[Feature idea] Sketcher L U and H shapes

[Shalmeneser]

L = rectangle + inside corner
U = rectangle + inside border
H = rectangle + 2×inside border

New menus :
* Fillet / Chamfer / inside corner
* Trimming / trimming-blue / inside border

trimming = replace part of line with nothing
trimming-blue = replace part of line with construction mode line (to keep maximum properties of the original line)

[adrianinsaval]

I meant outline, closed shape.

ok shalmenesr's image gave me the wrong impression (althought it wasn't clear in the text either). This makes more sense, H shape sounds overkill to me, L is more sensible, I assume it would be 4 clicks to define it or 3? How many for the U shape, 3? Also, are we talking shapes with rounded corners?

Perhaps it's not my place to comment on the maintainability since I'm not at all familiar with the sketcher code so I'll shut up about it.

Now, just my 2 cents, IMO it would be more productive to spend some time learning git and organizing your other tools than to work on these corner cases, would you consider polishing those before tackling this? Of course, I understand your time is your time, you're free to choose how you spend it, just my suggestion and my wish since I do want to see those others merged as they would be useful to me. Not trying to be demanding.

[chrisb]

Chrisb must be busy.

And so I am indeed (started at 6:30 with a lecture, online meeting at 14:00, next meeting at 16:00, having guests at 19:00, next lecture again at 6:30, ...)

Nevertheless: Perhaps the Polyline could have a special V/H submode, where simply only vertical and horizontal lines are drawn.
For editing such tools would help quite a lot, but I don't think that the use case is common enough to justify another tool. That would indeed rather be a macro.

[paddle]

It would be 3 clicks for all and not rounded corners. Though this could be an option but it makes things more complex.

Anyhow it's not such an important feature. I prefer to get offset and circular pattern working as those are much more useful. I might try to slip those L and U discretely later

V/H submode is an interesting idea.

[chrisb]

Anyhow it's not such an important feature. I prefer to get offset and circular pattern working as those are much more useful. I might try to slip those L and U discretely later

I would wait with implementing such features until Abdullah has implemented layers - or help him implmenting these first. Then it would be sensible to pattern such layers. Patterning with the current possibilities is either a horrible clutter of constraints or not really helpful as a parametric pattern.
Just my 2 cents.

[paddle]

Anyhow it's not such an important feature. I prefer to get offset and circular pattern working as those are much more useful. I might try to slip those L and U discretely later

I would wait with implementing such features until Abdullah has implemented layers - or help him implmenting these first. Then it would be sensible to pattern such layers. Patterning with the current possibilities is either a horrible clutter of constraints or not really helpful as a parametric pattern.
Just my 2 cents.

Thanks for feedback. I'm not sure what is layers and how it will help. But yes you're right, better discuss with him first. Maybe he even already started those.

[obelisk79]

I've been following the conversation, and while I often applaud new features, I'm just a bit skeptical on this one. The polyline tool itself is just too good already that I feel like this is approaching fluff territory. If anything, perhaps an optional macro for such shapes could be useful.

Now, an option mentioned that I think there could be some great value in would be the blue trim tool, although I'm not sure that's a great name for it. Perhaps "constraint preserving trim tool" or "construction trim". Maybe this should be discussed in a separate topic though.

[adrianinsaval]

regarding patterns, IMO it would be better if the positions of the patterned shapes are not determined by the solver but by a separate transformation function and they are handled as blocked geometry. This would make them easier to use as they won't dramatically increase solve time.