On 6/14/11 9:05 PM, JLuc wrote: > Here is maybe an answer : > the probleme is about the same as with undo. > > There are sOOOOooooo many actions that change the doc... and ALL might > not rise the flag. > So the danger is to believe that the doc has not changed, and to let > the user lieve without saving, > when it actualy has changed, unoticed by this flag. > > (But is there no unique low level interaction API with the data > structure ?) > > JLuc
Yes.. this is exactly it.. the changed() call within the doc gets called for every single action possible.. and there is something somewhere that calls that API once too often after the save process clears the flag. Its so low priority.. but maybe I will have another look. Once upon a time, we had literally 12-24+ man hours per day going into Scribus.. now.. it could be that per week, *if that*, due to family and work constraints. Priorities get placed where they *hopefully* should.. mostly going towards trying to get 1.4.0 released, hence the number of RC releases to try and get people to test as they hardly test until a release is made. Craig
