feedc0de/boost_array
1
0
Fork 0
forked from boostorg/array
Code Pull Requests Activity

array.xml: documentation updates from Nico

Browse Source 
[SVN r22066]
This commit is contained in:
Douglas Gregor
2004-01-30 03:51:06 +00:00
parent 0543286830
commit cfb6d6eae6
1 changed files with 31 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

70
doc/array.xml
View File

@ -66,9 +66,13 @@
<para>After considering different names, we decided to name this
class simply <code><classname>array</classname></code>.</para>
<para>Class <code><classname>array</classname></code> fulfills
most but not all of the requirements of "reversible containers"
(see Section 23.1, [lib.container.requirements] of the C++
<para>Note that this class is suggested to be part of the next
Technical Report, which will extend the C++ Standard (see
<ulink>http://std.dkuug.dk/jtc1/sc22/wg21/docs/papers/2003/n1548.htm</ulink>).</para>
<para>Class <code><classname>array</classname></code> fulfills most
but not all of the requirements of "reversible containers" (see
Section 23.1, [lib.container.requirements] of the C++
Standard). The reasons array is not an reversible STL container is
because:
<itemizedlist spacing="compact">
@ -478,51 +482,39 @@
<simpara>Note that if there are fewer elements in the
initializer list, then each remaining element gets
default-initialized (thus, it has a defined value). However,
passing no initializer list means that the elements have an
indetermined initial value.</simpara></listitem>
default-initialized (thus, it has a defined value).</simpara>
</listitem></itemizedlist></para>
<listitem><simpara>It has no user-declared constructors.</simpara></listitem>
<listitem><simpara>It has no private or protected non-static data members.</simpara></listitem>
<listitem><simpara>It has no base classes.</simpara></listitem>
<listitem><simpara>It has no virtual functions.</simpara></listitem>
<para>However, this approach has its drawbacks: <emphasis
role="bold"> passing no initializer list means that the elements
have an indetermined initial value</emphasis>, because the rule says
that aggregates may have:
<itemizedlist>
<listitem><simpara>No user-declared constructors.</simpara></listitem>
<listitem><simpara>No private or protected non-static data members.</simpara></listitem>
<listitem><simpara>No base classes.</simpara></listitem>
<listitem><simpara>No virtual functions.</simpara></listitem>
</itemizedlist>
</para>
<para>The current implementation uses this approach. However, being
able to have indeterminate initial values is a big drawback. So,
please give me some feedback, how useful you consider this feature
to be.</para>
</section>
<section id="array.issues">
<title>Open Issues</title>
<para>Nevertheless, The current implementation uses this approach.</para>
<itemizedlist>
<para>Note that for standard conforming compilers it is possible to
use fewer braces (according to 8.5.1 (11) of the Standard). That is,
you can initialize an array as follows:</para>
<listitem>
<simpara> Do we want initializer list support or would the
following be OK?:</simpara>
<programlisting>
<classname>boost::array</classname>&lt;int,4&gt; a = { 1, 2, 3 };
</programlisting>
<programlisting>int data[] = { 1, 2, 3, 4 }
<classname>boost::array</classname>&lt;int,5&gt; x(data); // or <classname>boost::array</classname>&lt;int,data&gt; x;</programlisting>
</listitem>
<para>I'd appreciate any constructive feedback. <emphasis
role="bold">Please note: I don't have time to read all boost
mails. Thus, to make sure that feedback arrives to me, please send
me a copy of each mail regarding this class.</emphasis></para>
<listitem><simpara>Could "{ ... }" be used portably instead of "{
{ ... } }" to initialize values? 8.5.1 (11) of the Standard seems
to allow it; however, gcc 2.95.2 prints a warning
message.</simpara></listitem>
<para>The code is provided "as is" without expressed or implied
warranty.</para>
<listitem><simpara>Any way to have determinate initial values and
initializer list support?</simpara></listitem>
<listitem><simpara>Static_casts for reverse iterator stuff?</simpara></listitem>
</itemizedlist>
<para>I'd appreciate any constructive feedback. Please note: I don't
have time to read all boost mails. Thus, to make sure that <ulink
url="mailto:solutions@josuttis.com">feedback</ulink> arrives me,
please send me a copy of each mail regarding this class.</para>
</section>
<section id="array.more.info">