New framework: KCalCore

[Aleix Pol]

On Wed, Apr 17, 2019 at 6:40 PM Volker Krause <vkrause at kde.org> wrote:
>
> On Sunday, 14 April 2019 13:31:41 CEST David Faure wrote:
> > On dimanche 7 avril 2019 14:45:09 CEST Volker Krause wrote:
> > > Hi,
> > >
> > > I'd like to propose KCalCore for review to move from KDE PIM to KF5.
> > >
> > > KCalCore is an implementation of the iCalendar standard based on libical,
> >
> > I wonder about the name, which doesn't mean much outside the circle of PIM
> > people. Shouldn't this be called KCalendar ?
> >
> > If the "Core" simply means non-GUI, we certainly don't have that word in
> > every non-GUI framework.
>
> Renaming the namespace should be manageable, we can soften the blow for
> external users with a namespace alias I guess, to at least keep SC until
> everyone has adapted.
>
> > > covering the data model, input/output and the rather complex recurrence
> > > algorithms defined in that standard. It's used outside of KDE PIM as well,
> > > e.g. by Zanshin or the Plasma Mobile calendar app.
> >
> > This makes me wonder: where does that mobile calendar app get the events
> > from? Akonadi? (then it still depends on kde/pim/*, and this move in itself
> > doesn't really remove the unwanted workspace->apps dependency?)
>
> So far it looked like just a local ical file, which I guess is temporary,
> while focusing on mobile UI development. Eventually I at least expect the need
> for KDav there too. So just moving KCalCore isn't going to be enough
> obviously, but it is a necessary first step either way.
>
> > Zanshin does use akonadi (though one could imagine a mobile version that
> > only uses KDav and KCalCore^H^H^H KCalendar).
> >
> > Some review:
> >
> > icalformat_p.h:    //TODO: KDE5, move this implementation to
> > icalformat_p.cpp incidencebase.h: * // TODO_KDE5: Provide a virtual
> > serialize() method, as done with assign() and equals(). incidencebase.h: *
> > // TODO_KDE5: Provide a virtual serialize() method, as done with assign()
> > and equals(). person.h:    // TODO_KDE5: FIXME: This operator does
> > slicing,if the object is in fact one of the derived classes (Attendee)
>
> Person/Attendee I'd like to ideally see split into two non-polymorphic value
> types, as that would make direct QML consumption easier. That would also
> address the slicing risk.

Aliases shouldn't be strictly necessary because the framework can be
called KCal and the target KF5::CalCore, much like in KConfig.

We could decide to have it this way though.

Aleix