Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / dc43b3424bb2b1b8b8b9cebf1986231ce4e53cd2 / . / docs / CodingStandards.html
blob: ff29b39d6b775a7a5216fcc30745b9a322dfffc8 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <link rel="stylesheet" href="llvm.css" type="text/css">
Misha Brukman4dbc9d52008-12-10 23:07:02 +0000[diff] [blame]6 <title>LLVM Coding Standards</title>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]7</head>
8<body>
9
10<div class="doc_title">
Misha Brukman4dbc9d52008-12-10 23:07:02 +0000[diff] [blame]11 LLVM Coding Standards
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]12</div>
13
14<ol>
15 <li><a href="#introduction">Introduction</a></li>
16 <li><a href="#mechanicalissues">Mechanical Source Issues</a>
17 <ol>
18 <li><a href="#sourceformating">Source Code Formatting</a>
19 <ol>
20 <li><a href="#scf_commenting">Commenting</a></li>
21 <li><a href="#scf_commentformat">Comment Formatting</a></li>
22 <li><a href="#scf_includes"><tt>#include</tt> Style</a></li>
23 <li><a href="#scf_codewidth">Source Code Width</a></li>
24 <li><a href="#scf_spacestabs">Use Spaces Instead of Tabs</a></li>
25 <li><a href="#scf_indentation">Indent Code Consistently</a></li>
26 </ol></li>
27 <li><a href="#compilerissues">Compiler Issues</a>
28 <ol>
29 <li><a href="#ci_warningerrors">Treat Compiler Warnings Like
30 Errors</a></li>
31 <li><a href="#ci_portable_code">Write Portable Code</a></li>
32 <li><a href="#ci_class_struct">Use of class/struct Keywords</a></li>
33 </ol></li>
34 </ol></li>
35 <li><a href="#styleissues">Style Issues</a>
36 <ol>
37 <li><a href="#macro">The High Level Issues</a>
38 <ol>
39 <li><a href="#hl_module">A Public Header File <b>is</b> a
40 Module</a></li>
41 <li><a href="#hl_dontinclude">#include as Little as Possible</a></li>
42 <li><a href="#hl_privateheaders">Keep "internal" Headers
43 Private</a></li>
44 <li><a href="#ll_iostream"><tt>#include &lt;iostream&gt;</tt> is
45 <em>forbidden</em></a></li>
46 </ol></li>
47 <li><a href="#micro">The Low Level Issues</a>
48 <ol>
49 <li><a href="#ll_assert">Assert Liberally</a></li>
50 <li><a href="#ll_ns_std">Do not use 'using namespace std'</a></li>
51 <li><a href="#ll_virtual_anch">Provide a virtual method anchor for
52 classes in headers</a></li>
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]53 <li><a href="#ll_end">Don't evaluate end() every time through a
54 loop</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]55 <li><a href="#ll_preincrement">Prefer Preincrement</a></li>
56 <li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
57 </ol></li>
58 </ol></li>
59 <li><a href="#seealso">See Also</a></li>
60</ol>
61
62<div class="doc_author">
63 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a> and
64 <a href="mailto:void@nondot.org">Bill Wendling</a></p>
65</div>
66
67
68<!-- *********************************************************************** -->
69<div class="doc_section">
70 <a name="introduction">Introduction</a>
71</div>
72<!-- *********************************************************************** -->
73
74<div class="doc_text">
75
76<p>This document attempts to describe a few coding standards that are being used
77in the LLVM source tree. Although no coding standards should be regarded as
78absolute requirements to be followed in all instances, coding standards can be
79useful.</p>
80
81<p>This document intentionally does not prescribe fixed standards for religious
82issues such as brace placement and space usage. For issues like this, follow
83the golden rule:</p>
84
85<blockquote>
86
87<p><b><a name="goldenrule">If you are adding a significant body of source to a
88project, feel free to use whatever style you are most comfortable with. If you
89are extending, enhancing, or bug fixing already implemented code, use the style
90that is already being used so that the source is uniform and easy to
91follow.</a></b></p>
92
93</blockquote>
94
95<p>The ultimate goal of these guidelines is the increase readability and
96maintainability of our common source base. If you have suggestions for topics to
97be included, please mail them to <a
98href="mailto:sabre@nondot.org">Chris</a>.</p>
99
100</div>
101
102<!-- *********************************************************************** -->
103<div class="doc_section">
104 <a name="mechanicalissues">Mechanical Source Issues</a>
105</div>
106<!-- *********************************************************************** -->
107
108<!-- ======================================================================= -->
109<div class="doc_subsection">
110 <a name="sourceformating">Source Code Formatting</a>
111</div>
112
113<!-- _______________________________________________________________________ -->
114<div class="doc_subsubsection">
115 <a name="scf_commenting">Commenting</a>
116</div>
117
118<div class="doc_text">
119
120<p>Comments are one critical part of readability and maintainability. Everyone
121knows they should comment, so should you. Although we all should probably
122comment our code more than we do, there are a few very critical places that
123documentation is very useful:</p>
124
125<b>File Headers</b>
126
127<p>Every source file should have a header on it that describes the basic
128purpose of the file. If a file does not have a header, it should not be
129checked into Subversion. Most source trees will probably have a standard
130file header format. The standard format for the LLVM source tree looks like
131this:</p>
132
133<div class="doc_code">
134<pre>
135//===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
Misha Brukman0fd5f3e2009-01-02 16:58:42 +0000[diff] [blame]136//
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]137// The LLVM Compiler Infrastructure
138//
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]139// This file is distributed under the University of Illinois Open Source
140// License. See LICENSE.TXT for details.
Misha Brukman0fd5f3e2009-01-02 16:58:42 +0000[diff] [blame]141//
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]142//===----------------------------------------------------------------------===//
143//
144// This file contains the declaration of the Instruction class, which is the
145// base class for all of the VM instructions.
146//
147//===----------------------------------------------------------------------===//
148</pre>
149</div>
150
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]151<p>A few things to note about this particular format: The "<tt>-*- C++
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]152-*-</tt>" string on the first line is there to tell Emacs that the source file
153is a C++ file, not a C file (Emacs assumes .h files are C files by default).
154Note that this tag is not necessary in .cpp files. The name of the file is also
155on the first line, along with a very short description of the purpose of the
156file. This is important when printing out code and flipping though lots of
157pages.</p>
158
Chris Lattnerfb5959f2007-12-29 19:56:08 +0000[diff] [blame]159<p>The next section in the file is a concise note that defines the license
160that the file is released under. This makes it perfectly clear what terms the
161source code can be distributed under and should not be modified in any way.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]162
163<p>The main body of the description does not have to be very long in most cases.
164Here it's only two lines. If an algorithm is being implemented or something
165tricky is going on, a reference to the paper where it is published should be
166included, as well as any notes or "gotchas" in the code to watch out for.</p>
167
168<b>Class overviews</b>
169
170<p>Classes are one fundamental part of a good object oriented design. As such,
171a class definition should have a comment block that explains what the class is
172used for... if it's not obvious. If it's so completely obvious your grandma
173could figure it out, it's probably safe to leave it out. Naming classes
174something sane goes a long ways towards avoiding writing documentation.</p>
175
176
177<b>Method information</b>
178
179<p>Methods defined in a class (as well as any global functions) should also be
180documented properly. A quick note about what it does any a description of the
181borderline behaviour is all that is necessary here (unless something
182particularly tricky or insideous is going on). The hope is that people can
183figure out how to use your interfaces without reading the code itself... that is
184the goal metric.</p>
185
186<p>Good things to talk about here are what happens when something unexpected
187happens: does the method return null? Abort? Format your hard disk?</p>
188
189</div>
190
191<!-- _______________________________________________________________________ -->
192<div class="doc_subsubsection">
193 <a name="scf_commentformat">Comment Formatting</a>
194</div>
195
196<div class="doc_text">
197
198<p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
199require less typing, don't have nesting problems, etc. There are a few cases
200when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
201
202<ol>
203 <li>When writing a C code: Obviously if you are writing C code, use C style
204 comments.</li>
205 <li>When writing a header file that may be <tt>#include</tt>d by a C source
206 file.</li>
207 <li>When writing a source file that is used by a tool that only accepts C
208 style comments.</li>
209</ol>
210
211<p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
212These nest properly and are better behaved in general than C style comments.</p>
213
214</div>
215
216<!-- _______________________________________________________________________ -->
217<div class="doc_subsubsection">
218 <a name="scf_includes"><tt>#include</tt> Style</a>
219</div>
220
221<div class="doc_text">
222
223<p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
224include guards if working on a header file), the <a
225href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
226file should be listed. We prefer these <tt>#include</tt>s to be listed in this
227order:</p>
228
229<ol>
230 <li><a href="#mmheader">Main Module header</a></li>
231 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
232 <li><tt>llvm/*</tt></li>
233 <li><tt>llvm/Analysis/*</tt></li>
234 <li><tt>llvm/Assembly/*</tt></li>
235 <li><tt>llvm/Bytecode/*</tt></li>
236 <li><tt>llvm/CodeGen/*</tt></li>
237 <li>...</li>
238 <li><tt>Support/*</tt></li>
239 <li><tt>Config/*</tt></li>
240 <li>System <tt>#includes</tt></li>
241</ol>
242
Nick Lewycky61819f72008-11-29 20:13:25 +0000[diff] [blame]243<p>... and each category should be sorted by name.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]244
245<p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
246which implement an interface defined by a .h file. This <tt>#include</tt>
247should always be included <b>first</b> regardless of where it lives on the file
248system. By including a header file first in the .cpp files that implement the
249interfaces, we ensure that the header does not have any hidden dependencies
250which are not explicitly #included in the header, but should be. It is also a
251form of documentation in the .cpp file to indicate where the interfaces it
252implements are defined.</p>
253
254</div>
255
256<!-- _______________________________________________________________________ -->
257<div class="doc_subsubsection">
258 <a name="scf_codewidth">Source Code Width</a>
259</div>
260
261<div class="doc_text">
262
263<p>Write your code to fit within 80 columns of text. This helps those of us who
264like to print out code and look at your code in an xterm without resizing
265it.</p>
266
Chris Lattnerb5daf702008-07-08 05:12:37 +0000[diff] [blame]267<p>The longer answer is that there must be some limit to the width of the code
268in order to reasonably allow developers to have multiple files side-by-side in
269windows on a modest display. If you are going to pick a width limit, it is
270somewhat arbitrary but you might as well pick something standard. Going with
27190 columns (for example) instead of 80 columns wouldn't add any significant
272value and would be detrimental to printing out code. Also many other projects
273have standardized on 80 columns, so some people have already configured their
274editors for it (vs something else, like 90 columns).</p>
275
276<p>This is one of many contentious issues in coding standards, but is not up
277for debate.</p>
278
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]279</div>
280
281<!-- _______________________________________________________________________ -->
282<div class="doc_subsubsection">
283 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
284</div>
285
286<div class="doc_text">
287
288<p>In all cases, prefer spaces to tabs in source files. People have different
289prefered indentation levels, and different styles of indentation that they
290like... this is fine. What isn't is that different editors/viewers expand tabs
291out to different tab stops. This can cause your code to look completely
292unreadable, and it is not worth dealing with.</p>
293
294<p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
295style of existing code if your are modifying and extending it. If you like four
296spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
297with two spaces of indentation. Also, do not reindent a whole source file: it
298makes for incredible diffs that are absolutely worthless.</p>
299
300</div>
301
302<!-- _______________________________________________________________________ -->
303<div class="doc_subsubsection">
304 <a name="scf_indentation">Indent Code Consistently</a>
305</div>
306
307<div class="doc_text">
308
309<p>Okay, your first year of programming you were told that indentation is
310important. If you didn't believe and internalize this then, now is the time.
311Just do it.</p>
312
313</div>
314
315
316<!-- ======================================================================= -->
317<div class="doc_subsection">
318 <a name="compilerissues">Compiler Issues</a>
319</div>
320
321
322<!-- _______________________________________________________________________ -->
323<div class="doc_subsubsection">
324 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
325</div>
326
327<div class="doc_text">
328
329<p>If your code has compiler warnings in it, something is wrong: you aren't
330casting values correctly, your have "questionable" constructs in your code, or
331you are doing something legitimately wrong. Compiler warnings can cover up
332legitimate errors in output and make dealing with a translation unit
333difficult.</p>
334
335<p>It is not possible to prevent all warnings from all compilers, nor is it
336desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
337a good thorough set of warnings, and stick to them. At least in the case of
338<tt>gcc</tt>, it is possible to work around any spurious errors by changing the
339syntax of the code slightly. For example, an warning that annoys me occurs when
340I write code like this:</p>
341
342<div class="doc_code">
343<pre>
344if (V = getValue()) {
345 ...
346}
347</pre>
348</div>
349
350<p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
351operator, and that I probably mistyped it. In most cases, I haven't, and I
352really don't want the spurious errors. To fix this particular problem, I
353rewrite the code like this:</p>
354
355<div class="doc_code">
356<pre>
357if ((V = getValue())) {
358 ...
359}
360</pre>
361</div>
362
363<p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
364be fixed by massaging the code appropriately.</p>
365
366<p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
367-Winline -W -Wwrite-strings -Wno-unused</tt></p>
368
369</div>
370
371<!-- _______________________________________________________________________ -->
372<div class="doc_subsubsection">
373 <a name="ci_portable_code">Write Portable Code</a>
374</div>
375
376<div class="doc_text">
377
378<p>In almost all cases, it is possible and within reason to write completely
379portable code. If there are cases where it isn't possible to write portable
380code, isolate it behind a well defined (and well documented) interface.</p>
381
382<p>In practice, this means that you shouldn't assume much about the host
383compiler, including its support for "high tech" features like partial
Chris Lattnerafec0252009-03-23 04:52:53 +0000[diff] [blame]384specialization of templates. If these features are used, they should only be
385an implementation detail of a library which has a simple exposed API.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]386
387</div>
388
389<!-- _______________________________________________________________________ -->
390<div class="doc_subsubsection">
391<a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
392</div>
393<div class="doc_text">
394
395<p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
396interchangeably. The only difference is when they are used to declare a class:
397<tt>class</tt> makes all members private by default while <tt>struct</tt> makes
398all members public by default.</p>
399
400<p>Unfortunately, not all compilers follow the rules and some will generate
401different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
402declare the symbol. This can lead to problems at link time.</p>
403
404<p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
405<b>all</b> members are public, in which case <tt>struct</tt> is allowed.</p>
406
407</div>
408
409<!-- *********************************************************************** -->
410<div class="doc_section">
411 <a name="styleissues">Style Issues</a>
412</div>
413<!-- *********************************************************************** -->
414
415
416<!-- ======================================================================= -->
417<div class="doc_subsection">
418 <a name="macro">The High Level Issues</a>
419</div>
420
421
422<!-- _______________________________________________________________________ -->
423<div class="doc_subsubsection">
424 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
425</div>
426
427<div class="doc_text">
428
429<p>C++ doesn't do too well in the modularity department. There is no real
430encapsulation or data hiding (unless you use expensive protocol classes), but it
431is what we have to work with. When you write a public header file (in the LLVM
432source tree, they live in the top level "include" directory), you are defining a
433module of functionality.</p>
434
435<p>Ideally, modules should be completely independent of each other, and their
436header files should only include the absolute minimum number of headers
437possible. A module is not just a class, a function, or a namespace: <a
438href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
439of these</a> that defines an interface. This interface may be several
440functions, classes or data structures, but the important issue is how they work
441together.</p>
442
443<p>In general, a module should be implemented with one or more <tt>.cpp</tt>
444files. Each of these <tt>.cpp</tt> files should include the header that defines
445their interface first. This ensure that all of the dependences of the module
446header have been properly added to the module header itself, and are not
447implicit. System headers should be included after user headers for a
448translation unit.</p>
449
450</div>
451
452<!-- _______________________________________________________________________ -->
453<div class="doc_subsubsection">
454 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
455</div>
456
457<div class="doc_text">
458
459<p><tt>#include</tt> hurts compile time performance. Don't do it unless you
460have to, especially in header files.</p>
461
462<p>But wait, sometimes you need to have the definition of a class to use it, or
463to inherit from it. In these cases go ahead and <tt>#include</tt> that header
464file. Be aware however that there are many cases where you don't need to have
465the full definition of a class. If you are using a pointer or reference to a
466class, you don't need the header file. If you are simply returning a class
467instance from a prototyped function or method, you don't need it. In fact, for
468most cases, you simply don't need the definition of a class... and not
469<tt>#include</tt>'ing speeds up compilation.</p>
470
471<p>It is easy to try to go too overboard on this recommendation, however. You
472<b>must</b> include all of the header files that you are using -- you can
473include them either directly
474or indirectly (through another header file). To make sure that you don't
475accidently forget to include a header file in your module header, make sure to
476include your module header <b>first</b> in the implementation file (as mentioned
477above). This way there won't be any hidden dependencies that you'll find out
478about later...</p>
479
480</div>
481
482<!-- _______________________________________________________________________ -->
483<div class="doc_subsubsection">
484 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
485</div>
486
487<div class="doc_text">
488
489<p>Many modules have a complex implementation that causes them to use more than
490one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
491internal communication interface (helper classes, extra functions, etc) in the
492public module header file. Don't do this.</p>
493
494<p>If you really need to do something like this, put a private header file in
495the same directory as the source files, and include it locally. This ensures
496that your private interface remains private and undisturbed by outsiders.</p>
497
498<p>Note however, that it's okay to put extra implementation methods a public
499class itself... just make them private (or protected), and all is well.</p>
500
501</div>
502
503<!-- _______________________________________________________________________ -->
504<div class="doc_subsubsection">
505 <a name="ll_iostream"><tt>#include &lt;iostream&gt;</tt> is forbidden</a>
506</div>
507
508<div class="doc_text">
509
510<p>The use of <tt>#include &lt;iostream&gt;</tt> in library files is
511hereby <b><em>forbidden</em></b>. The primary reason for doing this is to
512support clients using LLVM libraries as part of larger systems. In particular,
513we statically link LLVM into some dynamic libraries. Even if LLVM isn't used,
514the static c'tors are run whenever an application start up that uses the dynamic
515library. There are two problems with this:</p>
516
517<ol>
518 <li>The time to run the static c'tors impacts startup time of
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]519 applications&mdash;a critical time for GUI apps.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]520 <li>The static c'tors cause the app to pull many extra pages of memory off the
Bill Wendling71f8cf12007-11-06 09:36:34 +0000[diff] [blame]521 disk: both the code for the static c'tors in each <tt>.o</tt> file and the
522 small amount of data that gets touched. In addition, touched/dirty pages
523 put more pressure on the VM system on low-memory machines.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]524</ol>
525
Matthijs Kooijmanc4c0af22008-07-30 12:14:10 +0000[diff] [blame]526<p>Note that using the other stream headers (<tt>&lt;sstream&gt;</tt> for
527example) is allowed normally, it is just <tt>&lt;iostream&gt;</tt> that is
528causing problems.</p>
529
Chris Lattnerafec0252009-03-23 04:52:53 +0000[diff] [blame]530<p>The preferred replacement for stream functionality is the
Chris Lattner85187f12009-03-23 04:53:34 +0000[diff] [blame]531<tt>llvm::raw_ostream</tt> class (for writing to output streams of various
532sorts) and the <tt>llvm::MemoryBuffer</tt> API (for reading in files).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]533
534</div>
535
536
537<!-- ======================================================================= -->
538<div class="doc_subsection">
539 <a name="micro">The Low Level Issues</a>
540</div>
541
542
543<!-- _______________________________________________________________________ -->
544<div class="doc_subsubsection">
545 <a name="ll_assert">Assert Liberally</a>
546</div>
547
548<div class="doc_text">
549
550<p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
551preconditions and assumptions, you never know when a bug (not neccesarily even
552yours) might be caught early by an assertion, which reduces debugging time
553dramatically. The "<tt>&lt;cassert&gt;</tt>" header file is probably already
554included by the header files you are using, so it doesn't cost anything to use
555it.</p>
556
557<p>To further assist with debugging, make sure to put some kind of error message
558in the assertion statement (which is printed if the assertion is tripped). This
559helps the poor debugging make sense of why an assertion is being made and
560enforced, and hopefully what to do about it. Here is one complete example:</p>
561
562<div class="doc_code">
563<pre>
564inline Value *getOperand(unsigned i) {
565 assert(i &lt; Operands.size() &amp;&amp; "getOperand() out of range!");
566 return Operands[i];
567}
568</pre>
569</div>
570
571<p>Here are some examples:</p>
572
573<div class="doc_code">
574<pre>
575assert(Ty-&gt;isPointerType() &amp;&amp; "Can't allocate a non pointer type!");
576
577assert((Opcode == Shl || Opcode == Shr) &amp;&amp; "ShiftInst Opcode invalid!");
578
579assert(idx &lt; getNumSuccessors() &amp;&amp; "Successor # out of range!");
580
581assert(V1.getType() == V2.getType() &amp;&amp; "Constant types must be identical!");
582
583assert(isa&lt;PHINode&gt;(Succ-&gt;front()) &amp;&amp; "Only works on PHId BBs!");
584</pre>
585</div>
586
587<p>You get the idea...</p>
588
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]589<p>Please be aware when adding assert statements that not all compilers are aware of
590the semantics of the assert. In some places, asserts are used to indicate a piece of
591code that should not be reached. These are typically of the form:</p>
592
593<div class="doc_code">
594<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]595assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]596</pre>
597</div>
598
599<p>When used in a function that returns a value, they should be followed with a return
600statement and a comment indicating that this line is never reached. This will prevent
601a compiler which is unable to deduce that the assert statement never returns from
602generating a warning.</p>
603
604<div class="doc_code">
605<pre>
Dan Gohman4dfac702008-11-24 17:18:39 +0000[diff] [blame]606assert(0 &amp;&amp; "Some helpful error message");
Nick Lewycky009216a2008-05-31 23:54:55 +0000[diff] [blame]607// Not reached
608return 0;
609</pre>
610</div>
611
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]612</div>
613
614<!-- _______________________________________________________________________ -->
615<div class="doc_subsubsection">
616 <a name="ll_ns_std">Do not use '<tt>using namespace std</tt>'</a>
617</div>
618
619<div class="doc_text">
620<p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
621namespace with an "<tt>std::</tt>" prefix, rather than rely on
622"<tt>using namespace std;</tt>".</p>
623
624<p> In header files, adding a '<tt>using namespace XXX</tt>' directive pollutes
Misha Brukmanbcc53732008-12-11 19:37:04 +0000[diff] [blame]625the namespace of any source file that <tt>#include</tt>s the header. This is
626clearly a bad thing.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]627
628<p>In implementation files (e.g. .cpp files), the rule is more of a stylistic
629rule, but is still important. Basically, using explicit namespace prefixes
630makes the code <b>clearer</b>, because it is immediately obvious what facilities
631are being used and where they are coming from, and <b>more portable</b>, because
632namespace clashes cannot occur between LLVM code and other namespaces. The
633portability rule is important because different standard library implementations
634expose different symbols (potentially ones they shouldn't), and future revisions
635to the C++ standard will add more symbols to the <tt>std</tt> namespace. As
636such, we never use '<tt>using namespace std;</tt>' in LLVM.</p>
637
638<p>The exception to the general rule (i.e. it's not an exception for
639the <tt>std</tt> namespace) is for implementation files. For example, all of
640the code in the LLVM project implements code that lives in the 'llvm' namespace.
641As such, it is ok, and actually clearer, for the .cpp files to have a '<tt>using
642namespace llvm</tt>' directive at their top, after the <tt>#include</tt>s. The
643general form of this rule is that any .cpp file that implements code in any
644namespace may use that namespace (and its parents'), but should not use any
645others.</p>
646
647</div>
648
649<!-- _______________________________________________________________________ -->
650<div class="doc_subsubsection">
651 <a name="ll_virtual_anch">Provide a virtual method anchor for classes
652 in headers</a>
653</div>
654
655<div class="doc_text">
656
657<p>If a class is defined in a header file and has a v-table (either it has
658virtual methods or it derives from classes with virtual methods), it must
659always have at least one out-of-line virtual method in the class. Without
Misha Brukmanbcc53732008-12-11 19:37:04 +0000[diff] [blame]660this, the compiler will copy the vtable and RTTI into every <tt>.o</tt> file
661that <tt>#include</tt>s the header, bloating <tt>.o</tt> file sizes and
662increasing link times.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]663
664</div>
665
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]666<!-- _______________________________________________________________________ -->
667<div class="doc_subsubsection">
668 <a name="ll_end">Don't evaluate end() every time through a loop</a>
669</div>
670
671<div class="doc_text">
672
673<p>Because C++ doesn't have a standard "foreach" loop (though it can be emulated
674with macros and may be coming in C++'0x) we end up writing a lot of loops that
675manually iterate from begin to end on a variety of containers or through other
676data structures. One common mistake is to write a loop in this style:</p>
677
678<div class="doc_code">
679<pre>
680 BasicBlock *BB = ...
681 for (BasicBlock::iterator I = BB->begin(); I != <b>BB->end()</b>; ++I)
682 ... use I ...
683</pre>
684</div>
685
686<p>The problem with this construct is that it evaluates "<tt>BB->end()</tt>"
687every time through the loop. Instead of writing the loop like this, we strongly
688prefer loops to be written so that they evaluate it once before the loop starts.
689A convenient way to do this is like so:</p>
690
691<div class="doc_code">
692<pre>
693 BasicBlock *BB = ...
694 for (BasicBlock::iterator I = BB->begin(), E = <b>BB->end()</b>; I != E; ++I)
695 ... use I ...
696</pre>
697</div>
698
699<p>The observant may quickly point out that these two loops may have different
700semantics: if the container (a basic block in this case) is being mutated, then
701"<tt>BB->end()</tt>" may change its value every time through the loop and the
702second loop may not in fact be correct. If you actually do depend on this
Chris Lattnerdc43b342009-06-30 06:20:03 +0000[diff] [blame^]703behavior, please write the loop in the first form and add a comment indicating
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]704that you did it intentionally.</p>
705
706<p>Why do we prefer the second form (when correct)? Writing the loop in the
707first form has two problems: First it may be less efficient than evaluating it
708at the start of the loop. In this case, the cost is probably minor: a few extra
709loads every time through the loop. However, if the base expression is more
710complex, then the cost can rise quickly. I've seen loops where the end
711expression was actually something like: "<tt>SomeMap[x]->end()</tt>" and map
712lookups really aren't cheap. By writing it in the first form consistently, you
713eliminate the issue entirely and don't even have to think about it.</p>
714
Chris Lattnerdc43b342009-06-30 06:20:03 +0000[diff] [blame^]715<p>The second (even bigger) issue is that writing the loop in the first form
Chris Lattner01b8af32009-06-30 06:13:23 +0000[diff] [blame]716hints to the reader that the loop is mutating the container (a fact that a
717comment would handily confirm!). If you write the loop in the second form, it
718is immediately obvious without even looking at the body of the loop that the
719container isn't being modified, which makes it easier to read the code and
720understand what it does.</p>
721
722<p>While the second form of the loop is a few extra keystrokes, we do strongly
723prefer it.</p>
724
725</div>
726
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]727
728<!-- _______________________________________________________________________ -->
729<div class="doc_subsubsection">
730 <a name="ll_preincrement">Prefer Preincrement</a>
731</div>
732
733<div class="doc_text">
734
735<p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
736postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
737preincrementation whenever possible.</p>
738
739<p>The semantics of postincrement include making a copy of the value being
740incremented, returning it, and then preincrementing the "work value". For
741primitive types, this isn't a big deal... but for iterators, it can be a huge
742issue (for example, some iterators contains stack and set objects in them...
743copying an iterator could invoke the copy ctor's of these as well). In general,
744get in the habit of always using preincrement, and you won't have a problem.</p>
745
746</div>
747
748<!-- _______________________________________________________________________ -->
749<div class="doc_subsubsection">
750 <a name="ll_avoidendl">Avoid <tt>std::endl</tt></a>
751</div>
752
753<div class="doc_text">
754
755<p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
756to the output stream specified. In addition to doing this, however, it also
757flushes the output stream. In other words, these are equivalent:</p>
758
759<div class="doc_code">
760<pre>
761std::cout &lt;&lt; std::endl;
762std::cout &lt;&lt; '\n' &lt;&lt; std::flush;
763</pre>
764</div>
765
766<p>Most of the time, you probably have no reason to flush the output stream, so
767it's better to use a literal <tt>'\n'</tt>.</p>
768
769</div>
770
771
772<!-- *********************************************************************** -->
773<div class="doc_section">
774 <a name="seealso">See Also</a>
775</div>
776<!-- *********************************************************************** -->
777
778<div class="doc_text">
779
780<p>A lot of these comments and recommendations have been culled for other
781sources. Two particularly important books for our work are:</p>
782
783<ol>
784
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]785<li><a href="http://www.amazon.com/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0321334876">Effective
786C++</a> by Scott Meyers. Also
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]787interesting and useful are "More Effective C++" and "Effective STL" by the same
788author.</li>
789
Chris Lattnerb5955b72007-11-09 21:49:08 +0000[diff] [blame]790<li>Large-Scale C++ Software Design by John Lakos</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]791
792</ol>
793
794<p>If you get some free time, and you haven't read them: do so, you might learn
795something.</p>
796
797</div>
798
799<!-- *********************************************************************** -->
800
801<hr>
802<address>
803 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]804 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]805 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman947321d2008-12-11 17:34:48 +0000[diff] [blame]806 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]807
808 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
809 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
810 Last modified: $Date$
811</address>
812
813</body>
814</html>