<?xml version="1.0" encoding="ISO-8859-1"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> <head> <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> <meta name="AUTHOR" content="pme@gcc.gnu.org (Phil Edwards)" /> <meta name="KEYWORDS" content="HOWTO, libstdc++, GCC, g++, libg++, STL" /> <meta name="DESCRIPTION" content="Notes for the libstdc++ extensions." /> <meta name="GENERATOR" content="vi and eight fingers" /> <title>libstdc++-v3 HOWTO: Extensions</title> <link rel="StyleSheet" href="../lib3styles.css" type="text/css" /> <link rel="Start" href="../documentation.html" type="text/html" title="GNU C++ Standard Library" /> <link rel="Prev" href="../27_io/howto.html" type="text/html" title="Input/Output" /> <link rel="Bookmark" href="sgiexts.html" type="text/html" title="SGI extensions" /> <link rel="Bookmark" href="mt_allocator.html" type="text/html" title="__mt_alloc" /> <link rel="Copyright" href="../17_intro/license.html" type="text/html" /> </head> <body> <h1 class="centered"><a name="top">Extensions</a></h1> <p>Here we will make an attempt at describing the non-Standard extensions to the library. Some of these are from SGI's STL, some of these are GNU's, and some just seemed to appear on the doorstep. </p> <p><strong>Before you leap in and use these</strong>, be aware of two things: </p> <ol> <li>Non-Standard means exactly that. The behavior, and the very existence, of these extensions may change with little or no warning. (Ideally, the really good ones will appear in the next revision of C++.) Also, other platforms, other compilers, other versions of g++ or libstdc++-v3 may not recognize these names, or treat them differently, or... </li> <li>You should know how to <a href="../faq/index.html#5_4">access these headers properly</a>. </li> </ol> <!-- ####################################################### --> <hr /> <h1>Contents</h1> <ul> <li><a href="#1">Ropes and trees and hashes, oh my!</a></li> <li><a href="#2">Added members and types</a></li> <li><a href="mt_allocator.html"><code>__mt_alloc</code> </a></li> <li><a href="#4">Compile-time checks</a></li> <li><a href="#5">LWG Issues</a></li> <li><a href="../18_support/howto.html#6">Demangling</a></li> </ul> <hr /> <!-- ####################################################### --> <h2><a name="1">Ropes and trees and hashes, oh my!</a></h2> <p>The SGI headers</p> <pre> <bvector> <hash_map> <hash_set> <rope> <slist> <tree> </pre> <p>are all here; <code><bvector></code> exposes the old bit_vector class that was used before specialization of vector<bool> was available (it's actually a typedef for the specialization now). <code><hash_map></code> and <code><hash_set></code> are discussed further below. <code><rope></code> is the SGI specialization for large strings ("rope," "large strings," get it? love those SGI folks). <code><slist></code> is a singly-linked list, for when the doubly-linked <code>list<></code> is too much space overhead, and <code><tree></code> exposes the red-black tree classes used in the implementation of the standard maps and sets. </p> <p>Okay, about those hashing classes... I'm going to foist most of the work off onto SGI's own site. </p> <p>Each of the associative containers map, multimap, set, and multiset have a counterpart which uses a <a href="http://www.sgi.com/tech/stl/HashFunction.html">hashing function</a> to do the arranging, instead of a strict weak ordering function. The classes take as one of their template parameters a function object that will return the hash value; by default, an instantiation of <a href="http://www.sgi.com/tech/stl/hash.html">hash</a>. You should specialize this functor for your class, or define your own, before trying to use one of the hashing classes. </p> <p>The hashing classes support all the usual associative container functions, as well as some extra constructors specifying the number of buckets, etc. </p> <p>Why would you want to use a hashing class instead of the "normal" implementations? Matt Austern writes: </p> <blockquote><em>[W]ith a well chosen hash function, hash tables generally provide much better average-case performance than binary search trees, and much worse worst-case performance. So if your implementation has hash_map, if you don't mind using nonstandard components, and if you aren't scared about the possibility of pathological cases, you'll probably get better performance from hash_map.</em></blockquote> <p>(Side note: for those of you wondering, <strong>"Why wasn't a hash table included in the Standard in the first #!$@ place?"</strong> I'll give a quick answer: it was proposed, but too late and in too unorganized a fashion. Some sort of hashing will undoubtedly be included in a future Standard.) </p> <p>Return <a href="#top">to top of page</a> or <a href="../faq/index.html">to the FAQ</a>. </p> <hr /> <h2><a name="2">Added members and types</a></h2> <p>Some of the classes in the Standard Library have additional publicly-available members, and some classes are themselves not in the standard. Of those, some are intended purely for the implementors, for example, additional typedefs. Those won't be described here (or anywhere else). </p> <ul> <li>The extensions added by SGI are so numerous that they have <a href="sgiexts.html">their own page</a>. Since the SGI STL is no longer actively maintained, we will try and keep this code working ourselves.</li> <li>Extensions allowing <code>filebuf</code>s to be constructed from stdio types are described in the <a href="../27_io/howto.html#11">chapter 27 notes</a>.</li> </ul> <p>Return <a href="#top">to top of page</a> or <a href="../faq/index.html">to the FAQ</a>. </p> <hr /> <h2><a name="4">Compile-time checks</a></h2> <p>Currently libstdc++-v3 uses the concept checkers from the Boost library to perform <a href="../19_diagnostics/howto.html#3">optional compile-time checking</a> of template instantiations of the standard containers. They are described in the linked-to page. </p> <p>Return <a href="#top">to top of page</a> or <a href="../faq/index.html">to the FAQ</a>. </p> <hr /> <h2><a name="5">LWG Issues</a></h2> <p>Everybody's got issues. Even the C++ Standard Library. </p> <p>The Library Working Group, or LWG, is the ISO subcommittee responsible for making changes to the library. They periodically publish an Issues List containing problems and possible solutions. As they reach a consensus on proposed solutions, we often incorporate the solution into libstdc++-v3. </p> <p>Here are the issues which have resulted in code changes to the library. The links are to the specific defect reports from a <strong>partial copy</strong> of the Issues List. You can read the full version online at the <a href="http://www.open-std.org/jtc1/sc22/wg21/">ISO C++ Committee homepage</a>, linked to on the <a href="http://gcc.gnu.org/readings.html">GCC "Readings" page</a>. If you spend a lot of time reading the issues, we recommend downloading the ZIP file and reading them locally. </p> <p>(NB: <strong>partial copy</strong> means that not all links within the lwg-*.html pages will work. Specifically, links to defect reports that have not been accorded full DR status will probably break. Rather than trying to mirror the entire issues list on our overworked web server, we recommend you go to the LWG homepage instead.) </p> <p> If a DR is not listed here, we may simply not have gotten to it yet; feel free to submit a patch. Search the include/bits and src directories for appearances of _GLIBCXX_RESOLVE_LIB_DEFECTS for examples of style. Note that we usually do not make changes to the code until an issue has reached <a href="lwg-active.html#DR">DR</a> status. </p> <dl> <dt><a href="lwg-defects.html#5">5</a>: <em>string::compare specification questionable</em> </dt> <dd>This should be two overloaded functions rather than a single function. </dd> <dt><a href="lwg-defects.html#17">17</a>: <em>Bad bool parsing</em> </dt> <dd>Apparently extracting Boolean values was messed up... </dd> <dt><a href="lwg-defects.html#19">19</a>: <em>"Noconv" definition too vague</em> </dt> <dd>If <code>codecvt::do_in</code> returns <code>noconv</code> there are no changes to the values in <code>[to, to_limit)</code>. </dd> <dt><a href="lwg-defects.html#22">22</a>: <em>Member open vs flags</em> </dt> <dd>Re-opening a file stream does <em>not</em> clear the state flags. </dd> <dt><a href="lwg-defects.html#25">25</a>: <em>String operator<< uses width() value wrong</em> </dt> <dd>Padding issues. </dd> <dt><a href="lwg-defects.html#48">48</a>: <em>Use of non-existent exception constructor</em> </dt> <dd>An instance of <code>ios_base::failure</code> is constructed instead. </dd> <dt><a href="lwg-defects.html#49">49</a>: <em>Underspecification of ios_base::sync_with_stdio</em> </dt> <dd>The return type is the <em>previous</em> state of synchronization. </dd> <dt><a href="lwg-defects.html#50">50</a>: <em>Copy constructor and assignment operator of ios_base</em> </dt> <dd>These members functions are declared <code>private</code> and are thus inaccessible. Specifying the correct semantics of "copying stream state" was deemed too complicated. </dd> <dt><a href="lwg-defects.html#60">60</a>: <em>What is a formatted input function?</em> </dt> <dd>This DR made many widespread changes to <code>basic_istream</code> and <code>basic_ostream</code> all of which have been implemented. </dd> <dt><a href="lwg-defects.html#63">63</a>: <em>Exception-handling policy for unformatted output</em> </dt> <dd>Make the policy consistent with that of formatted input, unformatted input, and formatted output. </dd> <dt><a href="lwg-defects.html#68">68</a>: <em>Extractors for char* should store null at end</em> </dt> <dd>And they do now. An editing glitch in the last item in the list of [27.6.1.2.3]/7. </dd> <dt><a href="lwg-defects.html#74">74</a>: <em>Garbled text for codecvt::do_max_length</em> </dt> <dd>The text of the standard was gibberish. Typos gone rampant. </dd> <dt><a href="lwg-defects.html#75">75</a>: <em>Contradiction in codecvt::length's argument types</em> </dt> <dd>Change the first parameter to <code>stateT&</code> and implement the new effects paragraph. </dd> <dt><a href="lwg-defects.html#83">83</a>: <em>string::npos vs. string::max_size()</em> </dt> <dd>Safety checks on the size of the string should test against <code>max_size()</code> rather than <code>npos</code>. </dd> <dt><a href="lwg-defects.html#90">90</a>: <em>Incorrect description of operator>> for strings</em> </dt> <dd>The effect contain <code>isspace(c,getloc())</code> which must be replaced by <code>isspace(c,is.getloc())</code>. </dd> <dt><a href="lwg-defects.html#91">91</a>: <em>Description of operator>> and getline() for string<> might cause endless loop</em> </dt> <dd>They behave as a formatted input function and as an unformatted input function, respectively (except that <code>getline</code> is not required to set <code>gcount</code>). </dd> <dt><a href="lwg-defects.html#103">103</a>: <em>set::iterator is required to be modifiable, but this allows modification of keys.</em> </dt> <dd>For associative containers where the value type is the same as the key type, both <code>iterator</code> and <code>const_iterator </code> are constant iterators. </dd> <dt><a href="lwg-defects.html#109">109</a>: <em>Missing binders for non-const sequence elements</em> </dt> <dd>The <code>binder1st</code> and <code>binder2nd</code> didn't have an <code>operator()</code> taking a non-const parameter. </dd> <dt><a href="lwg-defects.html#110">110</a>: <em>istreambuf_iterator::equal not const</em> </dt> <dd>This was not a const member function. Note that the DR says to replace the function with a const one; we have instead provided an overloaded version with identical contents. </dd> <dt><a href="lwg-defects.html#117">117</a>: <em>basic_ostream uses nonexistent num_put member functions</em> </dt> <dd><code>num_put::put()</code> was overloaded on the wrong types. </dd> <dt><a href="lwg-defects.html#118">118</a>: <em>basic_istream uses nonexistent num_get member functions</em> </dt> <dd>Same as 117, but for <code>num_get::get()</code>. </dd> <dt><a href="lwg-defects.html#129">129</a>: <em>Need error indication from seekp() and seekg()</em> </dt> <dd>These functions set <code>failbit</code> on error now. </dd> <dt><a href="lwg-defects.html#136">136</a>: <em>seekp, seekg setting wrong streams?</em> </dt> <dd><code>seekp</code> should only set the output stream, and <code>seekg</code> should only set the input stream. </dd> <!--<dt><a href="lwg-defects.html#159">159</a>: <em>Strange use of underflow()</em> </dt> <dd>In fstream.tcc, the basic_filebuf<>::showmanyc() function should probably not be calling <code>underflow()</code>. </dd> --> <dt><a href="lwg-defects.html#167">167</a>: <em>Improper use of traits_type::length()</em> </dt> <dd><code>op<<</code> with a <code>const char*</code> was calculating an incorrect number of characters to write. </dd> <dt><a href="lwg-defects.html#171">171</a>: <em>Strange seekpos() semantics due to joint position</em> </dt> <dd>Quite complex to summarize... </dd> <dt><a href="lwg-defects.html#181">181</a>: <em>make_pair() unintended behavior</em> </dt> <dd>This function used to take its arguments as reference-to-const, now it copies them (pass by value). </dd> <dt><a href="lwg-defects.html#195">195</a>: <em>Should basic_istream::sentry's constructor ever set eofbit?</em> </dt> <dd>Yes, it can, specifically if EOF is reached while skipping whitespace. </dd> <dt><a href="lwg-defects.html#211">211</a>: <em>operator>>(istream&, string&) doesn't set failbit</em> </dt> <dd>If nothing is extracted into the string, <code>op>></code> now sets <code>failbit</code> (which can cause an exception, etc, etc). </dd> <dt><a href="lwg-defects.html#214">214</a>: <em>set::find() missing const overload</em> </dt> <dd>Both <code>set</code> and <code>multiset</code> were missing overloaded find, lower_bound, upper_bound, and equal_range functions for const instances. </dd> <dt><a href="lwg-defects.html#231">231</a>: <em>Precision in iostream?</em> </dt> <dd>For conversion from a floating-point type, <code>str.precision()</code> is specified in the conversion specification. </dd> <dt><a href="lwg-defects.html#235">235</a>: <em>No specification of default ctor for reverse_iterator</em> </dt> <dd>The declaration of <code>reverse_iterator</code> lists a default constructor. However, no specification is given what this constructor should do. </dd> <dt><a href="lwg-defects.html#243">243</a>: <em>get and getline when sentry reports failure</em> </dt> <dd>Store a null character only if the character array has a non-zero size. </dd> <dt><a href="lwg-defects.html#251">251</a>: <em>basic_stringbuf missing allocator_type</em> </dt> <dd>This nested typedef was originally not specified. </dd> <dt><a href="lwg-defects.html#253">253</a>: <em>valarray helper functions are almost entirely useless</em> </dt> <dd>Make the copy constructor and copy-assignment operator declarations public in gslice_array, indirect_array, mask_array, slice_array; provide definitions. </dd> <dt><a href="lwg-defects.html#265">265</a>: <em>std::pair::pair() effects overly restrictive</em> </dt> <dd>The default ctor would build its members from copies of temporaries; now it simply uses their respective default ctors. </dd> <dt><a href="lwg-defects.html#266">266</a>: <em>bad_exception::~bad_exception() missing Effects clause</em> </dt> <dd>The <code>bad_</code>* classes no longer have destructors (they are trivial), since no description of them was ever given. </dd> <dt><a href="lwg-defects.html#271">271</a>: <em>basic_iostream missing typedefs</em> </dt> <dd>The typedefs it inherits from its base classes can't be used, since (for example) <code>basic_iostream<T>::traits_type</code> is ambiguous. </dd> <dt><a href="lwg-defects.html#275">275</a>: <em>Wrong type in num_get::get() overloads</em> </dt> <dd>Similar to 118. </dd> <dt><a href="lwg-defects.html#292">292</a>: <em>Effects of a.copyfmt (a)</em> </dt> <dd>If <code>(this == &rhs)</code> do nothing. </dd> <dt><a href="lwg-defects.html#300">300</a>: <em>List::merge() specification incomplete</em> </dt> <dd>If <code>(this == &x)</code> do nothing. </dd> <dt><a href="lwg-defects.html#303">303</a>: <em>Bitset input operator underspecified</em> </dt> <dd>Basically, compare the input character to <code>is.widen(0)</code> and <code>is.widen(1)</code>. </dd> <dt><a href="lwg-defects.html#305">305</a>: <em>Default behavior of codecvt<wchar_t, char, mbstate_t>::length()</em> </dt> <dd>Do not specify what <code>codecvt<wchar_t, char, mbstate_t>::do_length</code> must return. </dd> <dt><a href="lwg-defects.html#328">328</a>: <em>Bad sprintf format modifier in money_put<>::do_put()</em> </dt> <dd>Change the format string to "%.0Lf". </dd> <dt><a href="lwg-defects.html#365">365</a>: <em>Lack of const-qualification in clause 27</em> </dt> <dd>Add const overloads of <code>is_open</code>. </dd> <dt><a href="lwg-defects.html#389">389</a>: <em>Const overload of valarray::operator[] returns by value</em> </dt> <dd>Change it to return a <code>const T&</code>. </dd> <dt><a href="lwg-defects.html#402">402</a>: <em>Wrong new expression in [some_]allocator::construct</em> </dt> <dd>Replace "new" with "::new". </dd> <dt><a href="lwg-active.html#409">409</a>: <em>Closing an fstream should clear the error state</em> </dt> <dd>Have <code>open</code> clear the error flags. </dd> <dt><a href="lwg-active.html#434">434</a>: <em>bitset::to_string() hard to use</em> </dt> <dd>Add three overloads, taking fewer template arguments. </dd> <dt><a href="lwg-active.html#453">453</a>: <em>basic_stringbuf::seekoff need not always fail for an empty stream</em> </dt> <dd>Don't fail if the next pointer is null and newoff is zero. </dd> <!-- <dt><a href="lwg-defects.html#"></a>: <em></em> </dt> <dd> </dd> --> </dl> <p>Return <a href="#top">to top of page</a> or <a href="../faq/index.html">to the FAQ</a>. </p> <!-- ####################################################### --> <hr /> <p class="fineprint"><em> See <a href="../17_intro/license.html">license.html</a> for copying conditions. Comments and suggestions are welcome, and may be sent to <a href="mailto:libstdc++@gcc.gnu.org">the libstdc++ mailing list</a>. </em></p> </body> </html>