libstdc++-v3/docs/html/22_locale/ctype.html

   1 <HTML>
   2 <HEAD>
   3   <H1>
   4   Notes on the ctype implementation.
   5   </H1>
   6 </HEAD>
   7 <I>
   8 prepared by Benjamin Kosnik (bkoz@redhat.com) on August 30, 2000
   9 </I>
  10
  11 <P>
  12 <H2>
  13 1. Abstract
  14 </H2>
  15 <P>
  16 Woe is me.
  17 </P>
  18
  19 <P>
  20 <H2>
  21 2. What the standard says
  22 </H2>
  23
  24
  25 <P>
  26 <H2>
  27 3. Problems with &quot;C&quot; ctype : global locales, termination.
  28 </H2>
  29
  30 <P>
  31 For the required specialization codecvt&lt;wchar_t, char, mbstate_t&gt; ,
  32 conversions are made between the internal character set (always UCS4
  33 on GNU/Linux) and whatever the currently selected locale for the
  34 LC_CTYPE category implements.
  35
  36 <P>
  37 <H2>
  38 4. Design
  39 </H2>
  40 The two required specializations are implemented as follows:
  41
  42 <P>
  43 <TT>
  44 ctype&lt;char&gt;
  45 </TT>
  46 <P>
  47 This is simple specialization. Implementing this was a piece of cake.
  48
  49 <P>
  50 <TT>
  51 ctype&lt;wchar_t&gt;
  52 </TT>
  53 <P>
  54 This specialization, by specifying all the template parameters, pretty
  55 much ties the hands of implementors. As such, the implementation is
  56 straightforward, involving mcsrtombs for the conversions between char
  57 to wchar_t and wcsrtombs for conversions between wchar_t and char.
  58
  59 <P>
  60 Neither of these two required specializations deals with Unicode
  61 characters. As such, libstdc++-v3 implements
  62
  63
  64
  65 <P>
  66 <H2>
  67 5.  Examples
  68 </H2>
  69
  70 <pre>
  71   typedef ctype<char> cctype;
  72 </pre>
  73
  74 More information can be found in the following testcases:
  75 <UL>
  76 <LI> testsuite/22_locale/ctype_char_members.cc
  77 <LI> testsuite/22_locale/ctype_wchar_t_members.cc
  78 </UL>
  79
  80 <P>
  81 <H2>
  82 6.  Unresolved Issues
  83 </H2>
  84
  85 <UL>
  86         <LI> how to deal with the global locale issue?
  87
  88         <LI> how to deal with different types than char, wchar_t?
  89
  90         <LI> codecvt/ctype overlap: narrow/widen
  91
  92         <LI> mask typedef in codecvt_base, argument types in codecvt.
  93         what is know about this type?
  94
  95         <LI> why mask* argument in codecvt?
  96
  97         <LI> can this be made (more) generic? is there a simple way to
  98         straighten out the configure-time mess that is a by-product of
  99         this class?
 100
 101         <LI> get the ctype<wchar_t>::mask stuff under control. Need to
 102         make some kind of static table, and not do lookup evertime
 103         somebody hits the do_is... functions. Too bad we can't just
 104         redefine mask for ctype<wchar_t>
 105
 106         <LI> rename abstract base class. See if just smash-overriding
 107         is a better approach. Clarify, add sanity to naming.
 108
 109 </UL>
 110
 111
 112 <P>
 113 <H2>
 114 7. Acknowledgments
 115 </H2>
 116 Ulrich Drepper for patient answering of late-night questions, skeletal
 117 examples, and C language expertise.
 118
 119 <P>
 120 <H2>
 121 8. Bibliography / Referenced Documents
 122 </H2>
 123
 124 Drepper, Ulrich, GNU libc (glibc) 2.2 manual. In particular, Chapters &quot;6. Character Set Handling&quot; and &quot;7 Locales and Internationalization&quot;
 125
 126 <P>
 127 Drepper, Ulrich, Numerous, late-night email correspondence
 128
 129 <P>
 130 ISO/IEC 14882:1998 Programming languages - C++
 131
 132 <P>
 133 ISO/IEC 9899:1999 Programming languages - C
 134
 135 <P>
 136 Langer, Angelika and Klaus Kreft, Standard C++ IOStreams and Locales, Advanced Programmer's Guide and Reference, Addison Wesley Longman, Inc. 2000
 137
 138 <P>
 139 Stroustrup, Bjarne, Appendix D, The C++ Programming Language, Special Edition, Addison Wesley, Inc. 2000
 140
 141 <P>
 142 System Interface Definitions, Issue 6 (IEEE Std. 1003.1-200x)
 143 The Open Group/The Institute of Electrical and Electronics Engineers, Inc.
 144 http://www.opennc.org/austin/docreg.html
 145
 146