| 1 | <?xml version="1.0" encoding="utf-8" ?> |
|---|
| 2 | <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> |
|---|
| 3 | <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> |
|---|
| 4 | <!-- Copyright Aleksey Gurtovoy 2006. Distributed under the Boost --> |
|---|
| 5 | <!-- Software License, Version 1.0. (See accompanying --> |
|---|
| 6 | <!-- file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) --> |
|---|
| 7 | <head> |
|---|
| 8 | <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> |
|---|
| 9 | <meta name="generator" content="Docutils 0.3.6: http://docutils.sourceforge.net/" /> |
|---|
| 10 | <title>THE BOOST MPL LIBRARY: Broken Integral Constant Expressions</title> |
|---|
| 11 | <link rel="stylesheet" href="../style.css" type="text/css" /> |
|---|
| 12 | </head> |
|---|
| 13 | <body class="docframe"> |
|---|
| 14 | <table class="header"><tr class="header"><td class="header-group navigation-bar"><span class="navigation-group"><a href="./portability.html" class="navigation-link">Prev</a> <a href="./incomplete-support-for.html" class="navigation-link">Next</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group">Back <a href="./incomplete-support-for.html" class="navigation-link">Along</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group"><a href="./portability.html" class="navigation-link">Up</a> <a href="../index.html" class="navigation-link">Home</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group"><a href="./tutorial_toc.html" class="navigation-link">Full TOC</a></span></td> |
|---|
| 15 | <td class="header-group page-location"><a href="../index.html" class="navigation-link">Front Page</a> / <a href="./technical-details.html" class="navigation-link">Technical Details</a> / <a href="./portability.html" class="navigation-link">Portability</a> / <a href="./broken-integral-constant.html" class="navigation-link">Broken Integral Constant Expressions</a></td> |
|---|
| 16 | </tr></table><div class="header-separator"></div> |
|---|
| 17 | <div class="section" id="broken-integral-constant"> |
|---|
| 18 | <h1><a class="toc-backref" href="./portability.html#id74" name="broken-integral-constant">Broken Integral Constant Expressions</a></h1> |
|---|
| 19 | <p>This is probably the most surprising of the portability issues |
|---|
| 20 | we're going to discuss, not least because for many C++ programmers, their everyday |
|---|
| 21 | experience seems to indicate no problems in this area whatsoever. After all, |
|---|
| 22 | integer compile-time computations along the lines of:</p> |
|---|
| 23 | <pre class="literal-block"> |
|---|
| 24 | enum flags { |
|---|
| 25 | flag1 = (1 << 0) |
|---|
| 26 | , flag2 = (1 << 1) |
|---|
| 27 | , flag3 = (1 << 2) |
|---|
| 28 | ... |
|---|
| 29 | }; |
|---|
| 30 | </pre> |
|---|
| 31 | <p>are <em>very</em> commonplace in C++, and there is hardly a compiler out |
|---|
| 32 | there that cannot handle this correctly. While arithmetic by |
|---|
| 33 | itself is indeed rarely problematic, when you are trying to mix it |
|---|
| 34 | with templates on certain deficient compilers, all kinds of new |
|---|
| 35 | issues arise. Fortunately, as with the rest of the portability |
|---|
| 36 | issues we're discussing here, the problem fades into past as new |
|---|
| 37 | compiler versions are released. The majority of most recent |
|---|
| 38 | compilers of many vendors are already free from these issues.</p> |
|---|
| 39 | <div class="section" id="the-problem"> |
|---|
| 40 | <h2><a name="the-problem">The Problem</a></h2> |
|---|
| 41 | <p>The problem is in fact multi-faceted; there are a number of |
|---|
| 42 | different subissues. Some are present in one set of compilers, |
|---|
| 43 | some are in another, and it's not uncommon for a code that works |
|---|
| 44 | for one compiler to break another one and vice-versa. If this |
|---|
| 45 | sounds like a maintenance nightmare to you, it is! If you are |
|---|
| 46 | interested in the specific list of issues, please refer to John |
|---|
| 47 | Maddock's excellent "<a class="reference" href="../../../../more/int_const_guidelines.htm" target="_top">Coding Guidelines for Integral Constant |
|---|
| 48 | Expressions</a>" summary. For the purpose of our discission here, it |
|---|
| 49 | is sufficient to say that if your code has to work on one of the |
|---|
| 50 | compilers listed as problematic in this area, you can safely assume |
|---|
| 51 | that if you decide to fight them on a case-by-case basis, chances |
|---|
| 52 | are that you won't be able to maintain your sanity for very long.</p> |
|---|
| 53 | </div> |
|---|
| 54 | <div class="section" id="the-symptoms"> |
|---|
| 55 | <h2><a name="the-symptoms">The Symptoms</a></h2> |
|---|
| 56 | <p>On the positive side, when you have an issue with integral |
|---|
| 57 | arithmetic, the diagnostics are almost always straightforward: |
|---|
| 58 | usually the error message refers you to the exact place in the code |
|---|
| 59 | that is causing problems, and the essence of issue is obvious from |
|---|
| 60 | the error's text, or it becomes obvious once you've encountered the |
|---|
| 61 | same error a couple of times. For instance, if we throw this |
|---|
| 62 | well-formed fragment at MSVC 7.1 (otherwise an excellent compiler!)</p> |
|---|
| 63 | <pre class="literal-block"> |
|---|
| 64 | void value(); |
|---|
| 65 | |
|---|
| 66 | // compares two Integral Constants |
|---|
| 67 | template< typename N1, typename N2 > struct less |
|---|
| 68 | : bool_< (N1::value < N2::value) > // line #8 |
|---|
| 69 | { |
|---|
| 70 | }; |
|---|
| 71 | </pre> |
|---|
| 72 | <p>we get:</p> |
|---|
| 73 | <pre class="literal-block"> |
|---|
| 74 | portability.cpp(8) : warning C4346: 'N2::value' : dependent name is not a type |
|---|
| 75 | prefix with 'typename' to indicate a type |
|---|
| 76 | portability.cpp(10) : see reference to class template instantiation 'less<N1,N2>' being compiled |
|---|
| 77 | portability.cpp(8) : error C2143: syntax error : missing ',' before ')' |
|---|
| 78 | portability.cpp(9) : error C2143: syntax error : missing ';' before '{' |
|---|
| 79 | portability.cpp(10) : error C2143: syntax error : missing ';' before '}' |
|---|
| 80 | portability.cpp(10) : fatal error C1004: unexpected end of file found |
|---|
| 81 | </pre> |
|---|
| 82 | <p>The errors themselves are far from being ideal, but at least we are |
|---|
| 83 | pointed at the correct line and even the correct part of the |
|---|
| 84 | line. The above basically reads as "something's wrong between the |
|---|
| 85 | parentheses", and that plus the "syntax error" part is usually |
|---|
| 86 | enough of the clue.</p> |
|---|
| 87 | </div> |
|---|
| 88 | <div class="section" id="the-solution"> |
|---|
| 89 | <h2><a name="the-solution">The Solution</a></h2> |
|---|
| 90 | <p>Despite the fact the problems are so numerous and multi-faceted and |
|---|
| 91 | the workarounds are conflicting, the problems can be hidden |
|---|
| 92 | reliably beneath a library abstraction layer. The underlaying idea |
|---|
| 93 | is very simple: we can always wrap the constants in types and pass |
|---|
| 94 | those around. Then all that is left is to implement algebraic |
|---|
| 95 | metafunctions that operate on such wrappers, once, and we are home |
|---|
| 96 | safe.</p> |
|---|
| 97 | <p>If this sounds familiar to you, probably it's because you have |
|---|
| 98 | already took a look at the MPL and know that the approach we just |
|---|
| 99 | described is in fact <em>the</em> standard way of doing arithmetic in the |
|---|
| 100 | library. Although it's motivated by more general observations, |
|---|
| 101 | this fact comes very handy for the library users that care about |
|---|
| 102 | portability of their numerically-heavy metaprograms. The MPL |
|---|
| 103 | primitives are already there, and more importantly, they already |
|---|
| 104 | implement the necessary workarounds, so your numeric code just |
|---|
| 105 | works. In fact, if you stay within the library's type-wrapper |
|---|
| 106 | idioms, these particular problems never "leak" out of its |
|---|
| 107 | abstraction layer.</p> |
|---|
| 108 | <p>On a final note, there is a price of avoiding built-in arithmetics |
|---|
| 109 | altogether, namely decreased readability and, on some compilers, |
|---|
| 110 | increased compile-time overhead. Still, in majority of cases, the |
|---|
| 111 | benefits of type-based arithmetics overweight its small |
|---|
| 112 | shortcomings.</p> |
|---|
| 113 | </div> |
|---|
| 114 | </div> |
|---|
| 115 | |
|---|
| 116 | <div class="footer-separator"></div> |
|---|
| 117 | <table class="footer"><tr class="footer"><td class="header-group navigation-bar"><span class="navigation-group"><a href="./portability.html" class="navigation-link">Prev</a> <a href="./incomplete-support-for.html" class="navigation-link">Next</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group">Back <a href="./incomplete-support-for.html" class="navigation-link">Along</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group"><a href="./portability.html" class="navigation-link">Up</a> <a href="../index.html" class="navigation-link">Home</a></span><span class="navigation-group-separator"> | </span><span class="navigation-group"><a href="./tutorial_toc.html" class="navigation-link">Full TOC</a></span></td> |
|---|
| 118 | </tr></table></body> |
|---|
| 119 | </html> |
|---|
