size_t
wxConvAuto::ToWChar(wchar_t *dst, size_t dstLen,
const char *src, size_t srcLen) const
{
// we check BOM and create the appropriate conversion the first time we're
// called but we also need to ensure that the BOM is skipped not only
// during this initial call but also during the first call with non-NULL
// dst as typically we're first called with NULL dst to calculate the
// needed buffer size
wxConvAuto *self = const_cast<wxConvAuto *>(this);
if ( !m_conv )
{
self->InitFromInput(&src, &srcLen);
if ( dst )
self->m_consumedBOM = true;
}
if ( !m_consumedBOM && dst )
{
self->m_consumedBOM = true;
SkipBOM(&src, &srcLen);
}
// try to convert using the auto-detected encoding
size_t rc = m_conv->ToWChar(dst, dstLen, src, srcLen);
if ( rc == wxCONV_FAILED && m_bomType == BOM_None )
{
// if the conversion failed but we didn't really detect anything and
// simply tried UTF-8 by default, retry it using the fall-back
if ( m_encDefault != wxFONTENCODING_MAX )
{
if ( m_ownsConv )
delete m_conv;
self->m_conv = new wxCSConv(m_encDefault == wxFONTENCODING_DEFAULT
? GetFallbackEncoding()
: m_encDefault);
self->m_ownsConv = true;
rc = m_conv->ToWChar(dst, dstLen, src, srcLen);
}
}
return rc;
}
size_t
wxConvAuto::ToWChar(wchar_t *dst, size_t dstLen,
const char *src, size_t srcLen) const
{
// we check BOM and create the appropriate conversion the first time we're
// called but we also need to ensure that the BOM is skipped not only
// during this initial call but also during the first call with non-NULL
// dst as typically we're first called with NULL dst to calculate the
// needed buffer size
wxConvAuto *self = const_cast<wxConvAuto *>(this);
if ( !m_conv )
{
if ( !self->InitFromInput(src, srcLen) )
{
// there is not enough data to determine whether we have a BOM or
// not, so fail for now -- the caller is supposed to call us again
// with more data
return wxCONV_FAILED;
}
}
if ( !m_consumedBOM )
{
SkipBOM(&src, &srcLen);
if ( srcLen == 0 )
{
// there is nothing left except the BOM so we'd return 0 below but
// this is unexpected: decoding a non-empty string must either fail
// or return something non-empty, in particular this would break
// the code in wxTextInputStream::NextChar()
//
// so still return an error as we need some more data to be able to
// decode it
return wxCONV_FAILED;
}
}
// try to convert using the auto-detected encoding
size_t rc = m_conv->ToWChar(dst, dstLen, src, srcLen);
if ( rc == wxCONV_FAILED && m_bomType == BOM_None )
{
// if the conversion failed but we didn't really detect anything and
// simply tried UTF-8 by default, retry it using the fall-back
if ( m_encDefault != wxFONTENCODING_MAX )
{
if ( m_ownsConv )
delete m_conv;
self->m_conv = new wxCSConv(m_encDefault == wxFONTENCODING_DEFAULT
? GetFallbackEncoding()
: m_encDefault);
self->m_ownsConv = true;
rc = m_conv->ToWChar(dst, dstLen, src, srcLen);
}
}
// don't skip the BOM again the next time if we really consumed it
if ( rc != wxCONV_FAILED && dst && !m_consumedBOM )
self->m_consumedBOM = true;
return rc;
}
void wxConvAuto::InitFromInput(const char **src, size_t *len)
{
m_bomType = DetectBOM(*src, *len);
InitFromBOM(m_bomType);
SkipBOM(src, len);
}
