PDA

View Full Version : Modeler: select by normal crashes



spigolo
03-02-2013, 11:09 PM
Hello,
is it just me?
Everytime I try select by normal in lw 11.5 with a polygon selected in order to tell what kind of normal i want modeler crashes...
It works if you use this tool indicating the kind of normal you want.

Waves of light
03-03-2013, 01:59 AM
Not here, I'm afraid.

1. Does this happen on simple geometry... so a newly created Ball?
2. What OS you on

spigolo
03-03-2013, 02:10 AM
this happen on existing geometry (complex one but not so complex)
I'm using 11.5 on win 7 64 bit on a new esacore 3930k

jeric_synergy
03-03-2013, 02:11 AM
It works if you use this tool indicating the kind of normal you want.
??? Which tool again?

gerry_g
03-03-2013, 04:06 AM
tried it on pre existing objects with counts of 280,000 & 850,000 polys having first selected a bunch of quads and worked fine (on a Mac running snowLeopard) but these were entirely quadratic objects not tri's or ngons, could this be where your problem lies, be more specific about your test object

Waves of light
03-03-2013, 04:11 AM
This is what I was getting at. Can he reproduce it with simple geometry or is it specific to the model he is working on.

Is it because the current model he is trying to do it on has unclean geometry (two point polys, etc.)

Andy Webb
03-03-2013, 05:53 AM
This is what I was getting at. Can he reproduce it with simple geometry or is it specific to the model he is working on.

Is it because the current model he is trying to do it on has unclean geometry (two point polys, etc.)

But should it matter if the geometry isn't clean, surely it should just ignore that geometry.

While I'm all for clean geometry, that doesn't mean tools have to not work, they should have built in code to deal with it.

Waves of light
03-03-2013, 07:10 AM
But should it matter if the geometry isn't clean, surely it should just ignore that geometry.

While I'm all for clean geometry, that doesn't mean tools have to not work, they should have built in code to deal with it.

Rounder - enough said.

E.g. what if the 'select normal' tool (and I'm no programmer) includes two point polys as part of the process when it's looking for polys, yet they don't have a normal. Therefore crashing the process and the software?

Whereas it would be nice to just assume that all functions and tools work under all circumstances, we know that is not the case. And all I'm trying to do is put in place some controls, that may eliminate the user's issue. If he can make 'select normal' work on simple geometry, but not on complex geometry, then that should point to his geometry?

Andy Webb
03-03-2013, 07:40 AM
But what if, instead of crashing, it gave a warning that some polygons in the mesh were 2 point and could not be selected.

In other words the tool was more informative and also avoided crashing.

Waves of light
03-03-2013, 07:58 AM
But what if, instead of crashing, it gave a warning that some polygons in the mesh were 2 point and could not be selected.

In other words the tool was more informative and also avoided crashing.

Don't get me wrong, that would be nice... and in some circumstances (and functions/plugins) that is true e.g. Press F4 without any geometry selected and you get error prompts and advice on what to do. But I know this isn't the norm for other functions.

Only recently, on another thread, Lewis was talking about how he had a script that cuts out all single points and two point polys, so that certain plugins/functions/workflows worked.

Anyway, our desire for alert changes is going off topic a little and probably not helping Spigolo find a solution to his problem.

Here's hoping someone with a little more knowledge of the 'select normals' functionality/programming than I will be able to sort this one out.

Andy Webb
03-03-2013, 08:31 AM
Yes I saw Lewis's thread which prompted my comments here.

Matt wrote a select by normals script sometime ago, I wouldn't be surprised if the new plug wasn't based on that.

Waves of light
03-03-2013, 08:36 AM
Yes I saw Lewis's thread which prompted my comments here.

Matt wrote a select by normals script sometime ago, I wouldn't be surprised if the new plug wasn't based on that.

Yer, that was it. Matt wrote a special plugin for him.

I don't know... if it ain't broken, don't fix it... but if it kind of works, then do what you need to make it work. Blimey, I remember when extend (or extender) left you with an adjoining unwanted poly between your extended points selection... you just had to delete it each time you used the function (which was a bloody lot).