void UnicodeTestCase::ConversionUTF16()
{
static const StringConversionData utf16data[] =
{
#ifdef wxHAVE_U_ESCAPE
StringConversionData(
"\x04\x1f\x04\x40\x04\x38\x04\x32\x04\x35\x04\x42\0\0",
L"\u041f\u0440\u0438\u0432\u0435\u0442"),
StringConversionData(
"\x01\0\0b\x01\0\0a\x01\0\0r\0\0",
L"\u0100b\u0100a\u0100r"),
#endif
StringConversionData("\0f\0o\0o\0\0", L"foo"),
};
wxCSConv conv(wxFONTENCODING_UTF16BE);
for ( size_t n = 0; n < WXSIZEOF(utf16data); n++ )
{
const StringConversionData& d = utf16data[n];
d.Test(n, conv);
}
// special case: this string has consecutive NULs inside it which don't
// terminate the string, this exposed a bug in our conversion code which
// got confused in this case
size_t len;
conv.cMB2WC("\x01\0\0B\0C" /* A macron BC */, 6, &len);
CPPUNIT_ASSERT_EQUAL( 3, len );
// When using UTF-16 internally (i.e. MSW), we don't have any surrogate
// support, so the length of the string below is 2, not 1.
#if SIZEOF_WCHAR_T == 4
// Another one: verify that the length of the resulting string is computed
// correctly when there is a surrogate in the input.
wxMBConvUTF16BE().cMB2WC("\xd8\x03\xdc\x01\0" /* OLD TURKIC LETTER YENISEI A */, wxNO_LEN, &len);
CPPUNIT_ASSERT_EQUAL( 1, len );
#endif // UTF-32 internal representation
#if SIZEOF_WCHAR_T == 2
// Verify that the length of UTF-32 string is correct even when converting
// to it from a longer UTF-16 string with surrogates.
// Construct CAT FACE U+1F431 without using \U which is not supported by
// ancient compilers and without using \u with surrogates which is
// (correctly) flagged as an error by the newer ones.
wchar_t ws[2];
ws[0] = 0xd83d;
ws[1] = 0xdc31;
CPPUNIT_ASSERT_EQUAL( 4, wxMBConvUTF32BE().FromWChar(NULL, 0, ws, 2) );
#endif // UTF-16 internal representation
}
void TextStreamTestCase::TestEmbeddedZerosUTF32BEInput()
{
TestInput(wxMBConvUTF32BE(), txtUtf32be, sizeof(txtUtf32be));
TestInput(wxCSConv(wxFONTENCODING_UTF32BE), txtUtf32be, sizeof(txtUtf32be));
}