[Kst] kst2 and kst 1.x/port
syntheticpp at gmx.net
Wed Dec 2 21:26:13 CET 2009
Barth Netterfield wrote:
> The structures in the kst1 scripting are only loosely associated with the
> internal structure of kst; the scripting is based on a bunch of intermediate
> objects which are mapped (sometimes in a contrived way) to the underlying kst
Fine, this makes it possible to also provide such intermediate objects for
kst2. It could be worse, a design like for kst2 scripiting would make it
nearly impossible to support kst1 scripts.
> Instead, we could decide to go to an automatically generated scripting
> paradigm for kst2, and expose the internal structures more directly. This
> would be far less work, but comes with the cost that internal structure
> changes would change the scripting language. At that point, we would have to
> build intermediate objects to protect the language from those changes. In our
> dreams, of course, we will have such a good internal structure before we start
> scripting, that we'll never want to change it :-)
> That being said, I would still welcome a port of the 1.x scripting to kst2.
Binding QObjects to QtScript doesn't need special code, so it is maybe
possible to avoid generating code for the binding.
But hiding internals is a problem, indeed.
> On Wednesday 02 December 2009 13:53:31 Peter Kümmel wrote:
>> Barth Netterfield wrote:
>>> Andrew felt that that kst1.x compatible scripting in kst2 would be very
>>> hard - perhaps even harder than the 1.x straight port, and would leave us
>>> with a worse product than a 'from scratch' scripting paradigm for kst2.
>> What's the difference between scripting in kst1 and kst2?
>> Kst2 uses QtScript, and kst1? Couldn't the support of kst1 scripts
>> be simply done by another binding to the new kst2 classes?
>> When support of kst1 scripting is such important it's maybe worth the
>> effort, especially when it focuses all developers onto trunk/kst2.
>> Kst mailing list
>> Kst at kde.org
> Kst mailing list
> Kst at kde.org
More information about the Kst