gEDA-dev: [gschem] Rounding error while dragging objects?

[Bernd Jendrissek]

Hmm, maybe not rounding.  Something wrong with expose events?  Can
somebody please see if they notice this weirdness too?

On 9/5/07, Bernd Jendrissek <bernd.jendrissek at gmail.com> wrote:
> I've noticed that when I drag things around (text and components)
> their sub-objects seem to move relative to each other.  The diagonals
> of an op-amp seem to move towards each other and so cross each other
> near the point of the triangle.  There are similar effects with text -
> the prim_objs lines seem to disconnect from each other.

These randomish offsets seem to be independent of any window clipping
as I describe it below.

> When I release the drag, everything's back to normal again: the text
> lines are contiguous and the op-amp is a neat triangle again.
>
> Even stranger: when I zoom way in and drag something, even a single
> line segment seems to develop a break, like there's some spacetime
> rift between the endpoints.

This spacetime rift is most easy to reproduce: select a line that is
not at a multiple of 45 degrees, obscure the drawing area with another
window, then drag the slanted line.  Text with lots of 'o' and 'e' is
good here.  The obscuring window seems to project rays of badness
horizontally from its top and bottom edges.

I also get parts of lines XOR-rendered onto the scrollbar at the
bottom of the drawing area, while I drag the slanted lines.  Only a
few pixels - a handful or two.

> I noticed in the coord window that when the breakage is at its worst,
> the world coord space is very sparse on the screen.  I can move the
> pointer across the whole window and the world coords will hardly
> change. That's why I'm guessing at rounding error.

No, I got fooled by the snap setting.  Forget this.