src/common/unichar.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        src/common/unichar.cpp
   3 // Purpose:     wxUniChar and wxUniCharRef classes
   4 // Author:      Vaclav Slavik
   5 // Created:     2007-03-19
   6 // RCS-ID:      $Id$
   7 // Copyright:   (c) 2007 REA Elektronik GmbH
   8 // Licence:     wxWindows licence
   9 ///////////////////////////////////////////////////////////////////////////////
  10
  11 // ===========================================================================
  12 // headers
  13 // ===========================================================================
  14
  15 // For compilers that support precompilation, includes "wx.h".
  16 #include "wx/wxprec.h"
  17
  18 #ifdef __BORLANDC__
  19     #pragma hdrstop
  20 #endif
  21
  22 #ifndef WX_PRECOMP
  23     #include "wx/strconv.h"  // wxConvLibc
  24     #include "wx/log.h"
  25 #endif
  26
  27 #include "wx/unichar.h"
  28 #include "wx/string.h"
  29
  30 // ===========================================================================
  31 // implementation
  32 // ===========================================================================
  33
  34 // ---------------------------------------------------------------------------
  35 // wxUniChar
  36 // ---------------------------------------------------------------------------
  37
  38 /* static */
  39 wxUniChar::value_type wxUniChar::FromHi8bit(char c)
  40 {
  41 #if wxUSE_UTF8_LOCALE_ONLY
  42     wxFAIL_MSG( "invalid UTF-8 character" );
  43     wxUnusedVar(c);
  44
  45     return wxT('?'); // FIXME-UTF8: what to use as failure character?
  46 #else
  47     wchar_t buf[2];
  48     if ( wxConvLibc.ToWChar(buf, 2, &c, 1) != 2 )
  49     {
  50         wxFAIL_MSG( "invalid multibyte character" );
  51         return wxT('?'); // FIXME-UTF8: what to use as failure character?
  52     }
  53     return buf[0];
  54 #endif
  55 }
  56
  57 /* static */
  58 char wxUniChar::ToHi8bit(wxUniChar::value_type c)
  59 {
  60 #if wxUSE_UTF8_LOCALE_ONLY
  61     wxFAIL_MSG( "character cannot be converted to single UTF-8 byte" );
  62     wxUnusedVar(c);
  63
  64     return '?'; // FIXME-UTF8: what to use as failure character?
  65 #else
  66     wchar_t in = c;
  67     char buf[2];
  68     if ( wxConvLibc.FromWChar(buf, 2, &in, 1) != 2 )
  69     {
  70         wxFAIL_MSG( "character cannot be converted to single byte" );
  71         return '?'; // FIXME-UTF8: what to use as failure character?
  72     }
  73     return buf[0];
  74 #endif
  75 }
  76
  77
  78 // ---------------------------------------------------------------------------
  79 // wxUniCharRef
  80 // ---------------------------------------------------------------------------
  81
  82 #if wxUSE_UNICODE_UTF8
  83 wxUniChar wxUniCharRef::UniChar() const
  84 {
  85     return wxStringOperations::DecodeChar(m_pos);
  86 }
  87
  88 wxUniCharRef& wxUniCharRef::operator=(const wxUniChar& c)
  89 {
  90     wxStringOperations::Utf8CharBuffer utf(wxStringOperations::EncodeChar(c));
  91     size_t lenOld = wxStringOperations::GetUtf8CharLength(*m_pos);
  92     size_t lenNew = wxStringOperations::GetUtf8CharLength(utf[0]);
  93
  94     if ( lenNew == lenOld )
  95     {
  96         // this is the simpler case: if the new value's UTF-8 code has the
  97         // same length, we can just replace it:
  98
  99         iterator pos(m_pos);
 100         for ( size_t i = 0; i < lenNew; ++i, ++pos )
 101             *pos = utf[i];
 102     }
 103     else // length of character encoding in UTF-8 changed
 104     {
 105         // the worse case is when the new value has either longer or shorter
 106         // code -- in that case, we have to use wxStringImpl::replace() and
 107         // this invalidates all iterators, so we have to update them too:
 108
 109         wxStringImpl& strimpl = m_str.m_impl;
 110
 111         int iterDiff = lenNew - lenOld;
 112         size_t posIdx = m_pos - strimpl.begin();
 113
 114         // compute positions of outstanding iterators for this string after the
 115         // replacement is done (there is only a small number of iterators at
 116         // any time, so we use an array on the stack to avoid unneeded
 117         // allocation):
 118         static const size_t STATIC_SIZE = 32;
 119         size_t indexes_a[STATIC_SIZE];
 120         size_t *indexes = indexes_a;
 121         size_t iterNum = 0;
 122         wxStringIteratorNode *it;
 123         for ( it = m_str.m_iterators.ptr; it; it = it->m_next, ++iterNum )
 124         {
 125             wxASSERT( it->m_iter || it->m_citer );
 126
 127             if ( iterNum == STATIC_SIZE )
 128             {
 129                 wxLogTrace( wxT("utf8"), wxT("unexpectedly many iterators") );
 130
 131                 size_t total = iterNum + 1;
 132                 for ( wxStringIteratorNode *it2 = it; it2; it2 = it2->m_next )
 133                     total++;
 134                 indexes = new size_t[total];
 135                 memcpy(indexes, indexes_a, sizeof(size_t) * STATIC_SIZE);
 136             }
 137
 138             size_t idx = it->m_iter
 139                          ? (*it->m_iter - strimpl.begin())
 140                          : (*it->m_citer - strimpl.begin());
 141
 142             if ( idx > posIdx )
 143                 idx += iterDiff;
 144
 145             indexes[iterNum] = idx;
 146         }
 147
 148         // update the string:
 149         strimpl.replace(m_pos, m_pos + lenOld, utf, lenNew);
 150
 151 #if wxUSE_STRING_POS_CACHE
 152         m_str.InvalidateCache();
 153 #endif // wxUSE_STRING_POS_CACHE
 154
 155         // finally, set the iterators to valid values again (note that this
 156         // updates m_pos as well):
 157         size_t i;
 158         for ( i = 0, it = m_str.m_iterators.ptr; it; it = it->m_next, ++i )
 159         {
 160             wxASSERT( i < iterNum );
 161             wxASSERT( it->m_iter || it->m_citer );
 162
 163             if ( it->m_iter )
 164                 *it->m_iter = strimpl.begin() + indexes[i];
 165             else // it->m_citer
 166                 *it->m_citer = strimpl.begin() + indexes[i];
 167         }
 168
 169         if ( indexes != indexes_a )
 170             delete[] indexes;
 171     }
 172
 173     return *this;
 174 }
 175 #endif // wxUSE_UNICODE_UTF8