https://www.ribbonsoft.com/bugtracker/i ... sk_id=2155
A nice example of such can be viewed in the mentioned topic:
https://qcad.org/rsforum/viewtopic.php? ... 889#p30905
Managed to explain it myself.
The inner offset shape of an arc with distance R is a Null-lenght arc.
The outer contour gets crippled and the fill is faulty.
Also meaning that polylines with width are rendered as solid hatched contours.
In the past there was a disagreement on that Null-entities should not be cast to a document.
Then again, I fully agree with Andrew that they have a meaning mathematically.
There are two kinds:
- Zero length entities having a certain position, visible as a blob.
- Invisible Null-entities only selectable with TA and those will turn up at the origin when a file is re-opened.
They also occur as a Null-length segment of a polyline.
Referring to the arc segment in red : (Select the poly-Arc to see its reference points
) The inner offset is disregarded and in the absence of its endpoints the outer offset arc is closed (Cyan) as an arc segment.
Not disregarded, that arc would have center=start=endpoint = (180;-30)
Even when avoiding to cast the Null-arc itself, the orange lines would be valid and it would display as an arc sector.
I used that as 2D centroid markers, knowing the issue and avoided W=2R.
But the error also turns up in any polyline with variable arc segments with a given width.
E.g.: The explosion of a spline or that of an ellipse.
Or on rounded polyline shapes where the width would match 2R.
- W=2R issue.png (4.28 KiB) Viewed 721 times
It seems that there is a mixed consensus on Null-entities.
I now also can explain the issues encountered when mixing widths and non-zero Z values.
A complete different topic and off topic for QCAD that is considered as 2D.
Regards,
CVH