Add an example of compareable solvers in a README?

[dimsuz]

I've only now watched Adam's presentation about Contour and while I was using it for some time now, this slide blew my mind

I've been occasionally struggling with the fact that sometimes I want one solver and sometimes (depending on view config) another. As a result I had to invent some clever tricks, while I didn't know about this feature.

Can it be somehow made more obvious? Perhaps a mention in README or in some other documentation source?

[dimsuz]

Another feature I constantly wish to have is having maxOf(), minOf() being available not only for 2 parameters, but any number of them (varargs?). This can be worked around with maxOf(xint1.value, xint2.value, xint3.value).toXInt() but feels hacky. If this is something possible I can create another issue (and perhaps a PR later).

[saket]

This would fit very well with #87.

Another feature I constantly wish to have is having maxOf(), minOf() being available not only for 2 parameters, but any number of them (varargs?).

I agree, but adding new solvers may not be very scalable. It'd be really nice to have an axis solver of type () -> XAxisSolver so that people can mix any kind of lazy-evaluated solvers depending on view config. We haven't felt a strong need for this yet because it's still indirectly doable (as you mentioned) but I would be interested in seeing this happen.

[dimsuz]

I now feel I shouldn't have mixed min/max in this issue. I actually was referring not to minOf/maxOf on solvers, but rather their overloads on XInt/YInt. I sometimes find myself having not 2, but 3 or 4 views where I want to write

view1.layoutBy(x = leftTo { maxOf(view2.left(), view3.left(), view4.left(), ...) })

As for solvers probably 2 would be the most common case here, so I agree that this perhaps has to be still proven as really needed to make it into a library...

As for the issue, if it's covered by #87 then this one should be closed, Recipes is a perfectly good place for this stuff! :)

[saket]

Sorry I forgot to follow-up. Yeah I wouldn't be against adding more overloads if others are running into it too. For now, I'd recommend maintaining an extension function for your project?

[dimsuz]

Yeah, I did extension function for minOf(XInt/YInt), like this:

fun ContourLayout.minOf(y1: YInt, y2: YInt, y3: YInt): YInt {
  return minOf(minOf(y1, y2), y3)
}

but unfortunately I can't do the same for YAxisSolver, because there's no minOf overload which accepts access solvers (which would be produced by the inner minOf above)