I agree with you all .

method 3 is better than method 1.

2010/10/22 Matti Airas <[email protected]>

> Hi,
>
> Basically, I agree with you and Lizardo. Here's my reasoning:
>
> On 21 October 2010 21:12, Renato Araujo Oliveira Filho
> <[email protected]> wrote:
>
> > 1 - We can try follow the same default signature in PySide and PyQt, I
> > do not know which rule PyQt follow but I think it is the first one
> > export to Python.
> > With this solution we will keep the compatibility with PyQt, but this
> > will generate a lot of work to check all signals if they are
> > registered in same order as PyQt, if PyQt change the order we will
> > keep this update.
>
> I think it's a bit problematic to define PySide's behaviour through
> PyQt, if the PyQt behaviour isn't explicitly defined. We're bound to
> end up with incompatible implementations anyway, if we implement some
> new classes first (as we tend to do), and PyQt then implements it in a
> different manner.
>
> > 2- Other option is to keep our order without any preoccupation with PyQt.
> > This will keep the things like they are now, without any extra work,
> > but with incompatibilities.
>
> This is slightly better than the first option, in my opinion. At least
> there's no extra work :-), we don't make any impossible-to-keep
> compatibility promises, and we are internally consistent. However, for
> people porting code from PyQt, this is going to be really suboptimal,
> with errors being quite non-obvious.
>
> > 3 - The last option is to make mandatory use of the signature when
> > connecting signals.
> > In my opinion this is the best solution, because this avoids mistakes
> > and make the things compatible with PySide and PyQt with less work.
> > The other good point of this option is: You can keep your code
> > compatible with other Qt versions even though new signals are inserted
> > before the old ones or if one signal with only one signature receive a
> > new signature in future Qt versions.
>
> Nothing much to add to your and Lizardo's arguments.
>
> Except: do you think we should make a PSEP for this design decision?
> If so, I can help you write it, if you'd like that.
>
> Cheers,
>
> ma.
> _______________________________________________
> PySide mailing list
> [email protected]
> http://lists.openbossa.org/listinfo/pyside
>
_______________________________________________
PySide mailing list
[email protected]
http://lists.openbossa.org/listinfo/pyside

Reply via email to