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"><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>
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>
32 <li><a href="#ci_class_struct">Use of class/struct Keywords</a></li>
35 <li><a href="#styleissues">Style Issues</a>
37 <li><a href="#macro">The High Level Issues</a>
39 <li><a href="#hl_module">A Public Header File <b>is</b> a
41 <li><a href="#hl_dontinclude">#include as Little as Possible</a></li>
42 <li><a href="#hl_privateheaders">Keep "internal" Headers
45 <li><a href="#micro">The Low Level Issues</a>
47 <li><a href="#hl_assert">Assert Liberally</a></li>
48 <li><a href="#hl_ns_std">Do not use 'using namespace std'</a></li>
49 <li><a href="#hl_preincrement">Prefer Preincrement</a></li>
50 <li><a href="#hl_avoidendl">Avoid <tt>std::endl</tt></a></li>
53 <li><a href="#seealso">See Also</a></li>
56 <div class="doc_author">
57 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a></p>
61 <!-- *********************************************************************** -->
62 <div class="doc_section">
63 <a name="introduction">Introduction</a>
65 <!-- *********************************************************************** -->
67 <div class="doc_text">
69 <p>This document attempts to describe a few coding standards that are being used
70 in the LLVM source tree. Although no coding standards should be regarded as
71 absolute requirements to be followed in all instances, coding standards can be
74 <p>This document intentionally does not prescribe fixed standards for religious
75 issues such as brace placement and space usage. For issues like this, follow
80 <p><b><a name="goldenrule">If you are adding a significant body of source to a
81 project, feel free to use whatever style you are most comfortable with. If you
82 are extending, enhancing, or bug fixing already implemented code, use the style
83 that is already being used so that the source is uniform and easy to
88 <p>The ultimate goal of these guidelines is the increase readability and
89 maintainability of our common source base. If you have suggestions for topics to
90 be included, please mail them to <a
91 href="mailto:sabre@nondot.org">Chris</a>.</p>
95 <!-- *********************************************************************** -->
96 <div class="doc_section">
97 <a name="mechanicalissues">Mechanical Source Issues</a>
99 <!-- *********************************************************************** -->
101 <!-- ======================================================================= -->
102 <div class="doc_subsection">
103 <a name="sourceformating">Source Code Formatting</a>
106 <!-- _______________________________________________________________________ -->
107 <div class="doc_subsubsection">
108 <a name="scf_commenting">Commenting</a>
111 <div class="doc_text">
113 <p>Comments are one critical part of readability and maintainability. Everyone
114 knows they should comment, so should you. Although we all should probably
115 comment our code more than we do, there are a few very critical places that
116 documentation is very useful:</p>
120 <p>Every source file should have a header on it that
121 describes the basic purpose of the file. If a file does not have a header, it
122 should not be checked into CVS. Most source trees will probably have a standard
123 file header format. The standard format for the LLVM source tree looks like
126 <div class="doc_code">
128 //===-- llvm/Instruction.h - Instruction class definition -------*- C++ -*-===//
130 // The LLVM Compiler Infrastructure
132 // This file was developed by the LLVM research group and is distributed under
133 // the University of Illinois Open Source License. See LICENSE.TXT for details.
135 //===----------------------------------------------------------------------===//
137 // This file contains the declaration of the Instruction class, which is the
138 // base class for all of the VM instructions.
140 //===----------------------------------------------------------------------===//
144 <p>A few things to note about this particular format: The "<tt>-*- C++
145 -*-</tt>" string on the first line is there to tell Emacs that the source file
146 is a C++ file, not a C file (Emacs assumes .h files are C files by default).
147 Note that this tag is not necessary in .cpp files. The name of the file is also
148 on the first line, along with a very short description of the purpose of the
149 file. This is important when printing out code and flipping though lots of
152 <p>The next section in the file is a concise note that defines the license that
153 the file is released under. This makes it perfectly clear what terms the source
154 code can be distributed under.</p>
156 <p>The main body of the description does not have to be very long in most cases.
157 Here it's only two lines. If an algorithm is being implemented or something
158 tricky is going on, a reference to the paper where it is published should be
159 included, as well as any notes or "gotchas" in the code to watch out for.</p>
161 <b>Class overviews</b>
163 <p>Classes are one fundemental part of a good object oriented design. As such,
164 a class definition should have a comment block that explains what the class is
165 used for... if it's not obvious. If it's so completely obvious your grandma
166 could figure it out, it's probably safe to leave it out. Naming classes
167 something sane goes a long ways towards avoiding writing documentation.</p>
170 <b>Method information</b>
172 <p>Methods defined in a class (as well as any global functions) should also be
173 documented properly. A quick note about what it does any a description of the
174 borderline behaviour is all that is necessary here (unless something
175 particularly tricky or insideous is going on). The hope is that people can
176 figure out how to use your interfaces without reading the code itself... that is
179 <p>Good things to talk about here are what happens when something unexpected
180 happens: does the method return null? Abort? Format your hard disk?</p>
184 <!-- _______________________________________________________________________ -->
185 <div class="doc_subsubsection">
186 <a name="scf_commentformat">Comment Formatting</a>
189 <div class="doc_text">
191 <p>In general, prefer C++ style (<tt>//</tt>) comments. They take less space,
192 require less typing, don't have nesting problems, etc. There are a few cases
193 when it is useful to use C style (<tt>/* */</tt>) comments however:</p>
196 <li>When writing a C code: Obviously if you are writing C code, use C style
198 <li>When writing a header file that may be <tt>#include</tt>d by a C source
200 <li>When writing a source file that is used by a tool that only accepts C
204 <p>To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>.
205 These nest properly and are better behaved in general than C style comments.</p>
209 <!-- _______________________________________________________________________ -->
210 <div class="doc_subsubsection">
211 <a name="scf_includes"><tt>#include</tt> Style</a>
214 <div class="doc_text">
216 <p>Immediately after the <a href="#scf_commenting">header file comment</a> (and
217 include guards if working on a header file), the <a
218 href="#hl_dontinclude">minimal</a> list of <tt>#include</tt>s required by the
219 file should be listed. We prefer these <tt>#include</tt>s to be listed in this
223 <li><a href="#mmheader">Main Module header</a></li>
224 <li><a href="#hl_privateheaders">Local/Private Headers</a></li>
225 <li><tt>llvm/*</tt></li>
226 <li><tt>llvm/Analysis/*</tt></li>
227 <li><tt>llvm/Assembly/*</tt></li>
228 <li><tt>llvm/Bytecode/*</tt></li>
229 <li><tt>llvm/CodeGen/*</tt></li>
231 <li><tt>Support/*</tt></li>
232 <li><tt>Config/*</tt></li>
233 <li>System <tt>#includes</tt></li>
236 <p>... and each catagory should be sorted by name.</p>
238 <p><a name="mmheader">The "Main Module Header"</a> file applies to .cpp file
239 which implement an interface defined by a .h file. This <tt>#include</tt>
240 should always be included <b>first</b> regardless of where it lives on the file
241 system. By including a header file first in the .cpp files that implement the
242 interfaces, we ensure that the header does not have any hidden dependencies
243 which are not explicitly #included in the header, but should be. It is also a
244 form of documentation in the .cpp file to indicate where the interfaces it
245 implements are defined.</p>
249 <!-- _______________________________________________________________________ -->
250 <div class="doc_subsubsection">
251 <a name="scf_codewidth">Source Code Width</a>
254 <div class="doc_text">
256 <p>Write your code to fit within 80 columns of text. This helps those of us who
257 like to print out code and look at your code in an xterm without resizing
262 <!-- _______________________________________________________________________ -->
263 <div class="doc_subsubsection">
264 <a name="scf_spacestabs">Use Spaces Instead of Tabs</a>
267 <div class="doc_text">
269 <p>In all cases, prefer spaces to tabs in source files. People have different
270 prefered indentation levels, and different styles of indentation that they
271 like... this is fine. What isn't is that different editors/viewers expand tabs
272 out to different tab stops. This can cause your code to look completely
273 unreadable, and it is not worth dealing with.</p>
275 <p>As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the
276 style of existing code if your are modifying and extending it. If you like four
277 spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code
278 with two spaces of indentation. Also, do not reindent a whole source file: it
279 makes for incredible diffs that are absolutely worthless.</p>
283 <!-- _______________________________________________________________________ -->
284 <div class="doc_subsubsection">
285 <a name="scf_indentation">Indent Code Consistently</a>
288 <div class="doc_text">
290 <p>Okay, your first year of programming you were told that indentation is
291 important. If you didn't believe and internalize this then, now is the time.
297 <!-- ======================================================================= -->
298 <div class="doc_subsection">
299 <a name="compilerissues">Compiler Issues</a>
303 <!-- _______________________________________________________________________ -->
304 <div class="doc_subsubsection">
305 <a name="ci_warningerrors">Treat Compiler Warnings Like Errors</a>
308 <div class="doc_text">
310 <p>If your code has compiler warnings in it, something is wrong: you aren't
311 casting values correctly, your have "questionable" constructs in your code, or
312 you are doing something legitimately wrong. Compiler warnings can cover up
313 legitimate errors in output and make dealing with a translation unit
316 <p>It is not possible to prevent all warnings from all compilers, nor is it
317 desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides
318 a good thorough set of warnings, and stick to them. At least in the case of
319 <tt>gcc</tt>, it is possible to work around any spurious errors by changing the
320 syntax of the code slightly. For example, an warning that annoys me occurs when
321 I write code like this:</p>
323 <div class="doc_code">
325 if (V = getValue()) {
331 <p><tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt>
332 operator, and that I probably mistyped it. In most cases, I haven't, and I
333 really don't want the spurious errors. To fix this particular problem, I
334 rewrite the code like this:</p>
336 <div class="doc_code">
338 if ((V = getValue())) {
344 <p>...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can
345 be fixed by massaging the code appropriately.</p>
347 <p>These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall
348 -Winline -W -Wwrite-strings -Wno-unused</tt></p>
352 <!-- _______________________________________________________________________ -->
353 <div class="doc_subsubsection">
354 <a name="ci_portable_code">Write Portable Code</a>
357 <div class="doc_text">
359 <p>In almost all cases, it is possible and within reason to write completely
360 portable code. If there are cases where it isn't possible to write portable
361 code, isolate it behind a well defined (and well documented) interface.</p>
363 <p>In practice, this means that you shouldn't assume much about the host
364 compiler, including its support for "high tech" features like partial
365 specialization of templates. In fact, Visual C++ 6 could be an important target
366 for our work in the future, and we don't want to have to rewrite all of our code
371 <!-- _______________________________________________________________________ -->
372 <div class="doc_subsubsection">
373 <a name="ci_class_struct">Use of <tt>class</tt> and <tt>struct</tt> Keywords</a>
375 <div class="doc_text">
377 <p>In C++, the <tt>class</tt> and <tt>struct</tt> keywords can be used almost
378 interchangeably. The only difference is when they are used to declare a class:
379 <tt>class</tt> makes all members private by default while <tt>struct</tt> makes
380 all members public by default.</p>
382 <p>Unfortunately, not all compilers follow the rules and some will generate
383 different symbols based on whether <tt>class</tt> or <tt>struct</tt> was used to
384 declare the symbol. This can lead to problems at link time.</p>
386 <p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
387 <b>all</b> members are public, in which case <tt>struct</tt> is allowed.</p>
391 <!-- *********************************************************************** -->
392 <div class="doc_section">
393 <a name="styleissues">Style Issues</a>
395 <!-- *********************************************************************** -->
398 <!-- ======================================================================= -->
399 <div class="doc_subsection">
400 <a name="macro">The High Level Issues</a>
404 <!-- _______________________________________________________________________ -->
405 <div class="doc_subsubsection">
406 <a name="hl_module">A Public Header File <b>is</b> a Module</a>
409 <div class="doc_text">
411 <p>C++ doesn't do too well in the modularity department. There is no real
412 encapsulation or data hiding (unless you use expensive protocol classes), but it
413 is what we have to work with. When you write a public header file (in the LLVM
414 source tree, they live in the top level "include" directory), you are defining a
415 module of functionality.</p>
417 <p>Ideally, modules should be completely independent of each other, and their
418 header files should only include the absolute minimum number of headers
419 possible. A module is not just a class, a function, or a namespace: <a
420 href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
421 of these</a> that defines an interface. This interface may be several
422 functions, classes or data structures, but the important issue is how they work
425 <p>In general, a module should be implemented with one or more <tt>.cpp</tt>
426 files. Each of these <tt>.cpp</tt> files should include the header that defines
427 their interface first. This ensure that all of the dependences of the module
428 header have been properly added to the module header itself, and are not
429 implicit. System headers should be included after user headers for a
430 translation unit.</p>
434 <!-- _______________________________________________________________________ -->
435 <div class="doc_subsubsection">
436 <a name="hl_dontinclude"><tt>#include</tt> as Little as Possible</a>
439 <div class="doc_text">
441 <p><tt>#include</tt> hurts compile time performance. Don't do it unless you
442 have to, especially in header files.</p>
444 <p>But wait, sometimes you need to have the definition of a class to use it, or
445 to inherit from it. In these cases go ahead and <tt>#include</tt> that header
446 file. Be aware however that there are many cases where you don't need to have
447 the full definition of a class. If you are using a pointer or reference to a
448 class, you don't need the header file. If you are simply returning a class
449 instance from a prototyped function or method, you don't need it. In fact, for
450 most cases, you simply don't need the definition of a class... and not
451 <tt>#include</tt>'ing speeds up compilation.</p>
453 <p>It is easy to try to go too overboard on this recommendation, however. You
454 <b>must</b> include all of the header files that you are using, either directly
455 or indirectly (through another header file). To make sure that you don't
456 accidently forget to include a header file in your module header, make sure to
457 include your module header <b>first</b> in the implementation file (as mentioned
458 above). This way there won't be any hidden dependencies that you'll find out
463 <!-- _______________________________________________________________________ -->
464 <div class="doc_subsubsection">
465 <a name="hl_privateheaders">Keep "internal" Headers Private</a>
468 <div class="doc_text">
470 <p>Many modules have a complex implementation that causes them to use more than
471 one implementation (<tt>.cpp</tt>) file. It is often tempting to put the
472 internal communication interface (helper classes, extra functions, etc) in the
473 public module header file. Don't do this.</p>
475 <p>If you really need to do something like this, put a private header file in
476 the same directory as the source files, and include it locally. This ensures
477 that your private interface remains private and undisturbed by outsiders.</p>
479 <p>Note however, that it's okay to put extra implementation methods a public
480 class itself... just make them private (or protected), and all is well.</p>
484 <!-- ======================================================================= -->
485 <div class="doc_subsection">
486 <a name="micro">The Low Level Issues</a>
490 <!-- _______________________________________________________________________ -->
491 <div class="doc_subsubsection">
492 <a name="hl_assert">Assert Liberally</a>
495 <div class="doc_text">
497 <p>Use the "<tt>assert</tt>" function to its fullest. Check all of your
498 preconditions and assumptions, you never know when a bug (not neccesarily even
499 yours) might be caught early by an assertion, which reduces debugging time
500 dramatically. The "<tt><cassert></tt>" header file is probably already
501 included by the header files you are using, so it doesn't cost anything to use
504 <p>To further assist with debugging, make sure to put some kind of error message
505 in the assertion statement (which is printed if the assertion is tripped). This
506 helps the poor debugging make sense of why an assertion is being made and
507 enforced, and hopefully what to do about it. Here is one complete example:</p>
509 <div class="doc_code">
511 inline Value *getOperand(unsigned i) {
512 assert(i < Operands.size() && "getOperand() out of range!");
518 <p>Here are some examples:</p>
520 <div class="doc_code">
522 assert(Ty->isPointerType() && "Can't allocate a non pointer type!");
524 assert((Opcode == Shl || Opcode == Shr) && "ShiftInst Opcode invalid!");
526 assert(idx < getNumSuccessors() && "Successor # out of range!");
528 assert(V1.getType() == V2.getType() && "Constant types must be identical!");
530 assert(isa<PHINode>(Succ->front()) && "Only works on PHId BBs!");
534 <p>You get the idea...</p>
538 <!-- _______________________________________________________________________ -->
539 <div class="doc_subsubsection">
540 <a name="hl_ns_std">Do not use 'using namespace std'</a>
543 <div class="doc_text">
544 <p>In LLVM, we prefer to explicitly prefix all identifiers from the standard
545 namespace with an "std::" prefix, rather than rely on "using namespace std;".
548 <p> In header files, adding a 'using namespace XXX' directive pollutes the
549 namespace of any source file that includes the header. This is clearly a bad
552 <p>In implementation files (e.g. .cpp files) the rule is more of a stylistic
553 rule, but is still important. Basically, using explicit namespace prefixes
555 the code <b>more clear</b> - because it is immediately obvious what facilities
556 are being used and where they are coming from - and <b>more portable</b> -
557 because namespace clashes cannot occur between LLVM code and other namespaces.
558 The portability rule is important because different standard library
559 implementations expose different symbols (potentially ones they shouldn't) and
560 future revisions to the C++ standard will add more symbols to the std
561 namespace. As such, we never 'using namespace std;' in LLVM.</p>
563 <p>The exception to the general rule (i.e. it's not an exception for the std
564 namespace) is for implementation files. For example, all of the code in the
565 LLVM project implements code that lives in the 'llvm' namespace. As such, it
566 is ok, and actually more clear, for the .cpp files to have a 'using namespace
567 llvm' directive at their top, after the #includes. The general form of this
568 rule is that any .cpp file that implements code in any namespace may use that
569 namespace (and its parents), but should not use any others.</p>
573 <!-- _______________________________________________________________________ -->
574 <div class="doc_subsubsection">
575 <a name="hl_preincrement">Prefer Preincrement</a>
578 <div class="doc_text">
580 <p>Hard fast rule: Preincrement (<tt>++X</tt>) may be no slower than
581 postincrement (<tt>X++</tt>) and could very well be a lot faster than it. Use
582 preincrementation whenever possible.</p>
584 <p>The semantics of postincrement include making a copy of the value being
585 incremented, returning it, and then preincrementing the "work value". For
586 primitive types, this isn't a big deal... but for iterators, it can be a huge
587 issue (for example, some iterators contains stack and set objects in them...
588 copying an iterator could invoke the copy ctor's of these as well). In general,
589 get in the habit of always using preincrement, and you won't have a problem.</p>
593 <!-- _______________________________________________________________________ -->
594 <div class="doc_subsubsection">
595 <a name="hl_avoidendl">Avoid <tt>std::endl</tt></a>
598 <div class="doc_text">
600 <p>The <tt>std::endl</tt> modifier, when used with iostreams outputs a newline
601 to the output stream specified. In addition to doing this, however, it also
602 flushes the output stream. In other words, these are equivalent:</p>
604 <div class="doc_code">
606 std::cout << std::endl;
607 std::cout << '\n' << std::flush;
611 <p>Most of the time, you probably have no reason to flush the output stream, so
612 it's better to use a literal <tt>'\n'</tt>.</p>
616 <!-- *********************************************************************** -->
617 <div class="doc_section">
618 <a name="seealso">See Also</a>
620 <!-- *********************************************************************** -->
622 <div class="doc_text">
624 <p>A lot of these comments and recommendations have been culled for other
625 sources. Two particularly important books for our work are:</p>
629 <li><a href="http://www.aw-bc.com/catalog/academic/product/0,1144,0201310155,00.html">Effective
630 C++</a> by Scott Meyers. There is an online version of the book (only some
632 href="http://www.awlonline.com/cseng/meyerscddemo/">available as well</a>. Also
633 interesting and useful are "More Effective C++" and "Effective STL" by the same
636 <li><a href="http://cseng.aw.com/book/0,3828,0201633620,00.html">Large-Scale C++
637 Software Design</a> by John Lakos</li>
641 <p>If you get some free time, and you haven't read them: do so, you might learn
646 <!-- *********************************************************************** -->
650 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
651 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
652 <a href="http://validator.w3.org/check/referer"><img
653 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
655 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
656 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
657 Last modified: $Date$