1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
2 <html><head><title>A Few Coding Standards</title></head>
5 <table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0>
6 <tr><td> <font size=+5 color="#EEEEFF" face="Georgia,Palatino,Times,Roman"><b>A Few Coding Standards</b></font></td>
10 <li><a href="#introduction">Introduction</a>
11 <li><a href="#mechanicalissues">Mechanical Source Issues</a>
13 <li><a href="#sourceformating">Source Code Formatting</a>
15 <li><a href="#scf_commenting">Commenting</a>
16 <li><a href="#scf_commentformat">Comment Formatting</a>
17 <li><a href="#scf_codewidth">Source Code Width</a>
18 <li><a href="#scf_spacestabs">Use Spaces Instead of Tabs</a>
19 <li><a href="#scf_indentation">Indent Code Consistently</a>
21 <li><a href="#compilerissues">Compiler Issues</a>
23 <li><a href="#ci_warningerrors">Treat Compiler Warnings Like Errors</a>
24 <li><a href="#ci_cpp_features">Which C++ features can I use?</a>
25 <li><a href="#ci_portable_code">Write Portable Code</a>
28 <li><a href="#styleissues">Style Issues</a>
30 <li><a href="#macro">The High Level Issues</a>
32 <li><a href="#hl_module">A Public Header File <b>is</b> a Module</a>
33 <li><a href="#hl_dontinclude">#include as Little as Possible</a>
34 <li><a href="#hl_privateheaders">Keep "internal" Headers Private</a>
36 <li><a href="#micro">The Low Level Issues</a>
38 <li><a href="#hl_assert">Assert Liberally</a>
39 <li><a href="#hl_preincrement">Prefer Preincrement</a>
40 <li><a href="#hl_avoidendl">Avoid endl</a>
41 <li><a href="#hl_exploitcpp">Exploit C++ to its Fullest</a>
43 <li><a href="#iterators">Writing Iterators</a>
45 <li><a href="#seealso">See Also</a>
49 <!-- *********************************************************************** -->
50 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0><tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
51 <a name="introduction">Introduction
52 </b></font></td></tr></table><ul>
53 <!-- *********************************************************************** -->
55 This document attempts to describe a few coding standards that are being used in the LLVM source tree. Although no coding standards should be regarded as absolute requirements to be followed in all instances, coding standards can be useful.<p>
57 This document intentionally does not prescribe fixed standards for religious issues such as brace placement and space usage. For issues like this, follow the golden rule:
60 <blockquote><b>If you are adding a significant body of source to a project, feel free to use whatever style you are most comfortable with. If you are extending, enhancing, or bug fixing already implemented code, use the style that is already being used so that the source is uniform and easy to follow.</b></blockquote>
62 The ultimate goal of these guidelines is the increase readability and maintainability of our common source base. If you have suggestions for topics to be included, please mail them to <a href="mailto:sabre@nondot.org">Chris</a>.<p>
65 <!-- *********************************************************************** -->
66 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0><tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
67 <a name="mechanicalissues">Mechanical Source Issues
68 </b></font></td></tr></table><ul>
69 <!-- *********************************************************************** -->
71 <!-- ======================================================================= -->
72 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
73 <a name="sourceformating">Source Code Formatting
74 </b></font></td></tr></table><ul>
77 <!-- _______________________________________________________________________ -->
78 </ul><a name="scf_commenting"><h4><hr size=0>Commenting</h4><ul>
80 Comments are one critical part of readability and maintainability. Everyone knows they should comment, so should you. :) Although we all should probably comment our code more than we do, there are a few very critical places that documentation is very useful:<p>
83 <h4><li>File Headers</h4>
84 Every source file should have a header on it that describes the basic purpose of the file. If a file does not have a header, it should not be checked into CVS. Most source trees will probably have a standard file header format. The standard format for the LLVM source tree looks like this:<p>
87 //===-- llvm/Instruction.h - Instruction class definition --------*- C++ -*--=//
89 // This file contains the declaration of the Instruction class, which is the
90 // base class for all of the VM instructions.
92 //===----------------------------------------------------------------------===//
95 A few things to note about this particular format. The "<tt>-*- C++ -*-</tt>" string on the first line is there to tell Emacs that the source file is a C++ file, not a C file (Emacs assumes .h files are C files by default [Note that tag this is not neccesary in .cpp files]). The name of the file is also on the first line, along with a very short description of the purpose of the file. This is important when printing out code and flipping though lots of pages.<p>
97 The main body of the description does not have to be very long in most cases. Here it's only two lines. If an algorithm is being implemented or something tricky is going on, a reference to the paper where it is published should be included, as well as any notes or "gotchas" in the code to watch out for.<p>
100 <h4><li>Class overviews</h4>
102 Classes are one fundemental part of a good object oriented design. As such, a class definition should have a comment block that explains what the class is used for... if it's not obvious. If it's so completely obvious your grandma could figure it out, it's probably safe to leave it out. Naming classes something sane goes a long ways towards avoiding writing documentation. :)<p>
105 <h4><li>Method information</h4>
107 Methods defined in a class (as well as any global functions) should also be documented properly. A quick note about what it does any a description of the borderline behaviour is all that is neccesary here (unless something particularly tricky or insideous is going on). The hope is that people can figure out how to use your interfaces without reading the code itself... that is the goal metric.<p>
109 Good things to talk about here are what happens when something unexpected happens: does the method return null? Abort? Format your hard disk?<p>
113 <!-- _______________________________________________________________________ -->
114 </ul><a name="scf_commentformat"><h4><hr size=0>Comment Formatting</h4><ul>
116 In general, prefer C++ style (<tt>//</tt>) comments. They take less space, require less typing, don't have nesting problems, etc. There are a few cases when it is useful to use C style (<tt>/* */</tt>) comments however:<p>
119 <li>When writing a C code: Obviously if you are writing C code, use C style comments. :)
120 <li>When writing a header file that may be #included by a C source file.
121 <li>When writing a source file that is used by a tool that only accepts C style comments.
124 To comment out a large block of code, use <tt>#if 0</tt> and <tt>#endif</tt>. These nest properly and are better behaved in general than C style comments.<p>
127 <!-- _______________________________________________________________________ -->
128 </ul><a name="scf_codewidth"><h4><hr size=0>Source Code Width</h4><ul>
130 Write your code to fit within 80 columns of text. This helps those of us who like to print out code and look at your code in an xterm without resizing it.
133 <!-- _______________________________________________________________________ -->
134 </ul><a name="scf_spacestabs"><h4><hr size=0>Use Spaces Instead of Tabs</h4><ul>
136 In all cases, prefer spaces to tabs in source files. People have different prefered indentation levels, and different styles of indentation that they like... this is fine. What isn't is that different editors/viewers expand tabs out to different tab stops. This can cause your code to look completely unreadable, and it is not worth dealing with.<p>
138 As always, follow the <a href="#goldenrule">Golden Rule</a> above: follow the style of existing code if your are modifying and extending it. If you like four spaces of indentation, <b>DO NOT</b> do that in the middle of a chunk of code with two spaces of indentation. Also, do not reindent a whole source file: it make for incredible diffs that are absolutely worthless.<p>
141 <!-- _______________________________________________________________________ -->
142 </ul><a name="scf_indentation"><h4><hr size=0>Indent Code Consistently</h4><ul>
144 Okay, your first year of programming you were told that indentation is important. If you didn't believe and internalize this then, now is the time. Just do it.<p>
149 <!-- ======================================================================= -->
150 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
151 <a name="compilerissues">Compiler Issues
152 </b></font></td></tr></table><ul>
155 <!-- _______________________________________________________________________ -->
156 </ul><a name="ci_warningerrors"><h4><hr size=0>Treat Compiler Warnings Like Errors</h4><ul>
158 If your code has compiler warnings in it, something is wrong: you aren't casting values correctly, your have "questionable" constructs in your code, or you are doing something legitimately wrong. Compiler warnings can cover up legitimate errors in output and make dealing with a translation unit difficult.<p>
160 It is not possible to prevent all warnings from all compilers, nor is it desirable. Instead, pick a standard compiler (like <tt>gcc</tt>) that provides a good thorough set of warnings, and stick to them. At least in the case of <tt>gcc</tt>, it is possible to work around any spurious errors by changing the syntax of the code slightly. For example, an warning that annoys me occurs when I write code like this:<p>
163 if (V = getValue()) {
168 <tt>gcc</tt> will warn me that I probably want to use the <tt>==</tt> operator, and that I probably mistyped it. In most cases, I haven't, and I really don't want the spurious errors. To fix this particular problem, I rewrite the code like this:<p>
171 if ((V = getValue())) {
176 ...which shuts <tt>gcc</tt> up. Any <tt>gcc</tt> warning that annoys you can be
177 fixed by massaging the code appropriately.<p>
179 These are the <tt>gcc</tt> warnings that I prefer to enable: <tt>-Wall -Winline
180 -W -Wwrite-strings -Wno-unused</tt><p>
183 <!-- _______________________________________________________________________ -->
184 </ul><a name="ci_cpp_features"><h4><hr size=0>Which C++ features can I use?</h4><ul>
186 Compilers are finally catching up to the C++ standard. Most compilers implement
187 most features, so you can use just about any features that you would like. In
188 the LLVM source tree, I have chosen to not use these features:<p>
191 <li>Exceptions: Exceptions are very useful for error reporting and handling
192 exceptional conditions. I do not use them in LLVM because they do have an
193 associated performance impact (by restricting restructuring of code), and parts
194 of LLVM are designed for performance critical purposes.<p>
196 Just like most of the rules in this document, this isn't a hard and fast
197 requirement. Exceptions are used in the Parser, because it simplifies error
198 reporting <b>significantly</b>, and the LLVM parser is not at all in the
201 <li>RTTI: RTTI has a large cost in terms of executable size, and compilers are
202 not yet very good at stomping out "dead" class information blocks. Because of
203 this, typeinfo and dynamic cast are not used.
206 Other features, such as templates (without partial specialization) can be used
207 freely. The general goal is to have clear, consise, performant code... if a
208 technique assists with that then use it.<p>
211 <!-- _______________________________________________________________________ -->
212 </ul><a name="ci_portable_code"><h4><hr size=0>Write Portable Code</h4><ul>
214 In almost all cases, it is possible and within reason to write completely
215 portable code. If there are cases where it isn't possible to write portable
216 code, isolate it behind a well defined (and well documented) interface.<p>
218 In practice, this means that you shouldn't assume much about the host compiler,
219 including its support for "high tech" features like partial specialization of
220 templates. In fact, Visual C++ 6 could be an important target for our work in
221 the future, and we don't want to have to rewrite all of our code to support
226 <!-- *********************************************************************** -->
227 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0><tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
228 <a name="styleissues">Style Issues
229 </b></font></td></tr></table><ul>
230 <!-- *********************************************************************** -->
233 <!-- ======================================================================= -->
234 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
235 <a name="macro">The High Level Issues
236 </b></font></td></tr></table><ul>
239 <!-- _______________________________________________________________________ -->
240 </ul><a name="hl_module"><h4><hr size=0>A Public Header File <b>is</b> a Module</h4><ul>
242 C++ doesn't do too well in the modularity department. There is no real
243 encapsulation or data hiding (unless you use expensive protocol classes), but it
244 is what we have to work with. When you write a public header file (in the LLVM
245 source tree, they live in the top level "include" directory), you are defining a
246 module of functionality.<p>
248 Ideally, modules should be completely independent of each other, and their
249 header files should only include the absolute minimum number of headers
250 possible. A module is not just a class, a function, or a namespace: <a
251 href="http://www.cuj.com/articles/2000/0002/0002c/0002c.htm">it's a collection
252 of these</a> that defines an interface. This interface may be several
253 functions, classes or data structures, but the important issue is how they work
256 <!--One example of this is the <tt>llvm/include/llvm/CFG.h</tt> file. It
257 defines a collection of global functions, template classes, and member functions
258 that are syntactically unrelated to each other. Semantically, however, they all
259 provide useful functionality for operating on a CFG, and so they are bound
262 In general, a module should be implemented with one or more <tt>.cpp</tt> files.
263 Each of these <tt>.cpp</tt> files should include the header that defines their
264 interface first. This ensure that all of the dependences of the module header
265 have been properly added to the module header itself, and are not implicit.
266 System headers should be included after user headers for a translation unit.<p>
269 <!-- _______________________________________________________________________ -->
270 </ul><a name="hl_dontinclude"><h4><hr size=0>#include as Little as Possible</h4><ul>
272 <tt>#include</tt> hurts compile time performance. Don't do it unless you have
273 to, especially in header files.<p>
275 But wait, sometimes you need to have the definition of a class to use it, or to
276 inherit from it. In these cases go ahead and #include that header file. Be
277 aware however that there are many cases where you don't need to have the full
278 definition of a class. If you are using a pointer or reference to a class, you
279 don't need the header file. If you are simply returning a class instance from a
280 prototyped function or method, you don't need it. In fact, for most cases, you
281 simply don't need the definition of a class... and not <tt>#include</tt>'ing
282 speeds up compilation.<p>
284 It is easy to try to go too overboard on this recommendation, however. You
285 <b>must</b> include all of the header files that you are using, either directly
286 or indirectly (through another header file). To make sure that you don't
287 accidently forget to include a header file in your module header, make sure to
288 include your module header <b>first</b> in the implementation file (as mentioned
289 above). This way there won't be any hidden dependencies that you'll find out
293 <!-- _______________________________________________________________________ -->
294 </ul><a name="hl_privateheaders"><h4><hr size=0>Keep "internal" Headers Private</h4><ul>
296 Many modules have a complex implementation that causes them to use more than one
297 implementation (<tt>.cpp</tt>) file. It is often tempting to put the internal
298 communication interface (helper classes, extra functions, etc) in the public
299 module header file. Don't do this. :)<p>
301 If you really need to do something like this, put a private header file in the
302 same directory as the source files, and include it locally. This ensures that
303 your private interface remains private and undisturbed by outsiders.<p>
305 Note however, that it's okay to put extra implementation methods a public class
306 itself... just make them private (or protected), and all is well.<p>
309 <!-- ======================================================================= -->
310 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
311 <a name="micro">The Low Level Issues
312 </b></font></td></tr></table><ul>
315 <!-- _______________________________________________________________________ -->
316 </ul><a name="hl_assert"><h4><hr size=0>Assert Liberally</h4><ul>
318 Use the "<tt>assert</tt>" function to its fullest. Check all of your
319 preconditions and assumptions, you never know when a bug (not neccesarily even
320 yours) might be caught early by an assertion, which reduces debugging time
321 dramatically. The "<tt><cassert></tt>" header file is probably already
322 included by the header files you are using, so it doesn't cost anything to use
325 To further assist with debugging, make sure to put some kind of error message in
326 the assertion statement (which is printed if the assertion is tripped). This
327 helps the poor debugging make sense of why an assertion is being made and
328 enforced, and hopefully what to do about it. Here is one complete example:<p>
331 inline Value *getOperand(unsigned i) {
332 assert(i < Operands.size() && "getOperand() out of range!");
337 Here are some examples:
340 assert(Ty->isPointerType() && "Can't allocate a non pointer type!");
342 assert((Opcode == Shl || Opcode == Shr) && "ShiftInst Opcode invalid!");
344 assert(idx < getNumSuccessors() && "Successor # out of range!");
346 assert(V1.getType() == V2.getType() && "Constant types must be identical!");
348 assert(isa<PHINode>(Succ->front()) && "Only works on PHId BBs!");
351 You get the idea...<p>
354 <!-- _______________________________________________________________________ -->
355 </ul><a name="hl_preincrement"><h4><hr size=0>Prefer Preincrement</h4><ul>
357 Hard fast rule: Preincrement (++X) may be no slower than postincrement (X++) and
358 could very well be a lot faster than it. Use preincrementation whenever
361 The semantics of postincrement include making a copy of the value being
362 incremented, returning it, and then preincrementing the "work value". For
363 primitive types, this isn't a big deal... but for iterators, it can be a huge
364 issue (for example, some iterators contains stack and set objects in them...
365 copying an iterator could invoke the copy ctor's of these as well). In general,
366 get in the habit of always using preincrement, and you won't have a problem.<p>
369 <!-- _______________________________________________________________________ -->
370 </ul><a name="hl_avoidendl"><h4><hr size=0>Avoid endl</h4><ul>
372 The <tt>endl</tt> modifier, when used with iostreams outputs a newline to the
373 output stream specified. In addition to doing this, however, it also flushes
374 the output stream. In other words, these are equivalent:<p>
378 cout << "\n" << flush;
381 Most of the time, you probably have no reason to flush the output stream, so it's better to use a literal <tt>"\n"</tt>.<p>
384 <!-- _______________________________________________________________________ -->
385 </ul><a name="hl_exploitcpp"><h4><hr size=0>Exploit C++ to its Fullest</h4><ul>
387 C++ is a powerful language. With a firm grasp on its capabilities, you can make
388 write effective, consise, readable and maintainable code all at the same time.
389 By staying consistent, you reduce the amount of special cases that need to be
390 remembered. Reducing the total number of lines of code you write is a good way
391 to avoid documentation, and avoid giving bugs a place to hide.<p>
393 For these reasons, come to know and love the contents of your local
394 <algorithm> header file. Know about <functional> and what it can do
395 for you. C++ is just a tool that wants you to master it. :)<p>
399 <!-- ======================================================================= -->
400 </ul><table width="100%" bgcolor="#441188" border=0 cellpadding=4 cellspacing=0><tr><td> </td><td width="100%"> <font color="#EEEEFF" face="Georgia,Palatino"><b>
401 <a name="iterators">Writing Iterators
402 </b></font></td></tr></table><ul>
404 Here's a pretty good summary of how to write your own data structure iterators in a way that is compatible with the STL, and with a lot of other code out there (slightly edited by Chris):<p>
407 From: Ross Smith <ross.s@ihug.co.nz>
408 Newsgroups: comp.lang.c++.moderated
409 Subject: Writing iterators (was: Re: Non-template functions that take iterators)
410 Date: 28 Jun 2001 12:07:10 -0400
413 > Any pointers handy on "writing STL-compatible iterators for
416 I'll give it a try...
418 The usual situation requiring user-defined iterators is that you have
419 a type that bears some resemblance to an STL container, and you want
420 to provide iterators so it can be used with STL algorithms. You need
421 to ask three questions:
423 First, is this simply a wrapper for an underlying collection of
424 objects that's held somewhere as a real STL container, or is it a
425 "virtual container" for which iteration is (under the hood) more
426 complicated than simply incrementing some underlying iterator (or
427 pointer or index or whatever)? In the former case you can frequently
428 get away with making your container's iterators simply typedefs for
429 those of the underlying container; your begin() function would call
430 member_container.begin(), and so on.
432 Second, do you only need read-only iterators, or do you need separate
433 read-only (const) and read-write (non-const) iterators?
435 Third, which kind of iterator (input, output, forward, bidirectional,
436 or random access) is appropriate? If you're familiar with the
437 properties of the iterator types (if not, visit
438 <a href="http://www.sgi.com/tech/stl/">http://www.sgi.com/tech/stl/</a>), the appropriate choice should be
439 obvious from the semantics of the container.
441 I'll start with forward iterators, as the simplest case that's likely
442 to come up in normal code. Input and output iterators have some odd
443 properties and rarely need to be implemented in user code; I'll leave
444 them out of discussion. Bidirectional and random access iterators are
447 The exact behaviour of a forward iterator is spelled out in the
448 Standard in terms of a set of expressions with specified behaviour,
449 rather than a set of member functions, which leaves some leeway in how
450 you actually implement it. Typically it looks something like this
451 (I'll start with the const-iterator-only situation):
457 typedef something_or_other value_type;
458 class const_iterator:
459 public std::iterator<std::forward_iterator_tag, value_type> {
460 friend class container;
462 const value_type& operator*() const;
463 const value_type* operator->() const;
464 const_iterator& operator++();
465 const_iterator operator++(int);
466 friend bool operator==(const_iterator lhs,
468 friend bool operator!=(const_iterator lhs,
476 An iterator should always be derived from an instantiation of the
477 std::iterator template. The iterator's life cycle functions
478 (constructors, destructor, and assignment operator) aren't declared
479 here; in most cases the compiler-generated ones are sufficient. The
480 container needs to be a friend of the iterator so that the container's
481 begin() and end() functions can fill in the iterator's private members
482 with the appropriate values.
484 <i>[Chris's Note: I prefer to not make my iterators friends. Instead, two
485 ctor's are provided for the iterator class: one to start at the end of the
486 container, and one at the beginning. Typically this is done by providing
487 two constructors with different signatures.]</i>
489 There are normally only three member functions that need nontrivial
490 implementations; the rest are just boilerplate.
492 const container::value_type&
493 container::const_iterator::operator*() const {
494 // find the element and return a reference to it
497 const container::value_type*
498 container::const_iterator::operator->() const {
502 If there's an underlying real container, operator*() can just return a
503 reference to the appropriate element. If there's no actual container
504 and the elements need to be generated on the fly -- what I think of as
505 a "virtual container" -- things get a bit more complicated; you'll
506 probably need to give the iterator a value_type member object, and
507 fill it in when you need to. This might be done as part of the
508 increment operator (below), or if the operation is nontrivial, you
509 might choose the "lazy" approach and only generate the actual value
510 when one of the dereferencing operators is called.
512 The operator->() function is just boilerplate around a call to
515 container::const_iterator&
516 container::const_iterator::operator++() {
517 // the incrementing logic goes here
521 container::const_iterator
522 container::const_iterator::operator++(int) {
523 const_iterator old(*this);
528 Again, the incrementing logic will usually be trivial if there's a
529 real container involved, more complicated if you're working with a
530 virtual container. In particular, watch out for what happens when you
531 increment past the last valid item -- this needs to produce an
532 iterator that will compare equal to container.end(), and making this
533 work is often nontrivial for virtual containers.
535 The post-increment function is just boilerplate again (and
536 incidentally makes it obvious why all the experts recommend using
537 pre-increment wherever possible).
539 bool operator==(container::const_iterator lhs,
540 container::const_iterator rhs) {
541 // equality comparison goes here
544 bool operator!=(container::const_iterator lhs,
545 container::const_iterator rhs) {
546 return !(lhs == rhs);
549 For a real container, the equality comparison will usually just
550 compare the underlying iterators (or pointers or indices or whatever).
551 The semantics of comparisons for virtual container iterators are often
552 tricky. Remember that iterator comparison only needs to be defined for
553 iterators into the same container, so you can often simplify things by
554 taking for granted that lhs and rhs both point into the same container
555 object. Again, the second function is just boilerplate.
557 It's a matter of taste whether iterator arguments are passed by value
558 or reference; I've shown tham passed by value to reduce clutter, but
559 if the iterator contains several data members, passing by reference
562 That convers the const-iterator-only situation. When we need separate
563 const and mutable iterators, one small complication is added beyond
564 the simple addition of a second class.
568 typedef something_or_other value_type;
569 class const_iterator;
571 public std::iterator<std::forward_iterator_tag, value_type> {
572 friend class container;
573 friend class container::const_iterator;
575 value_type& operator*() const;
576 value_type* operator->() const;
577 iterator& operator++();
578 iterator operator++(int);
579 friend bool operator==(iterator lhs, iterator rhs);
580 friend bool operator!=(iterator lhs, iterator rhs);
584 class const_iterator:
585 public std::iterator<std::forward_iterator_tag, value_type> {
586 friend class container;
589 const_iterator(const iterator& i);
590 const value_type& operator*() const;
591 const value_type* operator->() const;
592 const_iterator& operator++();
593 const_iterator operator++(int);
594 friend bool operator==(const_iterator lhs,
596 friend bool operator!=(const_iterator lhs,
604 There needs to be a conversion from iterator to const_iterator (so
605 that mixed-type operations, such as comparison between an iterator and
606 a const_iterator, will work). This is done here by giving
607 const_iterator a conversion constructor from iterator (equivalently,
608 we could have given iterator an operator const_iterator()), which
609 requires const_iterator to be a friend of iterator, so it can copy its
610 data members. (It also requires the addition of an explicit default
611 constructor to const_iterator, since the existence of another
612 user-defined constructor inhibits the compiler-defined one.)
614 Bidirectional iterators add just two member functions to forward
618 public std::iterator<std::bidirectional_iterator_tag, value_type> {
621 iterator& operator--();
622 iterator operator--(int);
626 I won't detail the implementations, they're obvious variations on
629 Random access iterators add several more member and friend functions:
632 public std::iterator<std::random_access_iterator_tag, value_type> {
635 iterator& operator+=(difference_type rhs);
636 iterator& operator-=(difference_type rhs);
637 friend iterator operator+(iterator lhs, difference_type rhs);
638 friend iterator operator+(difference_type lhs, iterator rhs);
639 friend iterator operator-(iterator lhs, difference_type rhs);
640 friend difference_type operator-(iterator lhs, iterator rhs);
641 friend bool operator<(iterator lhs, iterator rhs);
642 friend bool operator>(iterator lhs, iterator rhs);
643 friend bool operator<=(iterator lhs, iterator rhs);
644 friend bool operator>=(iterator lhs, iterator rhs);
648 container::iterator&
649 container::iterator::operator+=(container::difference_type rhs) {
650 // add rhs to iterator position
654 container::iterator&
655 container::iterator::operator-=(container::difference_type rhs) {
656 // subtract rhs from iterator position
660 container::iterator operator+(container::iterator lhs,
661 container::difference_type rhs) {
662 return iterator(lhs) += rhs;
665 container::iterator operator+(container::difference_type lhs,
666 container::iterator rhs) {
667 return iterator(rhs) += lhs;
670 container::iterator operator-(container::iterator lhs,
671 container::difference_type rhs) {
672 return iterator(lhs) -= rhs;
675 container::difference_type operator-(container::iterator lhs,
676 container::iterator rhs) {
677 // calculate distance between iterators
680 bool operator<(container::iterator lhs, container::iterator rhs) {
681 // perform less-than comparison
684 bool operator>(container::iterator lhs, container::iterator rhs) {
688 bool operator<=(container::iterator lhs, container::iterator rhs) {
692 bool operator>=(container::iterator lhs, container::iterator rhs) {
696 Four of the functions (operator+=(), operator-=(), the second
697 operator-(), and operator<()) are nontrivial; the rest are
700 One feature of the above code that some experts may disapprove of is
701 the declaration of all the free functions as friends, when in fact
702 only a few of them need direct access to the iterator's private data.
703 I originally got into the habit of doing this simply to keep the
704 declarations together; declaring some functions inside the class and
705 some outside seemed awkward. Since then, though, I've been told that
706 there's a subtle difference in the way name lookup works for functions
707 declared inside a class (as friends) and outside, so keeping them
708 together in the class is probably a good idea for practical as well as
711 I hope all this is some help to anyone who needs to write their own
712 STL-like containers and iterators.
715 Ross Smith <ross.s@ihug.co.nz> The Internet Group, Auckland, New Zealand
719 <!-- *********************************************************************** -->
720 </ul><table width="100%" bgcolor="#330077" border=0 cellpadding=4 cellspacing=0><tr><td align=center><font color="#EEEEFF" size=+2 face="Georgia,Palatino"><b>
721 <a name="seealso">See Also
722 </b></font></td></tr></table><ul>
723 <!-- *********************************************************************** -->
725 A lot of these comments and recommendations have been culled for other sources. Two particularly important books for our work are:<p>
728 <li><a href="http://www.aw.com/product/0,2627,0201924889,00.html">Effective C++</a> by Scott Meyers. There is an online version of the book (only some chapters though) <a href="http://www.awlonline.com/cseng/meyerscddemo/">available as well</a>.
729 <li><a href="http://cseng.aw.com/book/0,3828,0201633620,00.html">Large-Scale C++ Software Design</a> by John Lakos
732 If you get some free time, and you haven't read them: do so, you might learn something. :)
735 <!-- *********************************************************************** -->
737 <!-- *********************************************************************** -->
741 <address><a href="mailto:sabre@nondot.org">Chris Lattner</a></address>
742 <!-- Created: Tue Jan 23 15:19:28 CST 2001 -->
744 Last modified: Fri Jul 25 12:29:52 CDT 2003