(Bug) mirror subselection changes geometry / trimming

ups…
:cold_face:
sometimes simple things to wrong, thought i do a nice object for beginners tutorial…

Gastherme_Griff_B_00.3dm (3.5 MB)

Version 8 (8.23.25251.13002, 2025-09-08)

sub-select the violet Faces (per face color, same bug without per face color)
mirror
the mirrored surfaces have degenerated trims…

@Gijs - can you put this on the pile ?

thanks for a fix.

(workaround _extractSrf copy = yes mirror …)

kind regards - tom

to make it more beginner friendly, maybe increase the radius of this fillet so that the smaller fillet doesn’t end on the seam between surfaces: (here R15)

that being said, I was poking a little more and wonder why you would need to mirror these surfaces to start with. If I go to this step and then FilletEdge, things work normally:

i use this exercise to show how to construct with lines and arcs - and take care on Fillets already earlier.
i want to sharpen the eyes of the students to see those dependencies.

the way i want the students to construct the initial curve is to draw an arc tangent to the point, then a line tangent to this arc…

i ve this is to nit-picky - tell my - but i would love to see other options (@gustojunk ? or @davidcockey ?) before i show the students the cheating approach.

still think its a bug - wether the workflow is valid or not.

kind regards -tom

if you want the top to be a planar surface, then why create the revolved surface in the first place?

If I look at the construction of that form, I see this:

The mirror bug doesn’t happen in Rhino 9 WIP

RH-89458 Mirror creates bugged object

thanks.

can’t test this anymore - on a intel - sorry

Hi Tom, Since you asked… I think we need to stop teaching students developing form with degree 1 (planar sections) and degree 2 (fillets) surfaces only.

I interviewed someone today whose work looked promising, but in closer inspection it was all prismatic volumes pushed-pulled with fillets in Plasticity with fillets and booleans. It will take years to unteach those unsophisticated bad habits.

This is how I’d handle this exercise, and how I’d teach it to students to they are more useful, more employable.

below left: your original cap, middle blending the surface fillets (still shows weird next to completely planar sections), right, everything has slight curvature:

This is what my curves look like:

emap after also adding BlendEdge instead of fillet

Matcap shows this better.

This is what’s happening:

So in summary, I hope that better fillets in Rhino are not the reason students stop creating form in Rhino with care.

I hope this helps,

G

that s a great insight. many thanks. really appreciate.

I teach in a huge variety of backgrounds - all students already have a first job / apprenticeship / education. Only a few will become “hardcore” industrial designers. Some of them will stay carpenter or metal worker, but shift from the workshop to the planing / work preparation.
I think they should learn to construct proper G1-Lines and Arcs and understand where curvature needs to change … this is where is exercise is positioned. - for example understand the concept of a line being tangent to an circle (that was drawn first, not a fillet on a existing polyline)…

But I totally agree, that a few lines, arcs, extrusions / surfaces of revolution + push/Pull + fillet will not make great design in most scenarios.

have a nice day - kind regards -tom