ATTN: John Uhden @cv_inside bug.

R2004.
John.
While testing my area clear program that now implements the @cv_inside function I found this problem.
My program produces a point that is way out in space (-6.88997e+015 11.283 0.0). This is possible when the two segments that I use the @delta function on to get the deflection angle return a very sharp angle (near zero).
When this point is fed to @cv_inside it returns true.
I found that (setq Points (vlax-invoke Outer "IntersectWith" $cv_ray acExtendNone)) returns a list of about 11 points.
Command: (@cv_inside tp vlaobj-outline)

(49.5598 9.51014 0.0 48.5162 9.36869 0.0 48.5162 7.37672 0.0 43.7037 7.37672
0.0 43.7037 9.98428 0.0 43.7208 10.0058 0.0 43.8535 10.2601 0.0 44.1099 10.4392
0.0 48.6099 10.4392 0.0 48.5162 10.4392 0.0 48.5162 10.4472 0.0)
Thus cv_inside returns T.
The polygon outline is not very complex or odd (like a box with a bulge and an inverted corner) So the ray should not cross more than 2 or 3 times.
I may be able to fix my side so that it doesn't create a point like that but thought I should still say something in case others have trouble.

Bill

 

Hmm...
I'm looking at it with the ray made visible.
When I'm zoomed in it creates that point I talked about.
When I'm zoomed out a few X, it doesn't make the point way out in space. Works perfect.
I'm confused, why would the zoom affect a point calculation?
(setq ds2 (/ dis1 (cos (/ an3 2.0))) ;distance of offset.
ptx (polar p1 (+ (+ an1 ang1)(/ an3 2.0)) ds2) ;offset point from vertex at 1/2 included angle.
cnt (1+ cnt)
)
Even with that weird point, why would (vlax-invoke Outer "IntersectWith" $cv_ray acExtendNone) return a weird list on that point?

Bill

 

Bill,

See John's thread subject: @Inside "Look, Ma... No Rays!" dated 9/15/2003.

Joe Burke

 

Bill (and thanks Joe for the history):

I never finished it. Rays are NOT the answer. We had poor Doug chasing his
tail and I still feel bad about it.
I'm so blasted busy these days that I can't even play here too much. But please
feel free to pester me (even by e-mail; hell eveyone else does), and I'll try to
wrap it up. The real solution was to sum the deflections around the perimiter
(I know you know about deflections), but there were issues with polylines vs.
ellipses, I think. Then again, who needs ellipses? My PELLIPSE variable is set
to 1 so I can offset the blasted things and maintain the equidistance (izzat a
word?).

0.0). This is possible when the two segments that I use the @delta function on
to get the deflection angle return a very sharp angle (near zero).

acExtendNone)) returns a list of about 11 points.

inverted corner) So the ray should not cross more than 2 or 3 times.

thought I should still say something in case others have trouble.

 

John,

Glad to hear you've been blasted lately. I was beginning to wonder if we were simply
boring you.

Yup, that's a word.

Joe

 

Found time to play.
Check out the latest @cv_inside.lsp (attached).

0.0). This is possible when the two segments that I use the @delta function on
to get the deflection angle return a very sharp angle (near zero).

acExtendNone)) returns a list of about 11 points.

inverted corner) So the ray should not cross more than 2 or 3 times.

thought I should still say something in case others have trouble.

 

John,
Can you send me the new @cv_inside via email?
They have so protected with anti-virus and anti-spam software right now that I can't touch the attachments.
Or try posting in CF.
Thanks
Bill

 

Forget that last post John. I got the routine.

Thanks

Bill