1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <link rel="stylesheet" href="llvm.css" type="text/css">
6 <title>A Few Coding Standards</title>
10 <div class="doc_title">
11 A Few Coding Standards
15 <li><a href="#introduction">Introduction</a></li>
16 <li><a href="#mechanicalissues">Mechanical Source Issues</a>
18 <li><a href="#sourceformating">Source Code Formatting</a>
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">#include 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>
27 <li><a href="#compilerissues">Compiler Issues</a>
29 <li><a href="#ci_warningerrors">Treat Compiler Warnings Like
31 <li><a href="#ci_portable_code">Write Portable Code</a></li>
34 <li><a href="#styleissues">Style Issues</a>
36 <li><a href="#macro">The High Level Issues</a>
38 <li><a href="#hl_module">A Public Header File <b>is</b> a
40 <li><a href="#hl_dontinclude">#include as Little as Possible</a></li>
41 <li><a href="#hl_privateheaders">Keep "internal" Headers
44 <li><a href="#micro">The Low Level Issues</a>
46 <li><a href="#hl_assert">Assert Liberally</a></li>
47 <li><a href="#hl_preincrement">Prefer Preincrement</a></li>
48 <li><a href="#hl_avoidendl">Avoid std::endl</a></li>
49 <li><a href="#hl_exploitcpp">Exploit C++ to its Fullest</a></li>
52 <li><a href="#seealso">See Also</a></li>
55 <div class="doc_author">
56 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a></p>
60 <!-- *********************************************************************** -->
61 <div class="doc_section">
62 <a name="introduction">Introduction</a>
64 <!-- *********************************************************************** -->
66 <div class="doc_text">
68 <p>This document attempts to describe a few coding standards that are being used
69 in the LLVM source tree. Although no coding standards should be regarded as
70 absolute requirements to be followed in all instances, coding standards can be
73 <p>This document intentionally does not prescribe fixed standards for religious
74 issues such as brace placement and space usage. For issues like this, follow
79 <p><b><a name="goldenrule">If you are adding a significant body of source to a
80 project, feel free to use whatever style you are most comfortable with. If you
81 are extending, enhancing, or bug fixing already implemented code, use the style
82 that is already being used so that the source is uniform and easy to
87 <p>The ultimate goal of these guidelines is the increase readability and
88 maintainability of our common source base. If you have suggestions for topics to
89 be included, please mail them to <a
90 href="mailto:sabre@nondot.org">Chris</a>.</p>
94 <!-- *********************************************************************** -->
95 <div class="doc_section">
96 <a name="mechanicalissues">Mechanical Source Issues</a>
98 <!-- *********************************************************************** -->
100 <!-- ======================================================================= -->
101 <div class="doc_subsection">
102 <a name="sourceformating">Source Code Formatting</a>
105 <!-- _______________________________________________________________________ -->
106 <div class="doc_subsubsection">
107 <a name="scf_commenting">Commenting</a>
110 <div class="doc_text">
112 <p>Comments are one critical part of readability and maintainability. Everyone
113 knows they should comment, so should you. :) Although we all should probably
114 comment our code more than we do, there are a few very critical places that
115 documentation is very useful:</p>
119 <p>Every source file should have a header on it that
120 describes the basic purpose of the file. If a file does not have a header, it
121 should not be checked into CVS. Most source trees will probably have a standard
122 file header format. The standard format for the LLVM source tree looks like
125 <div class="doc_code">
127 //===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
129 // The LLVM Compiler Infrastructure
131 // This file was developed by the LLVM research group and is distributed under
132 // the University of Illinois Open Source License. See LICENSE.TXT for details.
134 //===----------------------------------------------------------------------===//
136 // This file contains the declaration of the Instruction class, which is the
137 // base class for all of the VM instructions.
139 //===----------------------------------------------------------------------===//
143 <p>A few things to note about this particular format: The "<tt>-*- C++
144 -*-</tt>" string on the first line is there to tell Emacs that the source file
145 is a C++ file, not a C file (Emacs assumes .h files are C files by default [Note
146 that tag this is not necessary in .cpp files]). The name of the file is also on
147 the first line, along with a very short description of the purpose of the file.
148 This is important when printing out code and flipping though lots of pages.</p>
150 <p>The next section in the file is a concise note that defines the license that
151 the file is released under. This makes it perfectly clear what terms the source
152 code can be distributed under.</p>
154 <p>The main body of the description does not have to be very long in most cases.
155 Here it's only two lines. If an algorithm is being implemented or something
156 tricky is going on, a reference to the paper where it is published should be
157 included, as well as any notes or "gotchas" in the code to watch out for.</p>
159 <b>Class overviews</b>
161 <p>Classes are one fundemental part of a good object oriented design. As such,
162 a class definition should have a comment block that explains what the class is
163 used for... if it's not obvious. If it's so completely obvious your grandma
164 could figure it out, it's probably safe to leave it out. Naming classes
165 something sane goes a long ways towards avoiding writing documentation. :)</p>
168 <b>Method information</b>
170 <p>Methods defined in a class (as well as any global functions) should also be
171 documented properly. A quick note about what it does any a description of the
172 borderline behaviour is all that is necessary here (unless something
173 particularly tricky or insideous is going on). The hope is that people can
174 figure out how to use your interfaces without reading the code itself... that is
177 <p>Good things to talk about here are what happens when something unexpected
178 happens: does the method return null? Abort? Format your hard disk?</p>
182 <!-- _______________________________________________________________________ -->
183 <div class="doc_subsubsection">
184 <a name="scf_commentformat">Comment Formatting</a>
187 <div class="doc_text">
189 <p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
190 require less typing, don't have nesting problems, etc. There are a few cases
191 when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
194 <li>When writing a C code: Obviously if you are writing C code, use C style
196 <li>When writing a header file that may be #included by a C source file.</li>
197 <li>When writing a source file that is used by a tool that only accepts C
201 <p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
202 These nest properly and are better behaved in general than C style comments.</p>
206 <!-- _______________________________________________________________________ -->
207 <div class="doc_subsubsection">
208 <a name="scf_includes">#include Style</a>
211 <div class="doc_text">
213 <p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
214 include guards if working on a header file), the <a
215 href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
216 file should be listed. We prefer these <tt>#include</tt>s to be listed in this
220 <li><a href="#mmheader">Main Module header</a></li>
221 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
222 <li><tt>llvm/*</tt></li>
223 <li><tt>llvm/Analysis/*</tt></li>
224 <li><tt>llvm/Assembly/*</tt></li>
225 <li><tt>llvm/Bytecode/*</tt></li>
226 <li><tt>llvm/CodeGen/*</tt></li>
228 <li><tt>Support/*</tt></li>
229 <li><tt>Config/*</tt></li>
230 <li>System <tt>#includes</tt></li>
233 <p>... and each catagory should be sorted by name.</p>
235 <p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
236 which implement an interface defined by a .h file. This #include should always
237 be included <b>first</b> regardless of where it lives on the file system. By
238 including a header file first in the .cpp files that implement the interfaces,
239 we ensure that the header does not have any hidden dependencies which are not
240 explicitly #included in the header, but should be. It is also a form of
241 documentation in the .cpp file to indicate where the interfaces it implements
246 <!-- _______________________________________________________________________ -->
247 <div class="doc_subsubsection">
248 <a name="scf_codewidth">Source Code Width</a>
251 <div class="doc_text">
253 <p>Write your code to fit within 80 columns of text. This helps those of us who
254 like to print out code and look at your code in an xterm without resizing
259 <!-- _______________________________________________________________________ -->
260 <div class="doc_subsubsection">
261 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
264 <div class="doc_text">
266 <p>In all cases, prefer spaces to tabs in source files. People have different
267 prefered indentation levels, and different styles of indentation that they
268 like... this is fine. What isn't is that different editors/viewers expand tabs
269 out to different tab stops. This can cause your code to look completely
270 unreadable, and it is not worth dealing with.</p>
272 <p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
273 style of existing code if your are modifying and extending it. If you like four
274 spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
275 with two spaces of indentation. Also, do not reindent a whole source file: it
276 makes for incredible diffs that are absolutely worthless.</p>
280 <!-- _______________________________________________________________________ -->
281 <div class="doc_subsubsection">
282 <a name="scf_indentation">Indent Code Consistently</a>
285 <div class="doc_text">
287 <p>Okay, your first year of programming you were told that indentation is
288 important. If you didn't believe and internalize this then, now is the time.
294 <!-- ======================================================================= -->
295 <div class="doc_subsection">
296 <a name="compilerissues">Compiler Issues</a>
300 <!-- _______________________________________________________________________ -->
301 <div class="doc_subsubsection">
302 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
305 <div class="doc_text">
307 <p>If your code has compiler warnings in it, something is wrong: you aren't
308 casting values correctly, your have "questionable" constructs in your code, or
309 you are doing something legitimately wrong. Compiler warnings can cover up
310 legitimate errors in output and make dealing with a translation unit
313 <p>It is not possible to prevent all warnings from all compilers, nor is it
314 desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
315 a good thorough set of warnings, and stick to them. At least in the case of
316 <tt>gcc</tt>, it is possible to work around any spurious errors by changing the
317 syntax of the code slightly. For example, an warning that annoys me occurs when
318 I write code like this:</p>
320 <div class="doc_code">
322 if (V = getValue()) {
328 <p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
329 operator, and that I probably mistyped it. In most cases, I haven't, and I
330 really don't want the spurious errors. To fix this particular problem, I
331 rewrite the code like this:</p>
333 <div class="doc_code">
335 if ((V = getValue())) {
341 <p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
342 be fixed by massaging the code appropriately.</p>
344 <p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
345 -Winline -W -Wwrite-strings -Wno-unused</tt></p>
349 <!-- _______________________________________________________________________ -->
350 <div class="doc_subsubsection">
351 <a name="ci_portable_code">Write Portable Code</a>
354 <div class="doc_text">
356 <p>In almost all cases, it is possible and within reason to write completely
357 portable code. If there are cases where it isn't possible to write portable
358 code, isolate it behind a well defined (and well documented) interface.</p>
360 <p>In practice, this means that you shouldn't assume much about the host
361 compiler, including its support for "high tech" features like partial
362 specialization of templates. In fact, Visual C++ 6 could be an important target
363 for our work in the future, and we don't want to have to rewrite all of our code
368 <!-- *********************************************************************** -->
369 <div class="doc_section">
370 <a name="styleissues">Style Issues</a>
372 <!-- *********************************************************************** -->
375 <!-- ======================================================================= -->
376 <div class="doc_subsection">
377 <a name="macro">The High Level Issues</a>
381 <!-- _______________________________________________________________________ -->
382 <div class="doc_subsubsection">
383 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
386 <div class="doc_text">
388 <p>C++ doesn't do too well in the modularity department. There is no real
389 encapsulation or data hiding (unless you use expensive protocol classes), but it
390 is what we have to work with. When you write a public header file (in the LLVM
391 source tree, they live in the top level "include" directory), you are defining a
392 module of functionality.</p>
394 <p>Ideally, modules should be completely independent of each other, and their
395 header files should only include the absolute minimum number of headers
396 possible. A module is not just a class, a function, or a namespace: <a
397 href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
398 of these</a> that defines an interface. This interface may be several
399 functions, classes or data structures, but the important issue is how they work
402 <p>In general, a module should be implemented with one or more <tt>.cpp</tt>
403 files. Each of these <tt>.cpp</tt> files should include the header that defines
404 their interface first. This ensure that all of the dependences of the module
405 header have been properly added to the module header itself, and are not
406 implicit. System headers should be included after user headers for a
407 translation unit.</p>
411 <!-- _______________________________________________________________________ -->
412 <div class="doc_subsubsection">
413 <a name="hl_dontinclude">#include as Little as Possible</a>
416 <div class="doc_text">
418 <p><tt>#include</tt> hurts compile time performance. Don't do it unless you
419 have to, especially in header files.</p>
421 <p>But wait, sometimes you need to have the definition of a class to use it, or
422 to inherit from it. In these cases go ahead and #include that header file. Be
423 aware however that there are many cases where you don't need to have the full
424 definition of a class. If you are using a pointer or reference to a class, you
425 don't need the header file. If you are simply returning a class instance from a
426 prototyped function or method, you don't need it. In fact, for most cases, you
427 simply don't need the definition of a class... and not <tt>#include</tt>'ing
428 speeds up compilation.</p>
430 <p>It is easy to try to go too overboard on this recommendation, however. You
431 <b>must</b> include all of the header files that you are using, either directly
432 or indirectly (through another header file). To make sure that you don't
433 accidently forget to include a header file in your module header, make sure to
434 include your module header <b>first</b> in the implementation file (as mentioned
435 above). This way there won't be any hidden dependencies that you'll find out
440 <!-- _______________________________________________________________________ -->
441 <div class="doc_subsubsection">
442 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
445 <div class="doc_text">
447 <p>Many modules have a complex implementation that causes them to use more than
448 one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
449 internal communication interface (helper classes, extra functions, etc) in the
450 public module header file. Don't do this. :)</p>
452 <p>If you really need to do something like this, put a private header file in
453 the same directory as the source files, and include it locally. This ensures
454 that your private interface remains private and undisturbed by outsiders.</p>
456 <p>Note however, that it's okay to put extra implementation methods a public
457 class itself... just make them private (or protected), and all is well.</p>
461 <!-- ======================================================================= -->
462 <div class="doc_subsection">
463 <a name="micro">The Low Level Issues</a>
467 <!-- _______________________________________________________________________ -->
468 <div class="doc_subsubsection">
469 <a name="hl_assert">Assert Liberally</a>
472 <div class="doc_text">
474 <p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
475 preconditions and assumptions, you never know when a bug (not neccesarily even
476 yours) might be caught early by an assertion, which reduces debugging time
477 dramatically. The "<tt><cassert></tt>" header file is probably already
478 included by the header files you are using, so it doesn't cost anything to use
481 <p>To further assist with debugging, make sure to put some kind of error message
482 in the assertion statement (which is printed if the assertion is tripped). This
483 helps the poor debugging make sense of why an assertion is being made and
484 enforced, and hopefully what to do about it. Here is one complete example:</p>
486 <div class="doc_code">
488 inline Value *getOperand(unsigned i) {
489 assert(i < Operands.size() && "getOperand() out of range!");
495 <p>Here are some examples:</p>
497 <div class="doc_code">
499 assert(Ty->isPointerType() && "Can't allocate a non pointer type!");
501 assert((Opcode == Shl || Opcode == Shr) && "ShiftInst Opcode invalid!");
503 assert(idx < getNumSuccessors() && "Successor # out of range!");
505 assert(V1.getType() == V2.getType() && "Constant types must be identical!");
507 assert(isa<PHINode>(Succ->front()) && "Only works on PHId BBs!");
511 <p>You get the idea...</p>
516 <!-- _______________________________________________________________________ -->
517 <div class="doc_subsubsection">
518 <a name="hl_preincrement">Prefer Preincrement</a>
521 <div class="doc_text">
523 <p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
524 postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
525 preincrementation whenever possible.</p>
527 <p>The semantics of postincrement include making a copy of the value being
528 incremented, returning it, and then preincrementing the "work value". For
529 primitive types, this isn't a big deal... but for iterators, it can be a huge
530 issue (for example, some iterators contains stack and set objects in them...
531 copying an iterator could invoke the copy ctor's of these as well). In general,
532 get in the habit of always using preincrement, and you won't have a problem.</p>
536 <!-- _______________________________________________________________________ -->
537 <div class="doc_subsubsection">
538 <a name="hl_avoidendl">Avoid std::endl</a>
541 <div class="doc_text">
543 <p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
544 to the output stream specified. In addition to doing this, however, it also
545 flushes the output stream. In other words, these are equivalent:</p>
547 <div class="doc_code">
549 std::cout << std::endl;
550 std::cout << '\n' << std::flush;
554 <p>Most of the time, you probably have no reason to flush the output stream, so
555 it's better to use a literal <tt>'\n'</tt>.</p>
559 <!-- _______________________________________________________________________ -->
560 <div class="doc_subsubsection">
561 <a name="hl_exploitcpp">Exploit C++ to its Fullest</a>
564 <div class="doc_text">
566 <p>C++ is a powerful language. With a firm grasp on its capabilities, you can
567 make write effective, consise, readable and maintainable code all at the same
568 time. By staying consistent, you reduce the amount of special cases that need
569 to be remembered. Reducing the total number of lines of code you write is a
570 good way to avoid documentation, and avoid giving bugs a place to hide.</p>
572 <p>For these reasons, come to know and love the contents of your local
573 <algorithm> header file. Know about <functional> and what it can do
574 for you. C++ is just a tool that wants you to master it. :)</p>
578 <!-- *********************************************************************** -->
579 <div class="doc_section">
580 <a name="seealso">See Also</a>
582 <!-- *********************************************************************** -->
584 <div class="doc_text">
586 <p>A lot of these comments and recommendations have been culled for other
587 sources. Two particularly important books for our work are:</p>
591 <li><a href="http://www.aw-bc.com/catalog/academic/product/0,1144,0201310155,00.html">Effective
592 C++</a> by Scott Meyers. There is an online version of the book (only some
594 href="http://www.awlonline.com/cseng/meyerscddemo/">available as well</a>. Also
595 interesting and useful are "More Effective C++" and "Effective STL" by the same
598 <li><a href="http://cseng.aw.com/book/0,3828,0201633620,00.html">Large-Scale C++
599 Software Design</a> by John Lakos</li>
603 <p>If you get some free time, and you haven't read them: do so, you might learn
608 <!-- *********************************************************************** -->
612 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
613 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
614 <a href="http://validator.w3.org/check/referer"><img
615 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
617 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
618 <a href="http://llvm.cs.uiuc.edu">LLVM Compiler Infrastructure</a><br>
619 Last modified: $Date$