Stopgap for #8862

[roed]

There has been no progress reported by Maxima on #8862 in the last year: ​ 3216684. I don't know exactly where we should add the warning: every time solve is called? Can we narrow down types of input to maxima that trigger this issue?

[kcrisman]

I doubt that this is possible to stopgap very well. I mean, the only way to really tell is to solve the equation! It's probably unsolvable in some appropriate sense even to determine when such missed solutions might happen.

I don't think it's appropriate to always add the warning. That would imply solve is totally unreliable, which isn't true - it is often quite good. But having a documenting of the bug seems appropriate. Maybe I'm already forgetting how stopgaps are implemented, though.

[roed]

Stopgaps are implemented by just showing a warning to the user (but only the first time that a warning for a given ticket would be triggered). At whatever point in the code where it's appropriate, you call stopgap. See #12691.

I would also like to avoid displaying a warning every time the user calls solve. But the point of stopgaps is to warn the user when they're doing something that might yield known incorrect results, and point them to a trac ticket where the issue is discussed. I'm not sure what the right solution is here though.