Bug #4123 - MacOS "Datum (whatever dimension) mm for the constraint with index (whatever constraint) is invalid."

[GeneFC]

From the console message it appears there is an exception which indicates some sort of unhandled situation in the underlying code.

Since you are a long-term user you are probably well aware that the error trapping and notification in the Sketcher is much more rigorous in the current versions than in the past.

The overriding rule now is, "Stop! Errors ignored will only cause increasing problems."

The message might be a bit clearer, but ignoring the error and charging ahead just does not work. In this case the problem happens before the attempt to add a vertical length constraint.

Gene

[steve123]

Agreed. I am simply bringing attention to the need to clarify the error message when someone has an opportunity to do so.

[steve123]

ignoring the error and charging ahead just does not work

I never interpreted the redundant constraints as "errors" ... is that in fact the case? They should be considered errors?

[chrisb]

I never interpreted the redundant constraints as "errors" ... is that in fact the case? They should be considered errors?

You should definitely remove them. They can prevent the sketch and subsequent features from updating.

[GeneFC]

I never interpreted the redundant constraints as "errors" ... is that in fact the case? They should be considered errors?

Redundant constraints definitely can hinder the solver, or even prevent it from reaching a solution. That is why they are considered bad.

Error?? I guess you can define it anyway you like.

Gene

[steve123]

Thank you for the clarification. I think in most cases I always fixed the redundant constraints. But in a few cases I "did not agree" with the solver's view that constraint was redundant Of course, I was wrong.

Notwithstanding my wrongness, there is usually a strategy I am attempting to use when constraining the sketch. Is there some sort of "priority" or "order" to the constraints? Sometimes I prefer to constrain a different way than the solver would appear to want to. So, understanding a bit more how the solver deals with the constraints might help my planning them.

[GeneFC]

Sometimes I prefer to constrain a different way than the solver would appear to want to.

That is the reason that "Auto remove redundants" is a checkbox. It seems to work correctly, but it has no way to know what is important to the user.

I am not the least bit qualified to talk about the Sketcher solver, but a simplified way to think about it is that it essentially tries to solve a system of linear equations. As you may recall from Algebra classes strange things can happen when two or more equations are co-dependent or are conflicting. Similar things happen to the Sketcher solver when there are conflicting or redundant constraints.

No doubt there are cases where the solution can be just fine when there are redundant constraints, but it is not always true. There have been numerous strange cases on the forum over the past few years that were eventually found to be due to bad constraints. As noted above, the Sketcher is much more careful now about allowing the redundancies to remain.

Gene

[steve123]

Thanks for taking the time to provide the feedback. I see your point regarding the solver. I'll be sure to be more careful about the redundant constraints.