06:58 <guso78> joseph_, the Only difference between PolySetIndexed and PolySet is that the points of the Variables not coordinates, but indices, which you can look up in an additional step. This change does yet yield the ability to "select/mark" faces.

06:59 <guso78> Rather it enables the ability to select/mark vertices as you could define a "select" property to each vertex. but this would be an additional step.

07:02 <guso78> My proposal is actually to make the "interface" more transparent by using some kind of setter/getter function which yields the "object" in either PolySet or PolySetIndexed, or other representation(whatever the algorithm can better process at this stage.

07:06 <guso78> Hardly changing between PolySet and PolySetIndexed is a heavy change IMHO which affects all OpenSCAD code and it will be even more hard to get it reviewe/accepted with good conscience. This is why I like to see both types present simultaniously so code can see a smooth transaction.

07:09 <guso78> teepee, i believe that c++ can check for the actual object type in an operation. can you explain which information your desired flag will contain ?

07:13 <guso78> yes, I have seen very weird code in GeometryEvaluator.cc where objects are converted to "asMutableGeometry" . Is this the "In-Place-Editing" you refer ? Due to lack of understanding , whats going on i got my functions working by mainly blindly copying the routines.

07:14 <guso78> teepee, i am lacking a complete overview, so i dont know if any code would appreciate PolySet in favor of PolySetIndexed, and yes: I also vote for step-by-step translation

07:16 <guso78> In order to progress I'd like to 1st create the required environment and after that use it step-by-step. My primary question is basically function names, and places where to locate the staffolding additions.

12:52 <kintel> guso78 I think _creating_ geometry often only has access to non-indexed geometry, and we use PolySet for that purpose and we don't know how many vertices we'll end up with.

12:53 <kintel> One improvement could be to have an efficient/ergonomic way of building a new polyset. Such a builder could just be a separate factory class from the actual polyset, in order to keep it slim.

12:55 <kintel> same for conversions really. Conversions can be expensive, so it would make sense to build components which makes it obvious to the calling code that they are expensive.

12:56 <kintel> Building such things "bottom up" is nice, but it's sometimes hard to keep the use-cases in mind. I usually approach such things using a "tracer bullet" approach:

12:58 <kintel> Start with the client code; where you wish you had a PolySetIndexed class, and write/modify that client code to do what you hope to achieve. Once you get a sense of the API, scaffold the new components and perhaps even hack in a quick and dirty implementation, to validate that it would actually do the trick, and if you're lucky even work just very slow or just with a hard-coded object.

12:59 <kintel> ..then look around on all the other uses of PolySet and see if you find some patterns or inconsistencies

13:00 <kintel> ^ my approach may not be the most efficient in the short term; some of my deep refactorings take months : /

14:03 <lf94-> guso78: you may have sent messages to lf94 - seems my bouncer disconnected and ive been up in the woods for a week x)

14:12 <lf94> Fixed

14:16 <guso78_> lf94 probably your bouncer killed your existance because i have written you some some things about fidget you did not want to realize. Actually i believed you went to vacation as its summer time.

14:19 <lf94> Ok good x)

14:20 <lf94> I figured you did :p

14:20 <guso78> haha

14:22 <guso78> znc log appears to miss some history ... is just this time ?

19:39 <guso78_> when i connected my laptop at home to my kvm switch,  neither keyboard nor mouse were working, only touchpad and laptop keyboard. all other ports were working, just not my laptop. i tried to exchange cables, ports .. connecting without docking station, nothing works. Finally I resignated and got another pair of keyboard and mouse and connected it

19:39 <guso78_> to the laptop direcly. did not work either. So finally it was time to reboot. When rebooting, i realized, that windows was about to complete the "update".  Windows has already started to exchange driver files in the background without letting me know and needed a final reboot to sync up everthing. I LOVE Windows!