behaviour of bytevector

[Lukáš Lalinský]

On Wed, Apr 21, 2010 at 10:06 PM, Matthieu Riolo
<matthieu.riolo at googlemail.com> wrote:
> String::String(const ByteVector &v, Type t)
> {
>   d = new StringPrivate;
>   if(v.isEmpty())
>     return;
>   if(t == Latin1 || t == UTF8) {
>     int length = 0;
>     d->data.resize(v.size());
>     wstring::iterator targetIt = d->data.begin();
>     for(ByteVector::ConstIterator it = v.begin(); it != v.end() && (*it);
> ++it) {
>       *targetIt = uchar(*it);
>       ++targetIt;
>       ++length;
>     }
>     d->data.resize(length);
>   }
>
> If I see that correct is the check && (*it)  a bit too much. Because the
> input is a ByteVector it should be clear that the binary can contain NULL
> values. But it should not matter as long we get the size (and this one we
> always have). I guess it was copy&pasted from the other constructors?

No, TagLib::String does not represent binary data, it represents
characters (Unicode code points) and stopping on a NULL byte is the
expected behavior.

As I asked in the other email, why are you converting the ByteVector
into a String?

Lukas