code is awkward to write, and is almost always a bad sign.
766Instead of this sort of loop, we strongly prefer to use a predicate function
767(which may be <a href="#micro_anonns">static</a>) that uses
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]768<a href="#hl_earlyexit">early exits</a> to compute the predicate. We prefer
769the code to be structured like this:
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]770</p>
771
772
773<div class="doc_code">
774<pre>
775/// ListContainsFoo - Return true if the specified list has an element that is
776/// a foo.
777static bool ListContainsFoo(const std::vector&lt;Bar*&gt; &amp;List) {
778 for (unsigned i = 0, e = List.size(); i != e; ++i)
779 if (List[i]-&gt;isFoo())
780 return true;
781 return false;
782}
783...
784
785 <b>if (ListContainsFoo(BarList)) {</b>
786 ...
787 }
788</pre>
789</div>
790
791<p>There are many reasons for doing this: it reduces indentation and factors out
792code which can often be shared by other code that checks for the same predicate.
793More importantly, it <em>forces you to pick a name</em> for the function, and
794forces you to write a comment for it. In this silly example, this doesn't add
795much value. However, if the condition is complex, this can make it a lot easier
796for the reader to understand the code that queries for this predicate. Instead
Chris Lattner71d8f3b2009-07-22 16:30:39 +0000[diff] [blame]797of being faced with the in-line details of how we check to see if the BarList
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]798contains a foo, we can trust the function name and continue reading with better
799locality.</p>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]800
801</div>
802
803
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]804<!-- ======================================================================= -->
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]805<div class="doc_subsection">
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]806 <a name="micro">The Low-Level Issues</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]807</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]808<!-- ======================================================================= -->
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]809
810
811<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]812<div class="doc_subsubsection">
Zhanyong Wan7fcd4dc2010-12-02 05:10:07 +0000[diff] [blame]813 <a name="ll_naming">Name Types, Functions, Variables, and Enumerators Properly</a>
814</div>
815
816<div class="doc_text">
817<p>Poorly-chosen names mislead the reader and cause bugs. We cannot
818stress enough how important it is to use <em>descriptive</em> names.
819Pick names that match the semantics and role of the underlying
820entities, within reason. Avoid abbreviations unless they are well
821known.</p>
822
823<p>In general, names of types, functions, variables, and enumerators
824should be in camel case (e.g. <tt>TextFileReader</tt>
825and <tt>isLValue()</tt>). Type names (including classes, structs,
826enums, typedefs, etc) should be nouns and start with an upper-case
Chris Lattner98bec512010-12-09 23:32:39 +0000[diff] [blame]827letter (e.g. <tt>TextFileReader</tt>). An <tt>enum</tt> for all the
828different kinds of something should be named with the <tt>Kind</tt>
829suffix. Function names should be verb phrases (as they represent
830actions) and start with a lower-case letter (e.g. a predicate may be
831named <tt>isFoo()</tt> or <tt>hasBar()</tt>, while the name of a
832command-like function should be imperative,
Zhanyong Wan7fcd4dc2010-12-02 05:10:07 +0000[diff] [blame]833like <tt>openFile()</tt>).</p>
834
835<p>Enumerators and public member variables should start with an
836upper-case letter, just like types. Unless the enumerators are
837defined in their own small namespace or inside a class, they should
838have a prefix. For example, <tt>enum ValueKind { ... };</tt> may
839contain enumerators like
840<tt>VK_Argument</tt>, <tt>VK_BasicBlock</tt>, etc. Enumerators that
841are just convenience constants are exempt from the requirement for a
842prefix. For instance:</p>
843<div class="doc_code">
844<pre>
845enum {
846 MaxSize = 42,
847 Density = 12
848};
849</pre>
850</div>
851
852<p>As an exception, classes that mimic STL classes can have member names
853in STL's style of lower-case words separated by underscores
854(e.g. <tt>begin()</tt>, <tt>push_back()</tt>, and <tt>empty()</tt>).</p>
855
856<p>Here are some examples of bad and good names:</p>
857<div class="doc_code">
858<pre>
859class VehicleMaker {
860 ...
861 Factory&lt;Tire&gt; f; // Bad -- abbreviation and non-descriptive.
862 Factory&lt;Tire&gt; factory; // Better.
863 Factory&lt;Tire&gt; tireFactory; // Even better -- if VehicleMaker has more than one
864 // kind of factories.
865};
866
867Vehicle MakeVehicle(VehicleType Type) {
868 VehicleMaker m; // Might be OK if having a short life-span.
869 Tire tmp1 = m.makeTire(); // Bad -- 'tmp1' provides no information.
870 Light headlight = m.makeLight("head"); // Good -- descriptive.
871 ...
872}
873</pre>
874</div>
875
876</div>
877
878
879<!-- _______________________________________________________________________ -->
880<div class="doc_subsubsection">
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]881 <a name="ll_assert">Assert Liberally</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]882</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]883
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]884<div class="doc_text">
885
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]886<p>Use the "<tt>assert</tt>" macro to its fullest. Check all of your
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]887preconditions and assumptions, you never know when a bug (not necessarily even
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]888yours) might be caught early by an assertion, which reduces debugging time
889dramatically. The "<tt>&lt;cassert&gt;</tt>" header file is probably already
890included by the header files you are using, so it doesn't cost anything to use
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]891it.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]892
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]893<p>To further assist with debugging, make sure to put some kind of error message
894in the assertion statement (which is printed if the assertion is tripped). This
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]895helps the poor debugger make sense of why an assertion is being made and
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]896enforced, and hopefully what to do about it. Here is one complete example:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]897
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]898<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]899<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]900inline Value *getOperand(unsigned i) {
901 assert(i &lt; Operands.size() &amp;&amp; "getOperand() out of range!");
902 return Operands[i];
903}
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]904</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]905</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]906
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]907<p>Here are some examples:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]908
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]909<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]910<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]911assert(Ty-&gt;isPointerType() &amp;&amp; "Can't allocate a non pointer type!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]912
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]913assert((Opcode == Shl || Opcode == Shr) &amp;&amp; "ShiftInst Opcode invalid!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]914
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]915assert(idx &lt; getNumSuccessors() &amp;&amp; "Successor # out of range!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]916
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]917assert(V1.getType() == V2.getType() &amp;&amp; "Constant types must be identical!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]918
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]919assert(isa&lt;PHINode&gt;(Succ-&gt;front()) &amp;&amp; "Only works on PHId BBs!");
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]920</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]921</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]922
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]923<p>You get the idea...</p>
924
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]925<p>Please be aware when adding assert statements that not all compilers are aware of
926the semantics of the assert. In some places, asserts are used to indicate a piece of
927code that should not be reached. These are typically of the form:</p>
928
929<div class="doc_code">
930<pre>
Dan Gohmanf0032762008-11-24 17:18:39 +0000[diff] [blame]931assert(0 &amp;&amp; "Some helpful error message");
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]932</pre>
933</div>
934
935<p>When used in a function that returns a value, they should be followed with a return
936statement and a comment indicating that this line is never reached. This will prevent
937a compiler which is unable to deduce that the assert statement never returns from
938generating a warning.</p>
939
940<div class="doc_code">
941<pre>
Dan Gohmanf0032762008-11-24 17:18:39 +0000[diff] [blame]942assert(0 &amp;&amp; "Some helpful error message");
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]943// Not reached
944return 0;
945</pre>
946</div>
947
Chris Lattner97928d12010-11-12 00:19:41 +0000[diff] [blame]948<p>Another issue is that values used only by assertions will produce an "unused
949 value" warning when assertions are disabled. For example, this code will warn:
950</p>
951
952<div class="doc_code">
953<pre>
954 unsigned Size = V.size();
955 assert(Size &gt; 42 &amp;&amp; "Vector smaller than it should be");
956
957 bool NewToSet = Myset.insert(Value);
958 assert(NewToSet &amp;&amp; "The value shouldn't be in the set yet");
959</pre>
960</div>
961
962<p>These are two interesting different cases: in the first case, the call to
963V.size() is only useful for the assert, and we don't want it executed when
964assertions are disabled. Code like this should move the call into the assert
965itself. In the second case, the side effects of the call must happen whether
966the assert is enabled or not. In this case, the value should be cast to void
967to disable the warning. To be specific, it is preferred to write the code
968like this:</p>
969
970<div class="doc_code">
971<pre>
972 assert(V.size() &gt; 42 &amp;&amp; "Vector smaller than it should be");
973
974 bool NewToSet = Myset.insert(Value); (void)NewToSet;
975 assert(NewToSet &amp;&amp; "The value shouldn't be in the set yet");
976</pre>
977</div>
978
979
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]980</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]981
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]982<!-- _______________________________________________________________________ -->
983<div class="doc_subsubsection">
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]984 <a name="ll_ns_std">Do not use '<tt>using namespace std</tt>'</a>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]985</div>
986
987<div class="doc_text">
988<p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]989namespace with an "<tt>std::</tt>" prefix, rather than rely on
990"<tt>using namespace std;</tt>".</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]991
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]992<p> In header files, adding a '<tt>using namespace XXX</tt>' directive pollutes
Misha Brukman2d33c542008-12-11 19:37:04 +0000[diff] [blame]993the namespace of any source file that <tt>#include</tt>s the header. This is
994clearly a bad thing.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]995
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]996<p>In implementation files (e.g. <tt>.cpp</tt> files), the rule is more of a stylistic
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]997rule, but is still important. Basically, using explicit namespace prefixes
998makes the code <b>clearer</b>, because it is immediately obvious what facilities
999are being used and where they are coming from, and <b>more portable</b>, because
1000namespace clashes cannot occur between LLVM code and other namespaces. The
1001portability rule is important because different standard library implementations
1002expose different symbols (potentially ones they shouldn't), and future revisions
1003to the C++ standard will add more symbols to the <tt>std</tt> namespace. As
1004such, we never use '<tt>using namespace std;</tt>' in LLVM.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]1005
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]1006<p>The exception to the general rule (i.e. it's not an exception for
1007the <tt>std</tt> namespace) is for implementation files. For example, all of
1008the code in the LLVM project implements code that lives in the 'llvm' namespace.
Chris Lattnereddd9692010-11-17 19:47:20 +0000[diff] [blame]1009As such, it is ok, and actually clearer, for the <tt>.cpp</tt> files to have a
1010'<tt>using namespace llvm</tt>' directive at their top, after the
1011<tt>#include</tt>s. This reduces indentation in the body of the file for source
1012editors that indent based on braces, and keeps the conceptual context cleaner.
1013The general form of this rule is that any <tt>.cpp</tt> file that implements
1014code in any namespace may use that namespace (and its parents'), but should not
1015use any others.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]1016
1017</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1018
1019<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1020<div class="doc_subsubsection">
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]1021 <a name="ll_virtual_anch">Provide a virtual method anchor for classes
1022 in headers</a>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1023</div>
1024
1025<div class="doc_text">
1026
1027<p>If a class is defined in a header file and has a v-table (either it has
1028virtual methods or it derives from classes with virtual methods), it must
1029always have at least one out-of-line virtual method in the class. Without
Misha Brukman2d33c542008-12-11 19:37:04 +0000[diff] [blame]1030this, the compiler will copy the vtable and RTTI into every <tt>.o</tt> file
1031that <tt>#include</tt>s the header, bloating <tt>.o</tt> file sizes and
1032increasing link times.</p>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1033
1034</div>
1035
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]1036<!-- _______________________________________________________________________ -->
1037<div class="doc_subsubsection">
Zhanyong Wancafe0b42010-11-23 05:03:07 +0000[diff] [blame]1038 <a name="ll_end">Don't evaluate <tt>end()</tt> every time through a loop</a>
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]1039</div>
1040
1041<div class="doc_text">
1042
1043<p>Because C++ doesn't have a standard "foreach" loop (though it can be emulated
1044with macros and may be coming in C++'0x) we end up writing a lot of loops that
1045manually iterate from begin to end on a variety of containers or through other
1046data structures. One common mistake is to write a loop in this style:</p>
1047
1048<div class="doc_code">
1049<pre>
1050 BasicBlock *BB = ...
1051 for (BasicBlock::iterator I = BB->begin(); I != <b>BB->end()</b>; ++I)
1052 ... use I ...
1053</pre>
1054</div>
1055
1056<p>The problem with this construct is that it evaluates "<tt>BB->end()</tt>"
1057every time through the loop. Instead of writing the loop like this, we strongly
1058prefer loops to be written so that they evaluate it once before the loop starts.
1059A convenient way to do this is like so:</p>
1060
1061<div class="doc_code">
1062<pre>
1063 BasicBlock *BB = ...
1064 for (BasicBlock::iterator I = BB->begin(), E = <b>BB->end()</b>; I != E; ++I)
1065 ... use I ...
1066</pre>
1067</div>
1068
1069<p>The observant may quickly point out that these two loops may have different
1070semantics: if the container (a basic block in this case) is being mutated, then
1071"<tt>BB->end()</tt>" may change its value every time through the loop and the
1072second loop may not in fact be correct. If you actually do depend on this
Chris Lattner09bc65e2009-06-30 06:20:03 +0000[diff] [blame]1073behavior, please write the loop in the first form and add a comment indicating
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]1074that you did it intentionally.</p>
1075
1076<p>Why do we prefer the second form (when correct)? Writing the loop in the
1077first form has two problems: First it may be less efficient than evaluating it
1078at the start of the loop. In this case, the cost is probably minor: a few extra
1079loads every time through the loop. However, if the base expression is more
1080complex, then the cost can rise quickly. I've seen loops where the end
1081expression was actually something like: "<tt>SomeMap[x]->end()</tt>" and map
Chris Lattnera44f87f2009-06-30 06:27:54 +0000[diff] [blame]1082lookups really aren't cheap. By writing it in the second form consistently, you
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]1083eliminate the issue entirely and don't even have to think about it.</p>
1084
Chris Lattner09bc65e2009-06-30 06:20:03 +0000[diff] [blame]1085<p>The second (even bigger) issue is that writing the loop in the first form
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]1086hints to the reader that the loop is mutating the container (a fact that a
1087comment would handily confirm!). If you write the loop in the second form, it
1088is immediately obvious without even looking at the body of the loop that the
1089container isn't being modified, which makes it easier to read the code and
1090understand what it does.</p>
1091
1092<p>While the second form of the loop is a few extra keystrokes, we do strongly
1093prefer it.</p>
1094
1095</div>
1096
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1097<!-- _______________________________________________________________________ -->
1098<div class="doc_subsubsection">
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1099 <a name="ll_iostream"><tt>#include &lt;iostream&gt;</tt> is forbidden</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1100</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1101
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1102<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1103
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1104<p>The use of <tt>#include &lt;iostream&gt;</tt> in library files is
1105hereby <b><em>forbidden</em></b>. The primary reason for doing this is to
1106support clients using LLVM libraries as part of larger systems. In particular,
1107we statically link LLVM into some dynamic libraries. Even if LLVM isn't used,
1108the static c'tors are run whenever an application start up that uses the dynamic
1109library. There are two problems with this:</p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1110
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1111<ol>
1112 <li>The time to run the static c'tors impacts startup time of
1113 applications&mdash;a critical time for GUI apps.</li>
1114 <li>The static c'tors cause the app to pull many extra pages of memory off the
1115 disk: both the code for the static c'tors in each <tt>.o</tt> file and the
1116 small amount of data that gets touched. In addition, touched/dirty pages
1117 put more pressure on the VM system on low-memory machines.</li>
1118</ol>
1119
1120<p>Note that using the other stream headers (<tt>&lt;sstream&gt;</tt> for
Chris Lattner983c5922009-08-23 21:53:47 +0000[diff] [blame]1121example) is not problematic in this regard (just <tt>&lt;iostream&gt;</tt>).
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]1122However, <tt>raw_ostream</tt> provides various APIs that are better performing for almost
1123every use than <tt>std::ostream</tt> style APIs.
1124<b>Therefore new code should always
Chris Lattner983c5922009-08-23 21:53:47 +0000[diff] [blame]1125use <a href="#ll_raw_ostream"><tt>raw_ostream</tt></a> for writing, or
1126the <tt>llvm::MemoryBuffer</tt> API for reading files.</b></p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1127
1128</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1129
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1130
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1131<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1132<div class="doc_subsubsection">
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]1133 <a name="ll_raw_ostream">Use <tt>raw_ostream</tt></a>
1134</div>
1135
1136<div class="doc_text">
1137
1138<p>LLVM includes a lightweight, simple, and efficient stream implementation
1139in <tt>llvm/Support/raw_ostream.h</tt> which provides all of the common features
1140of <tt>std::ostream</tt>. All new code should use <tt>raw_ostream</tt> instead
1141of <tt>ostream</tt>.</p>
1142
1143<p>Unlike <tt>std::ostream</tt>, <tt>raw_ostream</tt> is not a template and can
1144be forward declared as <tt>class raw_ostream</tt>. Public headers should
1145generally not include the <tt>raw_ostream</tt> header, but use forward
1146declarations and constant references to <tt>raw_ostream</tt> instances.</p>
1147
1148</div>
1149
1150
1151<!-- _______________________________________________________________________ -->
1152<div class="doc_subsubsection">
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1153 <a name="ll_avoidendl">Avoid <tt>std::endl</tt></a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1154</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1155
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1156<div class="doc_text">
1157
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]1158<p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
1159to the output stream specified. In addition to doing this, however, it also
1160flushes the output stream. In other words, these are equivalent:</p>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1161
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1162<div class="doc_code">
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1163<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1164std::cout &lt;&lt; std::endl;
1165std::cout &lt;&lt; '\n' &lt;&lt; std::flush;
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1166</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1167</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1168
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1169<p>Most of the time, you probably have no reason to flush the output stream, so
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1170it's better to use a literal <tt>'\n'</tt>.</p>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1171
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1172</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1173
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]1174
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1175<!-- ======================================================================= -->
1176<div class="doc_subsection">
1177 <a name="nano">Microscopic Details</a>
1178</div>
1179<!-- ======================================================================= -->
1180
1181<p>This section describes preferred low-level formatting guidelines along with
1182reasoning on why we prefer them.</p>
1183
1184<!-- _______________________________________________________________________ -->
1185<div class="doc_subsubsection">
1186 <a name="micro_spaceparen">Spaces Before Parentheses</a>
1187</div>
1188
1189<div class="doc_text">
1190
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]1191<p>We prefer to put a space before an open parenthesis only in control flow
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1192statements, but not in normal function call expressions and function-like
1193macros. For example, this is good:</p>
1194
1195<div class="doc_code">
1196<pre>
1197 <b>if (</b>x) ...
1198 <b>for (</b>i = 0; i != 100; ++i) ...
1199 <b>while (</b>llvm_rocks) ...
1200
1201 <b>somefunc(</b>42);
1202 <b><a href="#ll_assert">assert</a>(</b>3 != 4 &amp;&amp; "laws of math are failing me");
1203
1204 a = <b>foo(</b>42, 92) + <b>bar(</b>x);
1205 </pre>
1206</div>
1207
1208<p>... and this is bad:</p>
1209
1210<div class="doc_code">
1211<pre>
1212 <b>if(</b>x) ...
1213 <b>for(</b>i = 0; i != 100; ++i) ...
1214 <b>while(</b>llvm_rocks) ...
1215
1216 <b>somefunc (</b>42);
1217 <b><a href="#ll_assert">assert</a> (</b>3 != 4 &amp;&amp; "laws of math are failing me");
1218
1219 a = <b>foo (</b>42, 92) + <b>bar (</b>x);
1220</pre>
1221</div>
1222
1223<p>The reason for doing this is not completely arbitrary. This style makes
1224 control flow operators stand out more, and makes expressions flow better. The
1225 function call operator binds very tightly as a postfix operator. Putting
1226 a space after a function name (as in the last example) makes it appear that
1227 the code might bind the arguments of the left-hand-side of a binary operator
1228 with the argument list of a function and the name of the right side. More
1229 specifically, it is easy to misread the "a" example as:</p>
1230
1231<div class="doc_code">
1232<pre>
1233 a = foo <b>(</b>(42, 92) + bar<b>)</b> (x);
1234</pre>
1235</div>
1236
1237<p>... when skimming through the code. By avoiding a space in a function, we
1238avoid this misinterpretation.</p>
1239
1240</div>
1241
1242<!-- _______________________________________________________________________ -->
1243<div class="doc_subsubsection">
1244 <a name="micro_preincrement">Prefer Preincrement</a>
1245</div>
1246
1247<div class="doc_text">
1248
1249<p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
1250postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
1251preincrementation whenever possible.</p>
1252
1253<p>The semantics of postincrement include making a copy of the value being
1254incremented, returning it, and then preincrementing the "work value". For
1255primitive types, this isn't a big deal... but for iterators, it can be a huge
1256issue (for example, some iterators contains stack and set objects in them...
1257copying an iterator could invoke the copy ctor's of these as well). In general,
1258get in the habit of always using preincrement, and you won't have a problem.</p>
1259
1260</div>
1261
1262<!-- _______________________________________________________________________ -->
1263<div class="doc_subsubsection">
1264 <a name="micro_namespaceindent">Namespace Indentation</a>
1265</div>
1266
1267<div class="doc_text">
1268
1269<p>
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]1270In general, we strive to reduce indentation wherever possible. This is useful
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1271because we want code to <a href="#scf_codewidth">fit into 80 columns</a> without
1272wrapping horribly, but also because it makes it easier to understand the code.
1273Namespaces are a funny thing: they are often large, and we often desire to put
1274lots of stuff into them (so they can be large). Other times they are tiny,
1275because they just hold an enum or something similar. In order to balance this,
1276we use different approaches for small versus large namespaces.
1277</p>
1278
1279<p>
1280If a namespace definition is small and <em>easily</em> fits on a screen (say,
1281less than 35 lines of code), then you should indent its body. Here's an
1282example:
1283</p>
1284
1285<div class="doc_code">
1286<pre>
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1287namespace llvm {
1288 namespace X86 {
1289 /// RelocationType - An enum for the x86 relocation codes. Note that
1290 /// the terminology here doesn't follow x86 convention - word means
1291 /// 32-bit and dword means 64-bit.
1292 enum RelocationType {
1293 /// reloc_pcrel_word - PC relative relocation, add the relocated value to
1294 /// the value already in memory, after we adjust it for where the PC is.
1295 reloc_pcrel_word = 0,
1296
1297 /// reloc_picrel_word - PIC base relative relocation, add the relocated
1298 /// value to the value already in memory, after we adjust it for where the
1299 /// PIC base is.
1300 reloc_picrel_word = 1,
1301
1302 /// reloc_absolute_word, reloc_absolute_dword - Absolute relocation, just
1303 /// add the relocated value to the value already in memory.
1304 reloc_absolute_word = 2,
1305 reloc_absolute_dword = 3
1306 };
1307 }
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1308}
1309</pre>
1310</div>
1311
1312<p>Since the body is small, indenting adds value because it makes it very clear
1313where the namespace starts and ends, and it is easy to take the whole thing in
1314in one "gulp" when reading the code. If the blob of code in the namespace is
Chris Lattner684cf812010-11-17 17:14:55 +0000[diff] [blame]1315larger (as it typically is in a header in the <tt>llvm</tt> or <tt>clang</tt> namespaces), do not
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1316indent the code, and add a comment indicating what namespace is being closed.
1317For example:</p>
1318
1319<div class="doc_code">
1320<pre>
1321namespace llvm {
1322namespace knowledge {
1323
1324/// Grokable - This class represents things that Smith can have an intimate
1325/// understanding of and contains the data associated with it.
1326class Grokable {
1327...
1328public:
1329 explicit Grokable() { ... }
1330 virtual ~Grokable() = 0;
1331
1332 ...
1333
1334};
1335
1336} // end namespace knowledge
1337} // end namespace llvm
1338</pre>
1339</div>
1340
1341<p>Because the class is large, we don't expect that the reader can easily
1342understand the entire concept in a glance, and the end of the file (where the
1343namespaces end) may be a long ways away from the place they open. As such,
1344indenting the contents of the namespace doesn't add any value, and detracts from
1345the readability of the class. In these cases it is best to <em>not</em> indent
1346the contents of the namespace.</p>
1347
1348</div>
1349
1350<!-- _______________________________________________________________________ -->
1351<div class="doc_subsubsection">
1352 <a name="micro_anonns">Anonymous Namespaces</a>
1353</div>
1354
1355<div class="doc_text">
1356
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1357<p>After talking about namespaces in general, you may be wondering about
1358anonymous namespaces in particular.
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1359Anonymous namespaces are a great language feature that tells the C++ compiler
1360that the contents of the namespace are only visible within the current
1361translation unit, allowing more aggressive optimization and eliminating the
1362possibility of symbol name collisions. Anonymous namespaces are to C++ as
1363"static" is to C functions and global variables. While "static" is available
1364in C++, anonymous namespaces are more general: they can make entire classes
1365private to a file.</p>
1366
1367<p>The problem with anonymous namespaces is that they naturally want to
1368encourage indentation of their body, and they reduce locality of reference: if
1369you see a random function definition in a C++ file, it is easy to see if it is
1370marked static, but seeing if it is in an anonymous namespace requires scanning
1371a big chunk of the file.</p>
1372
1373<p>Because of this, we have a simple guideline: make anonymous namespaces as
1374small as possible, and only use them for class declarations. For example, this
1375is good:</p>
1376
1377<div class="doc_code">
1378<pre>
1379<b>namespace {</b>
1380 class StringSort {
1381 ...
1382 public:
1383 StringSort(...)
1384 bool operator&lt;(const char *RHS) const;
1385 };
1386<b>} // end anonymous namespace</b>
1387
1388static void Helper() {
1389 ...
1390}
1391
1392bool StringSort::operator&lt;(const char *RHS) const {
1393 ...
1394}
1395
1396</pre>
1397</div>
1398
1399<p>This is bad:</p>
1400
1401
1402<div class="doc_code">
1403<pre>
1404<b>namespace {</b>
1405class StringSort {
1406...
1407public:
1408 StringSort(...)
1409 bool operator&lt;(const char *RHS) const;
1410};
1411
1412void Helper() {
1413 ...
1414}
1415
1416bool StringSort::operator&lt;(const char *RHS) const {
1417 ...
1418}
1419
1420<b>} // end anonymous namespace</b>
1421
1422</pre>
1423</div>
1424
1425
1426<p>This is bad specifically because if you're looking at "Helper" in the middle
1427of a large C++ file, that you have no immediate way to tell if it is local to
1428the file. When it is marked static explicitly, this is immediately obvious.
1429Also, there is no reason to enclose the definition of "operator&lt;" in the
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1430namespace just because it was declared there.
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1431</p>
1432
1433</div>
1434
1435
1436
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1437<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1438<div class="doc_section">
1439 <a name="seealso">See Also</a>
1440</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1441<!-- *********************************************************************** -->
1442
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1443<div class="doc_text">
1444
1445<p>A lot of these comments and recommendations have been culled for other
1446sources. Two particularly important books for our work are:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1447
1448<ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1449
Chris Lattner788ee282007-11-09 21:49:08 +0000[diff] [blame]1450<li><a href="http://www.amazon.com/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0321334876">Effective
1451C++</a> by Scott Meyers. Also
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]1452interesting and useful are "More Effective C++" and "Effective STL" by the same
1453author.</li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1454
Chris Lattner788ee282007-11-09 21:49:08 +0000[diff] [blame]1455<li>Large-Scale C++ Software Design by John Lakos</li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1456
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1457</ol>
1458
1459<p>If you get some free time, and you haven't read them: do so, you might learn
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]1460something.</p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1461
1462</div>
1463
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1464<!-- *********************************************************************** -->
1465
1466<hr>
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1467<address>
1468 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1469 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1470 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1471 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1472
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1473 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1474 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1475 Last modified: $Date$
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1476</address>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1477
1478</body>
1479</html>