[kplato] commit of several changes

[Dag Andersen]

On Torsdag den 10. juli 2003 22:51, Heiko.Evermann at gmx.de wrote:
> Hi everyone,
>
> I just had a look at the mailing list archive. Quite a lot of time
> has passed
> since I last mailed. I have been working on some changes that I
> proposed:
>
> 1) delete node
> 2) move/indent/unindent node
> 3) => change in the object structure
> 3.1) use task for subtasks, tasks and milestones
> 3.2) use project only for the root model
> 3.3) hide the root model in the views
>
> I did not have much time for programming since our little daughter
> keeps us occupied at home. But by now I have collected so many
> changes that I would
> like to check them in. Otherwise, if someone else has any changes,
> the merge would be a mess. I know that not everything is perfect,
> but I hope You do not mind checking in.
No :)
>
> New features:
>
> * node delete
> * node move up
> * node move down
> * node indent
> * node unindent
> * insert  new subnode vs. insert  new sibling node by icon.
> * internal structure changes
> * some new icons for the new functions, including icons for the
> several insert options.
> * extended context menu for the functions
> * moved pert code into the pert view. (Thomas Zander was right to
> complain about
> mixing model and view code, so I used the opportunity to sort this
> thing out.
Exelent.
> The pert code is quite complicated. Is it documented anywhere? I
No.
> hope it still
> works as indended. My tests indicated that it is still working, but
> perhaps someone who programmed that in the first place could have a
> look?) remove milestone files (no longer needed, as handled in
> task)
Not quite, it crashes when loading from file ;) I've fixed that by 
moving drawing of relations to after all nodes have been drawn.
Known bug: Relations between sub-tasks are not loaded.
I don't know if I have the time to find it, I'm off on holyday et al 
tomorrow and probably won't be back before late august.
>
> things to do:
> get the time calculation working. I think it does not work at the
> moment. And
> I would prefer to move it out of the models. It is spread over
> several classes
> and IMHO it should be handled in project, or even better in a class
> of its own
> whith the sole responsability to perform these calculations. I
> think this is neccessary, as several algorithms for these
> calculations can be expected, and then we could have on class for
> each algorithm. It would be confusing to spread several calculation
> algorithms over several classes.)
>
> Delete can cause crash, when deleted object takes part in
> relations. (This should
> be handeled in project. All new actions are relayed to the project
> object, so that
> the view can just tell the project what to do, and project is
> supposed to keep everything consistent.
>
> With these changes, kplato can do some more things than before, and
> I hope you
> like these changes, even though not everything is perfect.
>
> Try the new things out and tell me what You think.
>
> Regards,
>
> Heiko
>
> PS: During checkin I noticed that some things have changed in the
> koffice interface.
> I had a conflict in kptpart.cc which was changed 9 days ago by
> faure. I think
> this is a global change in koffice. The end result is that I have
> to update all
> of koffice to test whether the current CVS of kplato still works.
> (As should be done after every checkin...)
> For this it is too late, so I can only check that tomorrow or over
> the weekend.
> Sorry. I did not realize this conflict until after I was in the
> midst of the checkin.

-- 
Mvh,
Dag Andersen