Re: [Rosegarden-devel] Undo loses information on cursor position

[Pedro Lopez-C. <ped...@gm...>]

On Saturday, 29 de April de 2006 10:32, Heikki Johannes Junes wrote:
> I have been using Rosegarden together with my piano, which is a great
> thing!
>
> When adding notes, it is possible to accidentally hit a wrong note.
> In that case, pressing undo is a great thing but, after pressing undo,
> the curson position is often lost.
>
> It is very easy to reproduce this problem, just open notation view and hit
> to quarter notes (by a mouse, for example), one after another, in the
> beginning of the composition. Undoing removes the last note and changes it
> to a rest, which is the correct behaviour, but it does not recover the
> previous cursor position.
>
> IMHO, the correct behaviour would be that undo would recover the program
> back to the state in which is was just before hitting a wrong key. Such a
> behaviour requires also an exact tracking of the curson position.

I basically agree. If the insert note command moves the cursor position when 
done, the undo operation for this same command should restore the cursor to 
its previous position. This is true for both the note and rest insertion 
tools. If there is no strong opposition, I would ask you to open a bug report 
in SourceForge:  
http://rosegarden.sourceforge.net/tutorial/bug-guidelines.html

You can also observe that the key insertion tool doesn't move the local cursor 
position, which is good :). And if you are in "chord mode", the cursor 
doesn't move either after adding a note, to help you to add notes to  the 
same chord easily. In "triplet mode", the note insertion tool also moves the 
cursor, but behaves a bit different. The "local cursor" is a very interesting 
feature, and there are useful commands for it (and also room for  
improvements) under the "tools->local cursor" menu, in both the notation and 
the matrix views. For instance, you can move the local cursor position using 
the arrow keys in the computer keyboard.

About the matrix view, you can observe that a note insertion in it doesn't 
modify the local cursor position; neither the "paste" operation does. The  
matrix view behaves very different from the notation view in many aspects, 
and also from the local cursor point of view, which also exists here. I think 
this is acceptable, because the two views are used for very different tasks 
and even there are people that only wants to work with one of them, ignoring 
the other view completely.

Returning to the main argument. I'm not sure about what the other developers 
think about this matter. There have been some discussions in this mailing 
list about controversial behavior in the program, usually ending with a 
consensus. Let's hear about what other people thinks.

Regards,
Pedro