[PyKDE] QString and + operator

[Gerard Vermeulen]

On Thu, 14 Apr 2005 13:29:56 +0100 (BST)
"Phil Thompson" <phil at riverbankcomputing.co.uk> wrote:

> > Phil Thompson <phil at riverbankcomputing.co.uk> wrote:
> >
> >>> I'm making a requirement for a project I'm working on to work with
> > QString
> >>> instead of the built-in type as much as possible. Some will obviously
> >>> complain about this, needing to cast to str() to keep some pythonic
> > spirit
> >>> for concatenation, augmented assignment, etc. I remember seeing in this
> >>> message
> >>> (http://mats.imk.fraunhofer.de/pipermail/pykde/2002-November/003954.html)
> >>> that Phil mentioned supporting the '+' operator for mixed types. I
> > imagine
> >>> the problem is determining what to return. Is there still an interest
> >>> in
> >>> exposing this?
> >>
> >> I thought I had implemented it - must have slipped through the cracks.
> >> It
> >> will be in the next snapshot.
> >>
> >> Note that...
> >>
> >> q = QString("foo") + QString("bar")
> >
> > This one is OK with me (it's a faitful binding of the existing C++
> > operator)
> > but:
> >
> >> q = QString("foo") + "bar"
> >>
> >> ...will both work, but...
> >>
> >> s = "foo" + QString("bar")
> >>
> >> ...will not.
> >
> > Is there a need for QString+str in the first place? If not, I'd rather not
> > have it: explicit is better than implicit :)
> > Also, having it working one way but not the other is confusing. So better
> > prohibit mixed types altogether IMO.
> 
> I'm inclined to agree. It works because Python strings are automatically
> converted to QStrings when needed - although it is trivial to suppress
> this in this case.
> 
> Opinions?
>
I agree with Giovanni, because of orthogonality.

Gerard