Thanks Mathew, it's not just me being stupid then.

Following up, I took a look at some old parts that I used fragments to make and they don't have this issue - even when placed in the same file. So it seems to be an issue with the part itself. As for printing, I tested that tonight and the part prints wrong as well (matching the screen display).

Just for kicks I have attached an image of showing the behavior I am seeing. The fragment portion is in a subroutine called by the main part with the intention of displaying the handrail section in plan. In this instance there is 2D line work that forms the outside of the rail profile and a fill for the body. If you look you can see that the line work displays correctly (essentially smooth) but the fill is heavily segmented.

I guess I will create a new part and rewrite the script to see if that helps.

For what it is worth, further investigation shows that this problem only occurs when you use a MUL2 command before the FRAGMENT2. Here is the code I was using:

PEN gs_cont_pen 
FILL gs_fill_type

dx = 1.701"    ! width
dy = 1.578"    ! height


MUL2 A/dx, B/dy
FRAGMENT2 1, 1

If I comment out the MUL2 all is well. Strange

dx = 1.701"    ! width
dy = 1.578"    ! height
MUL2 A/dx, B/dy
FRAGMENT2 1, 1

Matthew wrote:
Well at least it's an easy fix. If you need mirrored versions of the same fill you can always just use different fragments.

It looks like from his code snippet that he is resizing the fragment, not mirroring it, so the fix may be in converting the fill from a drawn symbol to a coded fill in the 2D Script.

Erich: have you tried dragging & dropping the drawn fill into the 2D Script Window to see if ArchiCAD will convert the fill to code automatically?

David

David,

Yes that is exactly the fix I used. It just takes a few additional steps I had hoped to avoid, not that it is difficult.

It is odd that this just happens to fills in the fragment and not the line work. So far I have not found documentation of this behavior and it feels like a bug to me. The effect seems to be more pronounced the further from the original size and proportions the MUL2 makes things.

Hi,

I tried to reproduce the problem having a MUL2 command in the script before the fragment but all worked fine for me.

Naturally, it is not the desired function that a gap between the line and the fill appears... Could you please send me a simplified version of the problem-prone object?

Thanks,