New framework: KCalCore

[Allen Winter]

On Thursday, July 18, 2019 12:18:36 PM EDT Volker Krause wrote:
> On Wednesday, 17 July 2019 01:51:42 CEST Aleix Pol wrote:
> > On Tue, Jul 16, 2019 at 6:10 PM Volker Krause <vkrause at kde.org> wrote:
> > > On Monday, 15 July 2019 18:43:42 CEST Aleix Pol wrote:
> > > > On Fri, Jul 12, 2019 at 9:03 PM Allen Winter <winter at kde.org> wrote:
> > > > > On Friday, July 12, 2019 12:23:58 PM EDT Volker Krause wrote:
> > > > > > With the 19.08 release approaching (and thus the deadline for
> > > > > > incompatible
> > > > > > changes if we go ahead with this plan), I'd like to raise this again
> > > > > > for
> > > > > > getting to a decision :)
> > > > > > 
> > > > > > Summary of what happened in the past weeks:
> > > > > > - the Person/Attendee slicing issue was fixed by making both
> > > > > > independent
> > > > > > types - several "leaf" types were turned into implicitly shared
> > > > > > value
> > > > > > types rather than being used heap-allocated inside shared pointers
> > > > > > - the dependency on the Akonadi supertrait.h header file was removed
> > > > > > - the virtual_hook usage in the incidence de/serialization code was
> > > > > > replaced by new virtual methods
> > > > > > 
> > > > > > Unless I missed something, the remaining unaddressed feedback is
> > > > > > down
> > > > > > to:
> > > > > > - Rename KCalCore to something else. I'm ok with executing the
> > > > > > rename,
> > > > > > but
> > > > > > somebody needs to tell me the new name :)
> > > > > 
> > > > > I don't remember the reason for changing the name.
> > > > > I vote for not changing the name. KCalCore is as good as any, imo
> > > > > 
> > > > > > - Alexander P's fundamental objections to the current KCalCore API
> > > > > > 
> > > > > > How do we proceed?
> > > > > > 
> > > > > > Thanks,
> > > > > > Volker
> > > > > > 
> > > > > > On Sunday, 7 April 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,
> > > > > > > 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.
> > > > > > > 
> > > > > > > KCalCore depends on Qt and libical only, making it a Tier 1
> > > > > > > functional
> > > > > > > framework.
> > > > > > > 
> > > > > > > KCalCore used to be part of part of kdepimlibs in the KDE4 era, so
> > > > > > > it's well prepared for complying with the API and ABI guarantees.
> > > > > > > 
> > > > > > > I'd suggest the same timeline as proposed for KContacts [1].
> > > > > > > During
> > > > > > > the PIM
> > > > > > > sprint we did a number of fixes and cleanups as part of the review
> > > > > > > for
> > > > > > > KF5
> > > > > > > that make 19.08 the earliest release after which we can switch as
> > > > > > > well, so
> > > > > > > we are looking at a switch in Sep/Oct this year.
> > > > 
> > > > If you don't remember, just scroll up a bit. :P
> > > > 
> > > > I went through the thread and that's what we said:
> > > > - It's a framework to understand the ical format, this is not conveyed
> > > > by the current name
> > > > - The Core postfix doesn't add anything and we are already using it
> > > > for differentiating different framework targets (e.g. KF5::ConfigCore
> > > > and KF5::ConfigGui)
> > > 
> > > "Core" had exactly the same meaning here when the widget parts were split
> > > out during the Nokia times. Less relevant today probably.
> > > 
> > > > I see KCal, KCalendar, KiCalendar, KiCal, KCalendarEvents,
> > > > KCalendarIncidences being suggested. It's not going to be me who
> > > > chooses one though.
> > > 
> > > If those are the options I'd vote for KCal or KCalendar, KiCal is too
> > > close to KiCad IMHO, and the rest is just unnecessarily long. Anyway,
> > > let's please have a quick decision on this, it's going to be a large
> > > change, so I'd like to get this in ASAP.
> > > 
> > > Thanks,
> > > Volker
> > 
> > I tend to agree. If so I'd suggest KCalendar. Following KContacts
> > (which should possibly be KVCards, but we assume vcards are the one
> > and only format ever for contacts).
> 
> It's not necessarily the one and only file format (and support for other file 
> formats is possible in those libraries, KCalCore e.g. can read some iCal 
> predecessor too), but both vcard and ical have effectively been the one 
> ubiquitous data model in the past two decades for this kind of data. That IMHO 
> justifies the very generic naming :)
> 
> In the earlier discussion David Jarvie raised concerns about the KCalendar 
> name being misleading towards a calendar systems related library. Valid point, 
> but IMHO the topic of calendar systems is more specialized (and typically an 
> implementation detail of other libs), compared to the calendar conecpt 
> KCalCore is referring to, so KCalendar vs KCalendarSystems would IMHO be 
> acceptable.
> 
> Other views/opinions on this?
> 
There was once a KCalendarSystem class.
We may need to resurrect it , but I forget why.  Maybe for holidays.

I don't think KCalendar or KCalendarSystem is good.

to make an analog of KContacts you'd have to call this KIncidences.
I don't hate that.  The library is really about Incidences.  A calendar
is basically a collection of incidences,