On 3/13/07, jiho <jo.irisson@...400...> wrote:
I know the question was not asked to me but I just step in to say that this will really be a nice improvement. Right now snapping behavior is quite difficult to understand and gives unexpected results when snapping both to nodes and bbox on objects with a large stroke (strange thing to do probably but the possibility is offered by the prefs dialog).
Agreed. This needs to be fixed.
However, now that I think of it some more, there's a potential problem. We actually have three kinds of things that snap:
1 bounding boxes
2 path nodes
3 some special points, some of them specific to an object type, such as geometric center, rotation axis, tips of a star, or the baseline origin in a text.
Now, our proposal (separate bbox modes) is aimed at cleanly separating 1 and 2 so that they are still both available but not get into the way of each other. But what about 3? Right now, as I understand it, in the snapping options the special points are covered by "snap nodes", i.e. are combined with 2. But I think they should snap in Selector tool regardless of what kind of bbox is chosen. I.e. we should separate them from the "nodes" category and always snap them with Selector. However, this potentially can lead to bugs similar to this one: for example, if I'm using visual bbox (with stroke) and one of a star's tips is adjacent to bbox, but not exactly (i.e. it's basically coincides with the node). Does anyone have an idea of how to deal with this?
In addition, will objects measurements also change with this bbox preference change?
Yes, measurements will depend on what kind of bbox is chosen.