[prev in list] [next in list] [prev in thread] [next in thread] 

List:       lyx-devel
Subject:    Re: feature request
From:       Andre Poenitz <andre.poenitz () mathematik ! tu-chemnitz ! de>
Date:       2006-04-10 10:25:47
Message-ID: 20060410102547.GD951 () millo ! mpi ! htwm ! de
[Download RAW message or body]

On Sun, Apr 09, 2006 at 02:15:51PM +0200, Leuven, E. wrote:
> > Also with active selection spanning more than the
> > current paragraph?
> 
> no, it only handles single paragraphs.
> 
> am trying to addresss this and have the following:

Better disable it when a selection is active. I doubt it is possible to
get (close to) right without using the cut&paste mechanism.

> 	case LFUN_PARAGRAPH_MOVE_DOWN: {
> 		pit_type begpit = cur.selBegin().pit()-1;
> 		pit_type pit = cur.selEnd().pit();

s/-/ - /

> 		bool counterUpdate = true;
> 		while (pit>begpit) {
> 			std::swap(pars_[pit], pars_[pit+1]);
> 			--pit;
> 		}
> 		++cur.pit();
> 		if (cur.selection()) {
> 			// adjust selection
> 		}
> 
> 		if (counterUpdate)
> 			updateCounters(cur.buffer());
> 		needsUpdate = true;
> 		break;
> 	}

Undo?

> i am not able to adjust the selection. 
> 
> ++cur.pit() moves the selection at the cursor side, but not at the
> point where the selection was started. hints?

Yes. Don't try it. There are dragons.
 
> second thing the code above does not do is flag whether we need to
> update the counters. the looping is over paragraphs, whereas
> needsCounterUpdate() wants a ParIterator...

I wonder (not related to this patch) whether the updateCounter stuff
should be made a bit more asynchronous from the user code's point of
view. I.e. just give the kernel a hint that certain counters are
out-of-date, but not actually re-create them.

Andre'
[prev in list] [next in list] [prev in thread] [next in thread]