Section plane in flipped component or group reverses cut direction unexpectedly


#1

Hi,

This is my first post here. I raised this issue on Sketucation, and the consensus was that this is a bug that has been around a while. I’m using SU Make 2015. The issue is that when a section plane is applied to or in a component and that component is flipped along an axis that should not change the cutting direction (as indicated by the arrows when in context), the cutting direction is nevertheless reversed. The same goes for groups, but obviously with groups you can correct that behaviour. I had hoped to use section cuts to make non destructive edits to components so that the same edit could be applied to a swapped component. This issue makes the approach less viable.

In my example, I was creating a 110 degree mitre joint on two copies of a component which is a profiled piece of timber to make a roof verge on either side of a shed. The containing component was the entire side of the shed, including the frame and the outer covering. The two sides are symmetrical and one is flipped. I wanted to keep them in synch. On a small scale it’s not too much of a hardship maintaining two groups to be virtually identical, but in a more complicated model this could get out of hand. Many thanks.

Regards,
Matt


#2

For example if you mirror a entity reversials happen, been that way since at least version 6 not a bug just a design artifact.


#3

I wanted to share this link with you, but then I realised that this might be redundant…


#4

[quote=“mac7595, post:2, topic:10037”]
…been that way since at least version 6 not a bug just a design artifact.
[/quote]Could you explain why this isn’t a bug and how we are supposed to deal with section planes inside a mirrored group or component. I’m just curious.


#5

Here’s the sketchUcation topic I posted.

http://sketchucation.com/forums/viewtopic.php?f=15&t=61586&sid=cdb31a3129892830aea2e6b6839c00f3

It contains a sample SKP file and a screen shot. When you flip a component or group, certain reversals do happen. That’s why you flip it after all. The other axes will maintain their integrity. If you flip along an axis that bears on the section plane, then of course the section plane will be affected. If you have for example a section plane that is on the red and blue axes with the arrows pointing along the green and you flip on the red or the blue, that should not affect the cut direction. I certainly wouldn’t expect it to, would you? But hey, if it did then I’d at least like to see the arrows change direction.


#6

Read my post only addressing the fact reversal has been around for long time. This is one of first issue I hit when starting and mortise and tenon joints surprised me. I know the issue and stay away.
My question is why you say this is bug but there other design artifacts you are mute about; What about small face formation at small dimensions; What about poor inferencing when one tries to do rotation intersect line with sphere.: What about …?
Why is there not a list of each of these so users know what to expect and do not have the time/ expense of finding out by surprise.


#7

I called it a bug because someone asked me about this specific weird behaviour. Editing context and being one level higher do not represent the same geometry and also selection options differ from what can be expected. (you can select an invisible side, yet you can’t select the visible side of a sectioned mirrored group or component).
Am I mute about other design artifacts? I didn’t know I was.
About small faces not being formed, … SketchUp wasn’t designed to do that. Once you know it was designed for larger geometry you can live with that. I hope one day it may be different.
Rotation edge (keeping its length!) to intersect another edge is something I mentioned often enough.
This has been fixed in the latest version(s), for when both edges are in the same context.
Intersection with a sphere? Unfortunately SketchUp does not have true curves and spheres. I wouldn’t know how to answer that question. The skin is a bunch of faces, not one smooth surface.


#8

“Design artifact” is sort of a euphemism, isn’t it? The design of the section plane functionality is incomplete or incorrect. For example, hidden parts of a cut component or group (and maybe raw geometry) cast shadows. Isn’t that a bug? So is this. It seems developers decided that it’s too hard to work out whether or not a section cut should be reversed when you flip, so just reverse it regardless. If the flip axis has no perpendicular projection onto the section plane, then it shouldn’t be reversed. The behaviour would be less weird if the arrows actually went in the direction of the cut after it becomes reversed. Then it would seem as though it was by design, even if the design is questionable. Unlike some of the other artifacts you mention, this one can’t be worked around when dealing with copies of a component. Maybe section planes were not designed to be used for component shaping, but it would be really powerful to be able to use section cuts this way. One of the many benefits would be reduction of file size, as more components could be reused with this sort of non-destructive edit. You just need to add faces. Even better would be if the section plane could add dynamic faces for you. That would be a new feature request though. This is more fundamental.


#9

And so is a “bug”. As noted it is not my “bug” but yours. That is reason use of that term has been phased out in cases. If you want to make case about file size then present some data vs. generalities and same with user support. You must have some skp file you can post showing the real increase in file size etc. Hopefully for you the case of "make play " can be made to supplier.


#10

Intersection with a sphere? Unfortunately SketchUp does not have true curves and spheres. I wouldn’t know how to answer that question. The skin is a bunch of faces, not one smooth surface.
Thanks there was posting on closed form solution for that. In fact there are five possible cases that have to be considered.
So your point is if SU was not designed for that,so we should accept it??. I have seen that about faces many times but not once have any hard facts been presented. It fact it has been stated it was a OpenGL issue.
[/quote]


#11

Mac, are you a Sketchup developer? If you work on Sketchup code, then a bug is more yours than mine. You can decide not to work on it for one reason or another, but it’s not really fair to deny its status. I take it you are not, from your remark about making a case to the supplier. I looked around for a bug tracking forum for Sketchup and was unable to find one. This is as close as I could find. If you are not a developer, then I don’t expect you (or anyone else necessarily) to take an interest in my findings. I have however found a bug, and I would hope that Sketchup developers would take it seriously. I like Sketchup Make a lot and I’ve thought about a Pro licence, but I’m not about to purchase something that doesn’t do quite what I want owing to some design limitation or oversight. I don’t think bug reports should be discouraged. The overall aim is to improve the quality of the software. If there is a better forum in which to raise this issue, then do please enlighten me. Thanks.


#12

No I am not and do not want to be.
Too many times folks state some thing is a bug w/o what I consider hard proof of that statement. I visit Sketchucation very often and do not remember the issue discussed there ( user as mac1 BTW maybe close to 1000 postings). Has someone there filed a bug report.). There are some very good modelers there and if you posted your model maybe they have a work around for you. One thought when reading your post is possible use of the SU capability of replacing one component with another. If there is any reasonable reduction in file size you should present the data other wise it is arm waiving. Trimble will have to expend $$ to under stand required changes, do the design, run regression test etc, etc.
so what may appear good for you may not be for them from business view point. Some time I speak from my technical manager view point and I am very big believer in regression testing. That has to be a cost driver for Trimble. Many times what is "better’ is determent to good enough.
I’ll look for your posting on Sketchucation.


#13

Hi. I referred to my SketUcation post earlier, as did someone else. Gerrit (Wo3Dan) has also confirmed that this is a problem that he has been aware of for some time. He also brought up the weird issue with reversals happening when you traverse up the context tree and with being unable to select the visible part of a section plane, but clicking on the invisible part selects it. I noticed this as well. It is weird and I don’t think it is intentional. Hidden parts of a section cut cast shadows. This all points to a buggy implementation of the section plane functionality. I’m not one to tell a software vendor what is important, but I’d think they might want to consider rewriting the section plane code. The other stuff about reducing file size was just a carrot. I’m thinking there is enough stick here. I’m looking forward to hearing from a developer at some point. There is a bug. There are in fact several, possibly related. They may want to fix them. I hope they do. If they do not, I’d appreciate some indication as to why. Higher priority items are understandable, but let’s see if we can get this on the roadmap.


#14

flayman:
Bad memory I did make comments to original post.
I need to spend more time on approach, but I keep asking my self why it is desirable to follow your approach which makes the model more complicated , I doubt saves file size because of additional edges created and the increased complexity will make it more prone to model errors.
I am sure you have heard of the term KISS.
If you have not found the outliner you should be using that. It shows a graphical picture of the model tree which I find very useful. When making you component and groups using meaningful names should help you.
Additionally I did find the selection issue but using the outliner over comes that very easily.
I am some what lost on the shadow issue so will look at that more.
The view direction of the section plane does not cause me any difficulty and I guessing this was designer choice. When one starts mirroring things etc. it can be easy to create a left hand coordinate system and I am sure a lots of folks would not like that. I need to look at more though.
You stated in you post you wanted the two verge to look same it appears all you have to do is reset scale on the back one?