[Kst] extragear/graphics/kst/src/libkstapp

[Andrew Walker]

I would agree with your first point. The second point works both ways though.

Imagine the case with the current code (after George reverted my change) where 
the user changes the x-label text in the plot dialog (without hitting Apply) 
and then zooms in on the same plot through the mouse. Hitting Apply now would 
revert the zoom. Almost certainly not would the user would expect.

It seems that there is always a case where the user is going to be mildly 
surprised. A better solution might be to disallow any changes to the plot 
selected in the plot dialog, while the plot dialog is open. Any attempt to 
make such changes could ask the user to either cancel or apply their changes
in the plot dialog first. I suspect the code for this would become somewhat 
messy, but if that is the right solution it shouldn't be a deterrent.

Andrew

On March 21, 2006 3:57 pm, Barth Netterfield wrote:
> modified() tells the doc that it needs to save on exit.
>
> Our policy is that we don't ask to save on exit for changes on zoom.  Not a
> big deal, but if all you did was load from the command line, look around a
> file, and exit (common in using kst to look at intermediate data) you
> probably don't want to save.
>
> I'm not sure how good an idea in a larger sense this fix is however.  The
> plot dialog is not really intended to be a property viewer.  The entries
> are filled with the current state for convenience, but I don't think that a
> zoom in the main window should change the entries (imagine that the user
> has been changing things, hasn't hit apply yet, zooms in to check more
> precisely where the upper range should be, and finds the other entries it
> put in
> over-ridden....
>
> cbn
>
> On Tuesday 21 March 2006 15:40, Andrew Walker wrote:
> > George, could you be more specific. In particular what are the side
> > effects besides "doc set dirty when it isn't". Could you also describe
> > some of the cases that it misses. Thanks,
> >
> > Andrew
> >
> > On March 21, 2006 3:27 pm, George Staikos wrote:
> > > SVN commit 521230 by staikos:
> > >
> > > I ran this patch through many tests here and they confirm that there
> > > are many side effects to this (doc set dirty when it isn't, for
> > > instance). It also misses more cases than it catches.  I think this is
> > > a losing cause, but if it's to be fixed, it needs to be done
> > > differently. REOPEN: 123568
> > >
> > >
> > >  M  +1 -2      kst2dplot.cpp
> > >
> > >
> > > --- trunk/extragear/graphics/kst/src/libkstapp/kst2dplot.cpp
> > > #521229:521230 @@ -4157,7 +4157,6 @@
> > >      kstdDebug() << "mouse release: do update" << endl;
> > >      setDirty();
> > >      static_cast<KstViewWidget*>(view)->paint();
> > > -    emit modified();
> > >    }
> > >  }
> > >
> > > @@ -5038,7 +5037,7 @@
> > >    if (handled) {
> > >      view->paint();
> > >      e->accept();
> > > -    emit modified();
> > > +    //emit modified();
> > >      return;
> > >    }
> > >
> > > _______________________________________________
> > > Kst mailing list
> > > Kst at kde.org
> > > https://mail.kde.org/mailman/listinfo/kst
> >
> > _______________________________________________
> > Kst mailing list
> > Kst at kde.org
> > https://mail.kde.org/mailman/listinfo/kst
>
> _______________________________________________
> Kst mailing list
> Kst at kde.org
> https://mail.kde.org/mailman/listinfo/kst