[Kexi-devel] Rewriting Flow & formulashape for 3.0? (Re: Qt5 Port Status)

[Friedrich W. H. Kossebau]

Hi Yue,

good to see that you among other things have already turned Karbon back to 
life in the Qt5/KF5 spheres :) Rock on!

Now, the plan for Calligra 3.0 was to focus on porting all code to Qt5 and 
KF5.

No refactoring or rewriting should be done ideally, as that will only 
complicate things, like history (including Calligra's) teaches.
One step at a time, they say surely in many languages.

With KoReport, KoProperty, CalligraDB we are breaking this initial idea that 
we informally agreed on, makes me not that happy, would have like to work on 
Plan porting already now, and also am slightly fearing how much things changed 
with those libs. But at least it seems things are getting ready almost in time 
now... not yet git-cloned the repos, but soon will do.

Seems you, Yue, want to do more for 3.0 as well, let me comment on that 
please:

Am Samstag, 21. März 2015, 20:02:32 schrieb Yue Liu:
> I want to add something to the "Stuff that can be removed" part of the
> porting plan.
> 
> plugins/formulashape/ - I will write a new formula plugin based on
> libmathview.

How broken is the current shape? What stopped you/us/somebody else from 
replacing the current with a better one before the Qt5 port? Unless it is 
really broken, I would propose to simply port the existing shape for now, 
given that porting does not really need such an amount of work. Once you have 
a replacement done, the current shape can be still replaced with it, no?
I see no problem doing that e.g. "only" for 3.1, not 3.0.

> flow/ - I want to re-write calligraflow based on Karbon
> libraries.

"want" means there is no code yet, only plans, right? And it surely is not a 
quick thing to do, otherwise you might have done it already.
Could we also perhaps just port Flow as it is now, and do the rewrite with all 
needed patience during 3.x times?
Otherwise Calligra 3.0 would either be delayed by waiting for the Flow rewrite 
(which also would be done) or end up without Flow (and dissapoint the existing 
users), and who knows when you/us actually will have time to do that big task 
completely.

Yue, what do you think? I would really like to get 3.0 done as quick and good 
as possible, and have us spent all available time on getting things to work 
properly in Qt5/KF5 spheres first. It surely means some work for Flow & the 
formula shape that might be useless some months later. But it keeps things on 
the safe side.
Rest assured that I will work on porting Flow and the current formula shape 
(already would have done, if it was not reserved), so it's not about telling 
you what you should do, but rather us :)

Cheers
Friedrich