2. reimplemented UTF-16/32 conversions using To/FromWChar() API instead
of MB2WC/WC2MB for Windows (or rather SIZEOF_WCHAR_T == 2 platforms),
the first tangible result is that reading UTF-32BE text streams now
works too
3. more off by 1 fixes
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@38585 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
2. this allows to use wxConvAuto() instead of wxConvUTF8 as default value
for this parameter in the classes which read text from the file: wxConvAuto
automatically recognizes the BOM at the start of file and uses the correct
conversion
3. don't use Windows for UTF-7 conversions as there is no way to make it
fail on invalid UTF-7 strings; use our own wxMBConvUtf7 instead
git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@38570 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
