Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 80f9e6724fa98b5af0bb43f58002c1ac3e9cb30b / . / docs / CodingStandards.html
blob: 416c29b7bedd35db687766745640b37bd95c393b [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>
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]32 <li><a href="#ci_class_struct">Use of class/struct Keywords</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]33 </ol></li>
34 </ol></li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]35 <li><a href="#styleissues">Style Issues</a>
36 <ol>
37 <li><a href="#macro">The High Level Issues</a>
38 <ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]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>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]44 <li><a href="#hl_earlyexit">Use Early Exits and 'continue' to Simplify
45 Code</a></li>
Chris Lattner9eb7e0a2009-07-28 22:54:04 +0000[diff] [blame]46 <li><a href="#hl_else_after_return">Don't use "else" after a
47 return</a></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]48 <li><a href="#hl_predicateloops">Turn Predicate Loops into Predicate
49 Functions</a></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]50 </ol></li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]51 <li><a href="#micro">The Low Level Issues</a>
52 <ol>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]53 <li><a href="#ll_assert">Assert Liberally</a></li>
54 <li><a href="#ll_ns_std">Do not use 'using namespace std'</a></li>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]55 <li><a href="#ll_virtual_anch">Provide a virtual method anchor for
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]56 classes in headers</a></li>
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]57 <li><a href="#ll_end">Don't evaluate end() every time through a
58 loop</a></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]59 <li><a href="#ll_iostream"><tt>#include &lt;iostream&gt;</tt> is
60 <em>forbidden</em></a></li>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]61 <li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
Daniel Dunbare8530a32009-07-24 23:04:51 +0000[diff] [blame]62 <li><a href="#ll_raw_ostream">Use <tt>raw_ostream</tt></a</li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]63 </ol></li>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]64
65 <li><a href="#nano">Microscopic Details</a>
66 <ol>
67 <li><a href="#micro_spaceparen">Spaces Before Parentheses</a></li>
68 <li><a href="#micro_preincrement">Prefer Preincrement</a></li>
69 <li><a href="#micro_namespaceindent">Namespace Indentation</a></li>
70 <li><a href="#micro_anonns">Anonymous Namespaces</a></li>
71 </ol></li>
72
73
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]74 </ol></li>
75 <li><a href="#seealso">See Also</a></li>
76</ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]77
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]78<div class="doc_author">
Chris Lattner219bd292009-07-22 05:43:01 +0000[diff] [blame]79 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a></p>
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]80</div>
81
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]82
83<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]84<div class="doc_section">
85 <a name="introduction">Introduction</a>
86</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]87<!-- *********************************************************************** -->
88
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]89<div class="doc_text">
90
91<p>This document attempts to describe a few coding standards that are being used
92in the LLVM source tree. Although no coding standards should be regarded as
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]93absolute requirements to be followed in all instances, coding standards can be
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]94useful.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]95
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]96<p>This document intentionally does not prescribe fixed standards for religious
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]97issues such as brace placement and space usage. For issues like this, follow
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]98the golden rule:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]99
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]100<blockquote>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]101
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]102<p><b><a name="goldenrule">If you are adding a significant body of source to a
103project, feel free to use whatever style you are most comfortable with. If you
104are extending, enhancing, or bug fixing already implemented code, use the style
105that is already being used so that the source is uniform and easy to
106follow.</a></b></p>
107
108</blockquote>
109
110<p>The ultimate goal of these guidelines is the increase readability and
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]111maintainability of our common source base. If you have suggestions for topics to
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]112be included, please mail them to <a
113href="mailto:sabre@nondot.org">Chris</a>.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]114
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]115</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]116
117<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]118<div class="doc_section">
119 <a name="mechanicalissues">Mechanical Source Issues</a>
120</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]121<!-- *********************************************************************** -->
122
123<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]124<div class="doc_subsection">
125 <a name="sourceformating">Source Code Formatting</a>
126</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]127
128<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]129<div class="doc_subsubsection">
130 <a name="scf_commenting">Commenting</a>
131</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]132
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]133<div class="doc_text">
134
135<p>Comments are one critical part of readability and maintainability. Everyone
Chris Lattnerb927ca82009-07-12 00:10:24 +0000[diff] [blame]136knows they should comment, so should you. When writing comments, write them as
137English prose, which means they should use proper capitalization, punctuation,
138etc. Although we all should probably
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]139comment our code more than we do, there are a few very critical places that
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]140documentation is very useful:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]141
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]142<b>File Headers</b>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]143
Reid Spencer669ed452007-07-09 08:04:31 +0000[diff] [blame]144<p>Every source file should have a header on it that describes the basic
145purpose of the file. If a file does not have a header, it should not be
146checked into Subversion. Most source trees will probably have a standard
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]147file header format. The standard format for the LLVM source tree looks like
148this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]149
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]150<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]151<pre>
Chris Lattnerc49fcd12003-10-13 14:58:11 +0000[diff] [blame]152//===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
Misha Brukman98117cd2009-01-02 16:58:42 +0000[diff] [blame]153//
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]154// The LLVM Compiler Infrastructure
155//
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]156// This file is distributed under the University of Illinois Open Source
157// License. See LICENSE.TXT for details.
Misha Brukman98117cd2009-01-02 16:58:42 +0000[diff] [blame]158//
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]159//===----------------------------------------------------------------------===//
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]160//
161// This file contains the declaration of the Instruction class, which is the
162// base class for all of the VM instructions.
163//
164//===----------------------------------------------------------------------===//
165</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]166</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]167
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]168<p>A few things to note about this particular format: The "<tt>-*- C++
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]169-*-</tt>" string on the first line is there to tell Emacs that the source file
Misha Brukman69c58892004-07-28 22:37:57 +0000[diff] [blame]170is a C++ file, not a C file (Emacs assumes .h files are C files by default).
171Note that this tag is not necessary in .cpp files. The name of the file is also
172on the first line, along with a very short description of the purpose of the
173file. This is important when printing out code and flipping though lots of
174pages.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]175
Chris Lattnerab2b10c2007-12-29 19:56:08 +0000[diff] [blame]176<p>The next section in the file is a concise note that defines the license
177that the file is released under. This makes it perfectly clear what terms the
178source code can be distributed under and should not be modified in any way.</p>
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]179
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]180<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]181Here it's only two lines. If an algorithm is being implemented or something
182tricky is going on, a reference to the paper where it is published should be
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]183included, 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]184
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]185<b>Class overviews</b>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]186
Jim Laskeyf55914a2006-07-31 20:18:49 +0000[diff] [blame]187<p>Classes are one fundamental part of a good object oriented design. As such,
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]188a class definition should have a comment block that explains what the class is
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]189used for... if it's not obvious. If it's so completely obvious your grandma
190could figure it out, it's probably safe to leave it out. Naming classes
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]191something sane goes a long ways towards avoiding writing documentation.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]192
193
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]194<b>Method information</b>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]195
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]196<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]197documented properly. A quick note about what it does and a description of the
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]198borderline behaviour is all that is necessary here (unless something
Dan Gohman621a4152010-02-26 20:18:32 +0000[diff] [blame]199particularly tricky or insidious is going on). The hope is that people can
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]200figure out how to use your interfaces without reading the code itself... that is
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]201the goal metric.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]202
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]203<p>Good things to talk about here are what happens when something unexpected
204happens: does the method return null? Abort? Format your hard disk?</p>
205
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]206</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]207
208<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]209<div class="doc_subsubsection">
210 <a name="scf_commentformat">Comment Formatting</a>
211</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]212
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]213<div class="doc_text">
214
215<p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]216require less typing, don't have nesting problems, etc. There are a few cases
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]217when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]218
219<ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]220 <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]221 comments.</li>
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]222 <li>When writing a header file that may be <tt>#include</tt>d by a C source
223 file.</li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]224 <li>When writing a source file that is used by a tool that only accepts C
225 style comments.</li>
226</ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]227
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]228<p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
229These nest properly and are better behaved in general than C style comments.</p>
230
231</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]232
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]233<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]234<div class="doc_subsubsection">
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]235 <a name="scf_includes"><tt>#include</tt> Style</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]236</div>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]237
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]238<div class="doc_text">
239
240<p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]241include guards if working on a header file), the <a
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]242href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
243file should be listed. We prefer these <tt>#include</tt>s to be listed in this
244order:</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]245
246<ol>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]247 <li><a href="#mmheader">Main Module header</a></li>
248 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]249 <li><tt>llvm/*</tt></li>
250 <li><tt>llvm/Analysis/*</tt></li>
251 <li><tt>llvm/Assembly/*</tt></li>
Duncan Sands18d52f22010-09-29 20:09:55 +0000[diff] [blame]252 <li><tt>llvm/Bitcode/*</tt></li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]253 <li><tt>llvm/CodeGen/*</tt></li>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]254 <li>...</li>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]255 <li><tt>Support/*</tt></li>
256 <li><tt>Config/*</tt></li>
257 <li>System <tt>#includes</tt></li>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]258</ol>
259
Nick Lewycky822293a2008-11-29 20:13:25 +0000[diff] [blame]260<p>... and each category should be sorted by name.</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]261
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]262<p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]263which implement an interface defined by a .h file. This <tt>#include</tt>
264should always be included <b>first</b> regardless of where it lives on the file
265system. By including a header file first in the .cpp files that implement the
266interfaces, we ensure that the header does not have any hidden dependencies
267which are not explicitly #included in the header, but should be. It is also a
268form of documentation in the .cpp file to indicate where the interfaces it
269implements are defined.</p>
Chris Lattner2e5cbc22003-08-07 21:45:47 +0000[diff] [blame]270
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]271</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]272
273<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]274<div class="doc_subsubsection">
275 <a name="scf_codewidth">Source Code Width</a>
276</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]277
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]278<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]279
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]280<p>Write your code to fit within 80 columns of text. This helps those of us who
281like to print out code and look at your code in an xterm without resizing
282it.</p>
283
Chris Lattner5822e9e2008-07-08 05:12:37 +0000[diff] [blame]284<p>The longer answer is that there must be some limit to the width of the code
285in order to reasonably allow developers to have multiple files side-by-side in
286windows on a modest display. If you are going to pick a width limit, it is
287somewhat arbitrary but you might as well pick something standard. Going with
28890 columns (for example) instead of 80 columns wouldn't add any significant
289value and would be detrimental to printing out code. Also many other projects
290have standardized on 80 columns, so some people have already configured their
291editors for it (vs something else, like 90 columns).</p>
292
293<p>This is one of many contentious issues in coding standards, but is not up
294for debate.</p>
295
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]296</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]297
298<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]299<div class="doc_subsubsection">
300 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
301</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]302
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]303<div class="doc_text">
304
305<p>In all cases, prefer spaces to tabs in source files. People have different
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]306preferred indentation levels, and different styles of indentation that they
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]307like... this is fine. What isn't is that different editors/viewers expand tabs
308out to different tab stops. This can cause your code to look completely
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]309unreadable, and it is not worth dealing with.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]310
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]311<p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]312style of existing code if your are modifying and extending it. If you like four
313spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
314with two spaces of indentation. Also, do not reindent a whole source file: it
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]315makes for incredible diffs that are absolutely worthless.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]316
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]317</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]318
319<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]320<div class="doc_subsubsection">
321 <a name="scf_indentation">Indent Code Consistently</a>
322</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]323
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]324<div class="doc_text">
325
326<p>Okay, your first year of programming you were told that indentation is
Misha Brukman1b36d672003-10-06 19:26:00 +0000[diff] [blame]327important. If you didn't believe and internalize this then, now is the time.
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]328Just do it.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]329
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]330</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]331
332
333<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]334<div class="doc_subsection">
335 <a name="compilerissues">Compiler Issues</a>
336</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]337
338
339<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]340<div class="doc_subsubsection">
341 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
342</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]343
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]344<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]345
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]346<p>If your code has compiler warnings in it, something is wrong: you aren't
347casting values correctly, your have "questionable" constructs in your code, or
348you are doing something legitimately wrong. Compiler warnings can cover up
349legitimate errors in output and make dealing with a translation unit
350difficult.</p>
351
352<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]353desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
354a good thorough set of warnings, and stick to them. At least in the case of
355<tt>gcc</tt>, it is possible to work around any spurious errors by changing the
356syntax of the code slightly. For example, an warning that annoys me occurs when
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]357I write code like this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]358
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]359<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]360<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]361if (V = getValue()) {
362 ...
363}
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]364</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]365</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]366
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]367<p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
368operator, and that I probably mistyped it. In most cases, I haven't, and I
369really don't want the spurious errors. To fix this particular problem, I
370rewrite the code like this:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]371
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]372<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]373<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]374if ((V = getValue())) {
375 ...
376}
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]377</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]378</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]379
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]380<p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
381be fixed by massaging the code appropriately.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]382
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]383<p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
384-Winline -W -Wwrite-strings -Wno-unused</tt></p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]385
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]386</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]387
388<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]389<div class="doc_subsubsection">
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]390 <a name="ci_portable_code">Write Portable Code</a>
391</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]392
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]393<div class="doc_text">
394
395<p>In almost all cases, it is possible and within reason to write completely
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]396portable code. If there are cases where it isn't possible to write portable
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]397code, isolate it behind a well defined (and well documented) interface.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]398
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]399<p>In practice, this means that you shouldn't assume much about the host
400compiler, including its support for "high tech" features like partial
Chris Lattner4bd3d7e2009-03-23 04:52:53 +0000[diff] [blame]401specialization of templates. If these features are used, they should only be
402an implementation detail of a library which has a simple exposed API.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]403
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]404</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]405
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]406<!-- _______________________________________________________________________ -->
407<div class="doc_subsubsection">
408<a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
409</div>
410<div class="doc_text">
Misha Brukmanedcc78a2004-10-26 15:45:13 +0000[diff] [blame]411
412<p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
413interchangeably. The only difference is when they are used to declare a class:
414<tt>class</tt> makes all members private by default while <tt>struct</tt> makes
415all members public by default.</p>
416
417<p>Unfortunately, not all compilers follow the rules and some will generate
418different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
419declare the symbol. This can lead to problems at link time.</p>
420
421<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]422<b>all</b> members are public and the type is a C++ "POD" type, in which case
423<tt>struct</tt> is allowed.</p>
Misha Brukmanedcc78a2004-10-26 15:45:13 +0000[diff] [blame]424
Reid Spencerb40df562004-09-23 16:03:48 +0000[diff] [blame]425</div>
426
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]427<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]428<div class="doc_section">
429 <a name="styleissues">Style Issues</a>
430</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]431<!-- *********************************************************************** -->
432
433
434<!-- ======================================================================= -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]435<div class="doc_subsection">
436 <a name="macro">The High Level Issues</a>
437</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]438<!-- ======================================================================= -->
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]439
440
441<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]442<div class="doc_subsubsection">
443 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
444</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]445
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]446<div class="doc_text">
447
448<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]449encapsulation or data hiding (unless you use expensive protocol classes), but it
450is what we have to work with. When you write a public header file (in the LLVM
451source tree, they live in the top level "include" directory), you are defining a
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]452module of functionality.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]453
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]454<p>Ideally, modules should be completely independent of each other, and their
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]455header files should only include the absolute minimum number of headers
456possible. A module is not just a class, a function, or a namespace: <a
457href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
458of these</a> that defines an interface. This interface may be several
459functions, classes or data structures, but the important issue is how they work
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]460together.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]461
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]462<p>In general, a module should be implemented with one or more <tt>.cpp</tt>
463files. Each of these <tt>.cpp</tt> files should include the header that defines
464their interface first. This ensure that all of the dependences of the module
465header have been properly added to the module header itself, and are not
466implicit. System headers should be included after user headers for a
467translation unit.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]468
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]469</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]470
471<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]472<div class="doc_subsubsection">
Misha Brukman91d64af2004-10-26 16:18:43 +0000[diff] [blame]473 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]474</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]475
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]476<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]477
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]478<p><tt>#include</tt> hurts compile time performance. Don't do it unless you
479have to, especially in header files.</p>
480
481<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]482to inherit from it. In these cases go ahead and <tt>#include</tt> that header
483file. Be aware however that there are many cases where you don't need to have
484the full definition of a class. If you are using a pointer or reference to a
485class, you don't need the header file. If you are simply returning a class
486instance from a prototyped function or method, you don't need it. In fact, for
487most cases, you simply don't need the definition of a class... and not
488<tt>#include</tt>'ing speeds up compilation.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]489
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]490<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]491<b>must</b> include all of the header files that you are using -- you can
492include them either directly
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]493or indirectly (through another header file). To make sure that you don't
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]494accidentally forget to include a header file in your module header, make sure to
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]495include your module header <b>first</b> in the implementation file (as mentioned
496above). This way there won't be any hidden dependencies that you'll find out
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]497about later...</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]498
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]499</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]500
501<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]502<div class="doc_subsubsection">
503 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
504</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]505
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]506<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]507
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]508<p>Many modules have a complex implementation that causes them to use more than
509one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
510internal communication interface (helper classes, extra functions, etc) in the
Misha Brukman3b53a262004-12-04 00:32:12 +0000[diff] [blame]511public module header file. Don't do this.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]512
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]513<p>If you really need to do something like this, put a private header file in
514the same directory as the source files, and include it locally. This ensures
515that your private interface remains private and undisturbed by outsiders.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]516
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]517<p>Note however, that it's okay to put extra implementation methods a public
518class itself... just make them private (or protected), and all is well.</p>
519
520</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]521
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]522<!-- _______________________________________________________________________ -->
523<div class="doc_subsubsection">
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]524 <a name="hl_earlyexit">Use Early Exits and 'continue' to Simplify Code</a>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]525</div>
526
527<div class="doc_text">
528
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]529<p>When reading code, keep in mind how much state and how many previous
530decisions have to be remembered by the reader to understand a block of code.
531Aim to reduce indentation where possible when it doesn't make it more difficult
532to understand the code. One great way to do this is by making use of early
533exits and the 'continue' keyword in long loops. As an example of using an early
534exit from a function, consider this "bad" code:</p>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]535
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]536<div class="doc_code">
537<pre>
538Value *DoSomething(Instruction *I) {
539 if (!isa&lt;TerminatorInst&gt;(I) &amp;&amp;
540 I-&gt;hasOneUse() &amp;&amp; SomeOtherThing(I)) {
541 ... some long code ....
542 }
543
544 return 0;
545}
546</pre>
547</div>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]548
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]549<p>This code has several problems if the body of the 'if' is large. When you're
550looking at the top of the function, it isn't immediately clear that this
551<em>only</em> does interesting things with non-terminator instructions, and only
552applies to things with the other predicates. Second, it is relatively difficult
553to describe (in comments) why these predicates are important because the if
554statement makes it difficult to lay out the comments. Third, when you're deep
555within the body of the code, it is indented an extra level. Finally, when
556reading the top of the function, it isn't clear what the result is if the
557predicate isn't true, you have to read to the end of the function to know that
558it returns null.</p>
Matthijs Kooijmaned58a972008-07-30 12:14:10 +0000[diff] [blame]559
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]560<p>It is much preferred to format the code like this:</p>
561
562<div class="doc_code">
563<pre>
564Value *DoSomething(Instruction *I) {
565 // Terminators never need 'something' done to them because, ...
566 if (isa&lt;TerminatorInst&gt;(I))
567 return 0;
568
569 // We conservatively avoid transforming instructions with multiple uses
570 // because goats like cheese.
571 if (!I-&gt;hasOneUse())
572 return 0;
573
574 // This is really just here for example.
575 if (!SomeOtherThing(I))
576 return 0;
577
578 ... some long code ....
579}
580</pre>
581</div>
582
583<p>This fixes these problems. A similar problem frequently happens in for
584loops. A silly example is something like this:</p>
585
586<div class="doc_code">
587<pre>
588 for (BasicBlock::iterator II = BB-&gt;begin(), E = BB-&gt;end(); II != E; ++II) {
589 if (BinaryOperator *BO = dyn_cast&lt;BinaryOperator&gt;(II)) {
590 Value *LHS = BO-&gt;getOperand(0);
591 Value *RHS = BO-&gt;getOperand(1);
592 if (LHS != RHS) {
593 ...
594 }
595 }
596 }
597</pre>
598</div>
599
600<p>When you have very very small loops, this sort of structure is fine, but if
601it exceeds more than 10-15 lines, it becomes difficult for people to read and
602understand at a glance.
603The problem with this sort of code is that it gets very nested very quickly,
604meaning that the reader of the code has to keep a lot of context in their brain
605to remember what is going immediately on in the loop, because they don't know
606if/when the if conditions will have elses etc. It is strongly preferred to
607structure the loop like this:</p>
608
609<div class="doc_code">
610<pre>
611 for (BasicBlock::iterator II = BB-&gt;begin(), E = BB-&gt;end(); II != E; ++II) {
612 BinaryOperator *BO = dyn_cast&lt;BinaryOperator&gt;(II);
613 if (!BO) continue;
614
615 Value *LHS = BO-&gt;getOperand(0);
616 Value *RHS = BO-&gt;getOperand(1);
617 if (LHS == RHS) continue;
618 }
619</pre>
620</div>
621
622<p>This has all the benefits of using early exits from functions: it reduces
623nesting of the loop, it makes it easier to describe why the conditions are true,
624and it makes it obvious to the reader that there is no "else" coming up that
625they have to push context into their brain for. If a loop is large, this can
626be a big understandability win.</p>
627
628</div>
629
Chris Lattner9eb7e0a2009-07-28 22:54:04 +0000[diff] [blame]630<!-- _______________________________________________________________________ -->
631<div class="doc_subsubsection">
632 <a name="hl_else_after_return">Don't use "else" after a return</a>
633</div>
634
635<div class="doc_text">
636
637<p>For similar reasons above (reduction of indentation and easier reading),
638 please do not use "else" or "else if" after something that interrupts
639 control flow like return, break, continue, goto, etc. For example, this is
640 "bad":</p>
641
642<div class="doc_code">
643<pre>
644 case 'J': {
645 if (Signed) {
646 Type = Context.getsigjmp_bufType();
647 if (Type.isNull()) {
648 Error = ASTContext::GE_Missing_sigjmp_buf;
649 return QualType();
650 } else {
651 break;
652 }
653 } else {
654 Type = Context.getjmp_bufType();
655 if (Type.isNull()) {
656 Error = ASTContext::GE_Missing_jmp_buf;
657 return QualType();
658 } else {
659 break;
660 }
661 }
662 }
663 }
664</pre>
665</div>
666
667<p>It is better to write this something like:</p>
668
669<div class="doc_code">
670<pre>
671 case 'J':
672 if (Signed) {
673 Type = Context.getsigjmp_bufType();
674 if (Type.isNull()) {
675 Error = ASTContext::GE_Missing_sigjmp_buf;
676 return QualType();
677 }
678 } else {
679 Type = Context.getjmp_bufType();
680 if (Type.isNull()) {
681 Error = ASTContext::GE_Missing_jmp_buf;
682 return QualType();
683 }
684 }
685 break;
686</pre>
687</div>
688
689<p>Or better yet (in this case), as:</p>
690
691<div class="doc_code">
692<pre>
693 case 'J':
694 if (Signed)
695 Type = Context.getsigjmp_bufType();
696 else
697 Type = Context.getjmp_bufType();
698
699 if (Type.isNull()) {
700 Error = Signed ? ASTContext::GE_Missing_sigjmp_buf :
701 ASTContext::GE_Missing_jmp_buf;
702 return QualType();
703 }
704 break;
705</pre>
706</div>
707
708<p>The idea is to reduce indentation and the amount of code you have to keep
709 track of when reading the code.</p>
710
711</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]712
713<!-- _______________________________________________________________________ -->
714<div class="doc_subsubsection">
715 <a name="hl_predicateloops">Turn Predicate Loops into Predicate Functions</a>
716</div>
717
718<div class="doc_text">
719
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]720<p>It is very common to write small loops that just compute a boolean
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]721 value. There are a number of ways that people commonly write these, but an
722 example of this sort of thing is:</p>
723
724<div class="doc_code">
725<pre>
726 <b>bool FoundFoo = false;</b>
727 for (unsigned i = 0, e = BarList.size(); i != e; ++i)
728 if (BarList[i]-&gt;isFoo()) {
729 <b>FoundFoo = true;</b>
730 break;
731 }
732
733 <b>if (FoundFoo) {</b>
734 ...
735 }
736</pre>
737</div>
738
739<p>This sort of code is awkward to write, and is almost always a bad sign.
740Instead of this sort of loop, we strongly prefer to use a predicate function
741(which may be <a href="#micro_anonns">static</a>) that uses
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]742<a href="#hl_earlyexit">early exits</a> to compute the predicate. We prefer
743the code to be structured like this:
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]744</p>
745
746
747<div class="doc_code">
748<pre>
749/// ListContainsFoo - Return true if the specified list has an element that is
750/// a foo.
751static bool ListContainsFoo(const std::vector&lt;Bar*&gt; &amp;List) {
752 for (unsigned i = 0, e = List.size(); i != e; ++i)
753 if (List[i]-&gt;isFoo())
754 return true;
755 return false;
756}
757...
758
759 <b>if (ListContainsFoo(BarList)) {</b>
760 ...
761 }
762</pre>
763</div>
764
765<p>There are many reasons for doing this: it reduces indentation and factors out
766code which can often be shared by other code that checks for the same predicate.
767More importantly, it <em>forces you to pick a name</em> for the function, and
768forces you to write a comment for it. In this silly example, this doesn't add
769much value. However, if the condition is complex, this can make it a lot easier
770for the reader to understand the code that queries for this predicate. Instead
Chris Lattner71d8f3b2009-07-22 16:30:39 +0000[diff] [blame]771of 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]772contains a foo, we can trust the function name and continue reading with better
773locality.</p>
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]774
775</div>
776
777
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]778<!-- ======================================================================= -->
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]779<div class="doc_subsection">
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]780 <a name="micro">The Low Level Issues</a>
781</div>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]782<!-- ======================================================================= -->
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]783
784
785<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]786<div class="doc_subsubsection">
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]787 <a name="ll_assert">Assert Liberally</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]788</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]789
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]790<div class="doc_text">
791
792<p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]793preconditions and assumptions, you never know when a bug (not necessarily even
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]794yours) might be caught early by an assertion, which reduces debugging time
795dramatically. The "<tt>&lt;cassert&gt;</tt>" header file is probably already
796included 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]797it.</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]798
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]799<p>To further assist with debugging, make sure to put some kind of error message
800in the assertion statement (which is printed if the assertion is tripped). This
Misha Brukman2d10d752003-07-28 21:57:18 +0000[diff] [blame]801helps the poor debugging make sense of why an assertion is being made and
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]802enforced, and hopefully what to do about it. Here is one complete example:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]803
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]804<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]805<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]806inline Value *getOperand(unsigned i) {
807 assert(i &lt; Operands.size() &amp;&amp; "getOperand() out of range!");
808 return Operands[i];
809}
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]810</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]811</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]812
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]813<p>Here are some examples:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]814
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]815<div class="doc_code">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]816<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]817assert(Ty-&gt;isPointerType() &amp;&amp; "Can't allocate a non pointer type!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]818
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]819assert((Opcode == Shl || Opcode == Shr) &amp;&amp; "ShiftInst Opcode invalid!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]820
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]821assert(idx &lt; getNumSuccessors() &amp;&amp; "Successor # out of range!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]822
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]823assert(V1.getType() == V2.getType() &amp;&amp; "Constant types must be identical!");
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]824
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]825assert(isa&lt;PHINode&gt;(Succ-&gt;front()) &amp;&amp; "Only works on PHId BBs!");
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]826</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]827</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]828
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]829<p>You get the idea...</p>
830
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]831<p>Please be aware when adding assert statements that not all compilers are aware of
832the semantics of the assert. In some places, asserts are used to indicate a piece of
833code that should not be reached. These are typically of the form:</p>
834
835<div class="doc_code">
836<pre>
Dan Gohmanf0032762008-11-24 17:18:39 +0000[diff] [blame]837assert(0 &amp;&amp; "Some helpful error message");
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]838</pre>
839</div>
840
841<p>When used in a function that returns a value, they should be followed with a return
842statement and a comment indicating that this line is never reached. This will prevent
843a compiler which is unable to deduce that the assert statement never returns from
844generating a warning.</p>
845
846<div class="doc_code">
847<pre>
Dan Gohmanf0032762008-11-24 17:18:39 +0000[diff] [blame]848assert(0 &amp;&amp; "Some helpful error message");
Nick Lewyckyf4408e82008-05-31 23:54:55 +0000[diff] [blame]849// Not reached
850return 0;
851</pre>
852</div>
853
Chris Lattner97928d12010-11-12 00:19:41 +0000[diff] [blame]854<p>Another issue is that values used only by assertions will produce an "unused
855 value" warning when assertions are disabled. For example, this code will warn:
856</p>
857
858<div class="doc_code">
859<pre>
860 unsigned Size = V.size();
861 assert(Size &gt; 42 &amp;&amp; "Vector smaller than it should be");
862
863 bool NewToSet = Myset.insert(Value);
864 assert(NewToSet &amp;&amp; "The value shouldn't be in the set yet");
865</pre>
866</div>
867
868<p>These are two interesting different cases: in the first case, the call to
869V.size() is only useful for the assert, and we don't want it executed when
870assertions are disabled. Code like this should move the call into the assert
871itself. In the second case, the side effects of the call must happen whether
872the assert is enabled or not. In this case, the value should be cast to void
873to disable the warning. To be specific, it is preferred to write the code
874like this:</p>
875
876<div class="doc_code">
877<pre>
878 assert(V.size() &gt; 42 &amp;&amp; "Vector smaller than it should be");
879
880 bool NewToSet = Myset.insert(Value); (void)NewToSet;
881 assert(NewToSet &amp;&amp; "The value shouldn't be in the set yet");
882</pre>
883</div>
884
885
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]886</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]887
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]888<!-- _______________________________________________________________________ -->
889<div class="doc_subsubsection">
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]890 <a name="ll_ns_std">Do not use '<tt>using namespace std</tt>'</a>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]891</div>
892
893<div class="doc_text">
894<p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]895namespace with an "<tt>std::</tt>" prefix, rather than rely on
896"<tt>using namespace std;</tt>".</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]897
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]898<p> In header files, adding a '<tt>using namespace XXX</tt>' directive pollutes
Misha Brukman2d33c542008-12-11 19:37:04 +0000[diff] [blame]899the namespace of any source file that <tt>#include</tt>s the header. This is
900clearly a bad thing.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]901
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]902<p>In implementation files (e.g. .cpp files), the rule is more of a stylistic
903rule, but is still important. Basically, using explicit namespace prefixes
904makes the code <b>clearer</b>, because it is immediately obvious what facilities
905are being used and where they are coming from, and <b>more portable</b>, because
906namespace clashes cannot occur between LLVM code and other namespaces. The
907portability rule is important because different standard library implementations
908expose different symbols (potentially ones they shouldn't), and future revisions
909to the C++ standard will add more symbols to the <tt>std</tt> namespace. As
910such, we never use '<tt>using namespace std;</tt>' in LLVM.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]911
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]912<p>The exception to the general rule (i.e. it's not an exception for
913the <tt>std</tt> namespace) is for implementation files. For example, all of
914the code in the LLVM project implements code that lives in the 'llvm' namespace.
915As such, it is ok, and actually clearer, for the .cpp files to have a '<tt>using
916namespace llvm</tt>' directive at their top, after the <tt>#include</tt>s. The
917general form of this rule is that any .cpp file that implements code in any
918namespace may use that namespace (and its parents'), but should not use any
919others.</p>
Chris Lattnerb0bff9e2006-01-01 21:59:22 +0000[diff] [blame]920
921</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]922
923<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]924<div class="doc_subsubsection">
Bill Wendling4ad6d612006-12-09 01:35:43 +0000[diff] [blame]925 <a name="ll_virtual_anch">Provide a virtual method anchor for classes
926 in headers</a>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]927</div>
928
929<div class="doc_text">
930
931<p>If a class is defined in a header file and has a v-table (either it has
932virtual methods or it derives from classes with virtual methods), it must
933always have at least one out-of-line virtual method in the class. Without
Misha Brukman2d33c542008-12-11 19:37:04 +0000[diff] [blame]934this, the compiler will copy the vtable and RTTI into every <tt>.o</tt> file
935that <tt>#include</tt>s the header, bloating <tt>.o</tt> file sizes and
936increasing link times.</p>
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]937
938</div>
939
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]940<!-- _______________________________________________________________________ -->
941<div class="doc_subsubsection">
942 <a name="ll_end">Don't evaluate end() every time through a loop</a>
943</div>
944
945<div class="doc_text">
946
947<p>Because C++ doesn't have a standard "foreach" loop (though it can be emulated
948with macros and may be coming in C++'0x) we end up writing a lot of loops that
949manually iterate from begin to end on a variety of containers or through other
950data structures. One common mistake is to write a loop in this style:</p>
951
952<div class="doc_code">
953<pre>
954 BasicBlock *BB = ...
955 for (BasicBlock::iterator I = BB->begin(); I != <b>BB->end()</b>; ++I)
956 ... use I ...
957</pre>
958</div>
959
960<p>The problem with this construct is that it evaluates "<tt>BB->end()</tt>"
961every time through the loop. Instead of writing the loop like this, we strongly
962prefer loops to be written so that they evaluate it once before the loop starts.
963A convenient way to do this is like so:</p>
964
965<div class="doc_code">
966<pre>
967 BasicBlock *BB = ...
968 for (BasicBlock::iterator I = BB->begin(), E = <b>BB->end()</b>; I != E; ++I)
969 ... use I ...
970</pre>
971</div>
972
973<p>The observant may quickly point out that these two loops may have different
974semantics: if the container (a basic block in this case) is being mutated, then
975"<tt>BB->end()</tt>" may change its value every time through the loop and the
976second loop may not in fact be correct. If you actually do depend on this
Chris Lattner09bc65e2009-06-30 06:20:03 +0000[diff] [blame]977behavior, please write the loop in the first form and add a comment indicating
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]978that you did it intentionally.</p>
979
980<p>Why do we prefer the second form (when correct)? Writing the loop in the
981first form has two problems: First it may be less efficient than evaluating it
982at the start of the loop. In this case, the cost is probably minor: a few extra
983loads every time through the loop. However, if the base expression is more
984complex, then the cost can rise quickly. I've seen loops where the end
985expression was actually something like: "<tt>SomeMap[x]->end()</tt>" and map
Chris Lattnera44f87f2009-06-30 06:27:54 +0000[diff] [blame]986lookups really aren't cheap. By writing it in the second form consistently, you
Chris Lattner0af39ea2009-06-30 06:13:23 +0000[diff] [blame]987eliminate the issue entirely and don't even have to think about it.</p>
988
Chris Lattner09bc65e2009-06-30 06:20:03 +0000[diff] [blame]989<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]990hints to the reader that the loop is mutating the container (a fact that a
991comment would handily confirm!). If you write the loop in the second form, it
992is immediately obvious without even looking at the body of the loop that the
993container isn't being modified, which makes it easier to read the code and
994understand what it does.</p>
995
996<p>While the second form of the loop is a few extra keystrokes, we do strongly
997prefer it.</p>
998
999</div>
1000
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1001<!-- _______________________________________________________________________ -->
1002<div class="doc_subsubsection">
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1003 <a name="ll_iostream"><tt>#include &lt;iostream&gt;</tt> is forbidden</a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1004</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1005
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1006<div class="doc_text">
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1007
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1008<p>The use of <tt>#include &lt;iostream&gt;</tt> in library files is
1009hereby <b><em>forbidden</em></b>. The primary reason for doing this is to
1010support clients using LLVM libraries as part of larger systems. In particular,
1011we statically link LLVM into some dynamic libraries. Even if LLVM isn't used,
1012the static c'tors are run whenever an application start up that uses the dynamic
1013library. There are two problems with this:</p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1014
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1015<ol>
1016 <li>The time to run the static c'tors impacts startup time of
1017 applications&mdash;a critical time for GUI apps.</li>
1018 <li>The static c'tors cause the app to pull many extra pages of memory off the
1019 disk: both the code for the static c'tors in each <tt>.o</tt> file and the
1020 small amount of data that gets touched. In addition, touched/dirty pages
1021 put more pressure on the VM system on low-memory machines.</li>
1022</ol>
1023
1024<p>Note that using the other stream headers (<tt>&lt;sstream&gt;</tt> for
Chris Lattner983c5922009-08-23 21:53:47 +0000[diff] [blame]1025example) is not problematic in this regard (just <tt>&lt;iostream&gt;</tt>).
1026However, raw_ostream provides various APIs that are better performing for almost
1027every use than std::ostream style APIs, so you should just use it for new
1028code.</p>
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1029
Chris Lattner983c5922009-08-23 21:53:47 +0000[diff] [blame]1030<p><b>New code should always
1031use <a href="#ll_raw_ostream"><tt>raw_ostream</tt></a> for writing, or
1032the <tt>llvm::MemoryBuffer</tt> API for reading files.</b></p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1033
1034</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1035
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1036
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1037<!-- _______________________________________________________________________ -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1038<div class="doc_subsubsection">
Chris Lattner6883a882006-07-27 04:24:14 +0000[diff] [blame]1039 <a name="ll_avoidendl">Avoid <tt>std::endl</tt></a>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1040</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1041
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1042<div class="doc_text">
1043
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]1044<p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
1045to the output stream specified. In addition to doing this, however, it also
1046flushes the output stream. In other words, these are equivalent:</p>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1047
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1048<div class="doc_code">
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1049<pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1050std::cout &lt;&lt; std::endl;
1051std::cout &lt;&lt; '\n' &lt;&lt; std::flush;
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1052</pre>
Misha Brukmana8c760a2004-07-28 22:31:54 +0000[diff] [blame]1053</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1054
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1055<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]1056it's better to use a literal <tt>'\n'</tt>.</p>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1057
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1058</div>
Chris Lattner13c46592002-01-20 19:01:26 +0000[diff] [blame]1059
Bill Wendling7f564c02006-12-09 01:20:34 +0000[diff] [blame]1060
Daniel Dunbare8530a32009-07-24 23:04:51 +0000[diff] [blame]1061<!-- _______________________________________________________________________ -->
1062<div class="doc_subsubsection">
1063 <a name="ll_raw_ostream">Use <tt>raw_ostream</tt></a>
1064</div>
1065
1066<div class="doc_text">
1067
1068<p>LLVM includes a lightweight, simple, and efficient stream implementation
1069in <tt>llvm/Support/raw_ostream.h</tt> which provides all of the common features
Daniel Dunbar26fe8662009-07-24 23:54:34 +0000[diff] [blame]1070of <tt>std::ostream</tt>. All new code should use <tt>raw_ostream</tt> instead
1071of <tt>ostream</tt>.</p>
Daniel Dunbare8530a32009-07-24 23:04:51 +0000[diff] [blame]1072
Daniel Dunbar26fe8662009-07-24 23:54:34 +0000[diff] [blame]1073<p>Unlike <tt>std::ostream</tt>, <tt>raw_ostream</tt> is not a template and can
Daniel Dunbare8530a32009-07-24 23:04:51 +0000[diff] [blame]1074be forward declared as <tt>class raw_ostream</tt>. Public headers should
1075generally not include the <tt>raw_ostream</tt> header, but use forward
1076declarations and constant references to <tt>raw_ostream</tt> instances.</p>
1077
1078</div>
1079
1080
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1081<!-- ======================================================================= -->
1082<div class="doc_subsection">
1083 <a name="nano">Microscopic Details</a>
1084</div>
1085<!-- ======================================================================= -->
1086
1087<p>This section describes preferred low-level formatting guidelines along with
1088reasoning on why we prefer them.</p>
1089
1090<!-- _______________________________________________________________________ -->
1091<div class="doc_subsubsection">
1092 <a name="micro_spaceparen">Spaces Before Parentheses</a>
1093</div>
1094
1095<div class="doc_text">
1096
1097<p>We prefer to put a space before a parentheses only in control flow
1098statements, but not in normal function call expressions and function-like
1099macros. For example, this is good:</p>
1100
1101<div class="doc_code">
1102<pre>
1103 <b>if (</b>x) ...
1104 <b>for (</b>i = 0; i != 100; ++i) ...
1105 <b>while (</b>llvm_rocks) ...
1106
1107 <b>somefunc(</b>42);
1108 <b><a href="#ll_assert">assert</a>(</b>3 != 4 &amp;&amp; "laws of math are failing me");
1109
1110 a = <b>foo(</b>42, 92) + <b>bar(</b>x);
1111 </pre>
1112</div>
1113
1114<p>... and this is bad:</p>
1115
1116<div class="doc_code">
1117<pre>
1118 <b>if(</b>x) ...
1119 <b>for(</b>i = 0; i != 100; ++i) ...
1120 <b>while(</b>llvm_rocks) ...
1121
1122 <b>somefunc (</b>42);
1123 <b><a href="#ll_assert">assert</a> (</b>3 != 4 &amp;&amp; "laws of math are failing me");
1124
1125 a = <b>foo (</b>42, 92) + <b>bar (</b>x);
1126</pre>
1127</div>
1128
1129<p>The reason for doing this is not completely arbitrary. This style makes
1130 control flow operators stand out more, and makes expressions flow better. The
1131 function call operator binds very tightly as a postfix operator. Putting
1132 a space after a function name (as in the last example) makes it appear that
1133 the code might bind the arguments of the left-hand-side of a binary operator
1134 with the argument list of a function and the name of the right side. More
1135 specifically, it is easy to misread the "a" example as:</p>
1136
1137<div class="doc_code">
1138<pre>
1139 a = foo <b>(</b>(42, 92) + bar<b>)</b> (x);
1140</pre>
1141</div>
1142
1143<p>... when skimming through the code. By avoiding a space in a function, we
1144avoid this misinterpretation.</p>
1145
1146</div>
1147
1148<!-- _______________________________________________________________________ -->
1149<div class="doc_subsubsection">
1150 <a name="micro_preincrement">Prefer Preincrement</a>
1151</div>
1152
1153<div class="doc_text">
1154
1155<p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
1156postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
1157preincrementation whenever possible.</p>
1158
1159<p>The semantics of postincrement include making a copy of the value being
1160incremented, returning it, and then preincrementing the "work value". For
1161primitive types, this isn't a big deal... but for iterators, it can be a huge
1162issue (for example, some iterators contains stack and set objects in them...
1163copying an iterator could invoke the copy ctor's of these as well). In general,
1164get in the habit of always using preincrement, and you won't have a problem.</p>
1165
1166</div>
1167
1168<!-- _______________________________________________________________________ -->
1169<div class="doc_subsubsection">
1170 <a name="micro_namespaceindent">Namespace Indentation</a>
1171</div>
1172
1173<div class="doc_text">
1174
1175<p>
1176In general, we strive to reduce indentation where ever possible. This is useful
1177because we want code to <a href="#scf_codewidth">fit into 80 columns</a> without
1178wrapping horribly, but also because it makes it easier to understand the code.
1179Namespaces are a funny thing: they are often large, and we often desire to put
1180lots of stuff into them (so they can be large). Other times they are tiny,
1181because they just hold an enum or something similar. In order to balance this,
1182we use different approaches for small versus large namespaces.
1183</p>
1184
1185<p>
1186If a namespace definition is small and <em>easily</em> fits on a screen (say,
1187less than 35 lines of code), then you should indent its body. Here's an
1188example:
1189</p>
1190
1191<div class="doc_code">
1192<pre>
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1193namespace llvm {
1194 namespace X86 {
1195 /// RelocationType - An enum for the x86 relocation codes. Note that
1196 /// the terminology here doesn't follow x86 convention - word means
1197 /// 32-bit and dword means 64-bit.
1198 enum RelocationType {
1199 /// reloc_pcrel_word - PC relative relocation, add the relocated value to
1200 /// the value already in memory, after we adjust it for where the PC is.
1201 reloc_pcrel_word = 0,
1202
1203 /// reloc_picrel_word - PIC base relative relocation, add the relocated
1204 /// value to the value already in memory, after we adjust it for where the
1205 /// PIC base is.
1206 reloc_picrel_word = 1,
1207
1208 /// reloc_absolute_word, reloc_absolute_dword - Absolute relocation, just
1209 /// add the relocated value to the value already in memory.
1210 reloc_absolute_word = 2,
1211 reloc_absolute_dword = 3
1212 };
1213 }
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1214}
1215</pre>
1216</div>
1217
1218<p>Since the body is small, indenting adds value because it makes it very clear
1219where the namespace starts and ends, and it is easy to take the whole thing in
1220in one "gulp" when reading the code. If the blob of code in the namespace is
1221larger (as it typically is in a header in the llvm or clang namespaces), do not
1222indent the code, and add a comment indicating what namespace is being closed.
1223For example:</p>
1224
1225<div class="doc_code">
1226<pre>
1227namespace llvm {
1228namespace knowledge {
1229
1230/// Grokable - This class represents things that Smith can have an intimate
1231/// understanding of and contains the data associated with it.
1232class Grokable {
1233...
1234public:
1235 explicit Grokable() { ... }
1236 virtual ~Grokable() = 0;
1237
1238 ...
1239
1240};
1241
1242} // end namespace knowledge
1243} // end namespace llvm
1244</pre>
1245</div>
1246
1247<p>Because the class is large, we don't expect that the reader can easily
1248understand the entire concept in a glance, and the end of the file (where the
1249namespaces end) may be a long ways away from the place they open. As such,
1250indenting the contents of the namespace doesn't add any value, and detracts from
1251the readability of the class. In these cases it is best to <em>not</em> indent
1252the contents of the namespace.</p>
1253
1254</div>
1255
1256<!-- _______________________________________________________________________ -->
1257<div class="doc_subsubsection">
1258 <a name="micro_anonns">Anonymous Namespaces</a>
1259</div>
1260
1261<div class="doc_text">
1262
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1263<p>After talking about namespaces in general, you may be wondering about
1264anonymous namespaces in particular.
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1265Anonymous namespaces are a great language feature that tells the C++ compiler
1266that the contents of the namespace are only visible within the current
1267translation unit, allowing more aggressive optimization and eliminating the
1268possibility of symbol name collisions. Anonymous namespaces are to C++ as
1269"static" is to C functions and global variables. While "static" is available
1270in C++, anonymous namespaces are more general: they can make entire classes
1271private to a file.</p>
1272
1273<p>The problem with anonymous namespaces is that they naturally want to
1274encourage indentation of their body, and they reduce locality of reference: if
1275you see a random function definition in a C++ file, it is easy to see if it is
1276marked static, but seeing if it is in an anonymous namespace requires scanning
1277a big chunk of the file.</p>
1278
1279<p>Because of this, we have a simple guideline: make anonymous namespaces as
1280small as possible, and only use them for class declarations. For example, this
1281is good:</p>
1282
1283<div class="doc_code">
1284<pre>
1285<b>namespace {</b>
1286 class StringSort {
1287 ...
1288 public:
1289 StringSort(...)
1290 bool operator&lt;(const char *RHS) const;
1291 };
1292<b>} // end anonymous namespace</b>
1293
1294static void Helper() {
1295 ...
1296}
1297
1298bool StringSort::operator&lt;(const char *RHS) const {
1299 ...
1300}
1301
1302</pre>
1303</div>
1304
1305<p>This is bad:</p>
1306
1307
1308<div class="doc_code">
1309<pre>
1310<b>namespace {</b>
1311class StringSort {
1312...
1313public:
1314 StringSort(...)
1315 bool operator&lt;(const char *RHS) const;
1316};
1317
1318void Helper() {
1319 ...
1320}
1321
1322bool StringSort::operator&lt;(const char *RHS) const {
1323 ...
1324}
1325
1326<b>} // end anonymous namespace</b>
1327
1328</pre>
1329</div>
1330
1331
1332<p>This is bad specifically because if you're looking at "Helper" in the middle
1333of a large C++ file, that you have no immediate way to tell if it is local to
1334the file. When it is marked static explicitly, this is immediately obvious.
1335Also, there is no reason to enclose the definition of "operator&lt;" in the
Chris Lattner59fec6a2009-07-22 16:54:14 +0000[diff] [blame]1336namespace just because it was declared there.
Chris Lattnerd283cb12009-07-22 05:40:54 +0000[diff] [blame]1337</p>
1338
1339</div>
1340
1341
1342
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1343<!-- *********************************************************************** -->
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1344<div class="doc_section">
1345 <a name="seealso">See Also</a>
1346</div>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1347<!-- *********************************************************************** -->
1348
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1349<div class="doc_text">
1350
1351<p>A lot of these comments and recommendations have been culled for other
1352sources. Two particularly important books for our work are:</p>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1353
1354<ol>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1355
Chris Lattner788ee282007-11-09 21:49:08 +0000[diff] [blame]1356<li><a href="http://www.amazon.com/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0321334876">Effective
1357C++</a> by Scott Meyers. Also
Chris Lattner999cf092004-05-23 21:05:07 +0000[diff] [blame]1358interesting and useful are "More Effective C++" and "Effective STL" by the same
1359author.</li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1360
Chris Lattner788ee282007-11-09 21:49:08 +0000[diff] [blame]1361<li>Large-Scale C++ Software Design by John Lakos</li>
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1362
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1363</ol>
1364
1365<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]1366something.</p>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1367
1368</div>
1369
Chris Lattnerd80c43c2001-07-09 03:27:08 +0000[diff] [blame]1370<!-- *********************************************************************** -->
1371
1372<hr>
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1373<address>
1374 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1375 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1376 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]1377 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]1378
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1379 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]1380 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1381 Last modified: $Date$
Misha Brukman6c16c4b2004-05-12 18:37:22 +0000[diff] [blame]1382</address>
Misha Brukman75f258e2003-10-24 17:57:33 +0000[diff] [blame]1383
1384</body>
1385</html>