[Development] QUtf8String{, View}

[Thiago Macieira]

On sexta-feira, 15 de maio de 2020 10:52:49 PDT Matthew Woehlke wrote:
> > Like that, it's just "array of bytes of an arbitrary encoding (or none)".
> > There's still a reason to have QByteArray and it'll need to exist in
> > networking and file I/O code. That means the string classes, if any, need
> > to be convertible to QByteArray anyway.
> 
> I think we can learn from Python 3 here... QByteArray should go the way
> of QStringList, i.e. yes, it *should* be a QVector<byte>. Like
> QVector<QString>, it might (should) have additional methods, such as
> explicit conversion to/from QString (a la Python's encode/decode), but
> it should *not* have string-like manipulation (e.g. toUpper).

Those are all Qt 7 work. We can deprecate those methods in Qt 6, but not 
remove them in 6.0.

Python's bstr still behaves string-like and has methods like 
QByteArray::indexOf(const char *).  QVector has no such methods.

But since we do have QListSpecialMethods, we can add inject them into QVector 
too.

> >> So, assuming the premise that QByteArray should not be string-ish
> >> anymore, what do we want to have as the result type of QString::toUtf8()
> >> and QString::toLatin1()? Do we really want mere bytes?
> 
> Yes. Maybe. Again, this is how Python 3 works.
> 
> It might make sense to have a QUtf8String class, but that should be
> distinct from, and not implicitly constructible from, QByteArray a.k.a.
> QVector<byte>. (Implicit conversion *to* QByteArray might be okay.)
> 
> (BTW, is 'byte' QByte or std::byte? Can we possibly achieve the latter?)

There's no QByte and we shouldn't have that type.

std::byte is an enum around the actual byte type (unsigned char) and char is 
also a byte.

-- 
Thiago Macieira - thiago.macieira (AT) intel.com
  Software Architect - Intel System Software Products