ExtendSrf - Why are extensions different lengths?

Rhino
1

image
image1893Ă—1092 338 KB

ExtendSrf Different Langths DC01.3dm (2.0 MB)
Version 8 SR30
(8.30.26090.13001, 2026-03-31)

Simple single span, untrimmed symmetric rational surface. Degrees 2 x 1.

ExtendSrf opposite edges by the surface by the same amount and the extensions are different lengths. Because the surface is untrimmed and symmetric I expect the extension lengths to be the same. (Results are the same with Merge=No or Merge=Yes)

This looks like a bug.

Command line history:

Command: ExtendSrf
Select an edge to extend ( Type=Smooth Merge=Yes ): Merge=No
Select an edge to extend ( Type=Smooth Merge=No )
Extension point <2.000> ( SetBasePoint Type=Smooth Merge=No )
Command: ExtendSrf
Select an edge to extend ( Type=Smooth Merge=No )
Extension point <2.000> ( SetBasePoint Type=Smooth Merge=No )
Command: ExtendSrf
Select an edge to extend ( Type=Smooth Merge=No )

Updated (Credit to @sjb for identifying the problem as dependent on surface location.)

The extension lengths of rational surfaces created by ExtendSrf is dependent on the location of the surface. Non-rational surfaces work correctly. This bug identified a long time ago. A YT item was created over 8 years ago and assigned to Greg Arden, who has since retired. It needs to be assigned to someone else with priority. https://mcneel.myjetbrains.com/youtrack/issue/RH-41887

The results of ExtendSrf should be totally independent of the surface location. In addition if opposite edges of a symmetric surface are extended with the same “extension point” the resulting extension lengths should be the same.

Example file of symmetric surface in rational and non-rational forms in multiple locations.
ExtendSrf Bug DC10.3dm (2.1 MB)
Rational surfaces are blue, non-rational surfaces are green:

image
image1920Ă—1121 456 KB

2

The _ExtendSrf results appear to be location dependent. For example, copy the original surface to another location and extend both it and the original the same amount.

The inconsistency doesn’t occur when the surfaces are non-rational.

_DimCurveLength output shown:

image
image607Ă—433 94.9 KB

ExtendSrf_Rational_Location_dependent.3dm (98.0 KB)

1 Like
3

Thanks. I’ve confirmed that behavior. This definitely looks like a bug. Surface position should have zero effect on extension length.

4

In my experience, this is one of the commands that’s always bug: in one version it doesn’t extend properly, in another build a border doesn’t work, in another it has some other anomaly. You can’t set fixed points. Yet I believe there are other, much more complex commands that work better.

5

Bumping this thread so someone from McNeel will respond.

6

like you have seen this is a very old bug. I’ve reassigned the issue