preview tolerance of line segments

[Julian Todd]

probably by plotting all the node points in advance

[timswait]

This could be really useful, together with some way of checking and telling me about errors in outlines. It's very easy to end up with a tiny gap or overlap in an outline which makes the laserplacer refuse to group it but is very time consuming to locate in CAD. If the laserplacer could say "there's a gap here" or "there's a double line section here" it would save A LOT of time finding the mistake. Even better would be an option to fix it in laserplacer (e.g. fill the gap with a straight line or delete duplicate lines), but just identifying and flagging the reason it won't group would be a big improvement.

[Julian Todd]

Duplicate edges seems to be a pretty common thing in dxf files. I've got some good examples.

It's a better form to get the problems fixed upstream than bodging stuff downstream.

[timswait]

True, but if it would at least tell me where the gaps/duplicates are then it makes them much easier to correct upstream. Last set, I spent ages on some of the more complex panels trying to work out what the problem was that was stopping it grouping. Especially if there was more than one issue. I would find what I thought was the issue, fix it, convert to SVG, reload into laser placer, it still wouldn't group so I'd have to do it again.

[Julian Todd]

I'm going to start on this one first, as it is the least fun and probably causes the most frustration

The plan is to create the capability to break down the function function ProcessToPathGroupings(res, rlistb, closedist, spnumCSP, fadivid, elprocessstatus, groupingprocessFinalize)

into it's component stages, and show them happening step by step.

PolySorting.FindClosedPathSequencesD() GetSingletsListCSP() MakeContourcurvesFromSequences() PolySorting.FindAreaGroupingsD() PolySorting.FindPathOrientation() PolySorting.SingletsToGroupingsD()

We have a sort of crude toolpath preview system to show the order of cuts in the finished output, so it should be possible to replay the closed path sequence algorithm to find the points of failure in some way.