dbickin wrote:This looks like exactly what I want.
This Pascal solution implements a complex custom type for the RPN stack, which makes the solver complicated. IMHO it's better to operate with two stacks (value/RPN).
This answer confuses me.... not too hard to do.
I suspect because you are answering a question I never really asked. Frankly I don't care what kind of stack the solution uses behind the scene.
The given link was to the RPN facet of the pascal class, but I was looking at the whole thing: namely a spreadsheet component that I think lets me embed it in a form, does functions, and other things a good spreadsheet would do.
It sounds like what I want. (Actually sounds too good to be true. Testing will tell.)
So, were you saying the way they do RPN is not as good as the way you would do it, or that the way they do it simply DOESN'T work, or that you have tried it and it is too slow to be useful?
No one has piped up if I would be somehow able to use this control, which I gather is a part of LCL, from a freebasic program. From searching the forum here (and slightly the freepascal forum) it sounds like the answer is no. (But the most solid answers were from as far back as 2011 so that may have changed.)
If the answer is No, then unless I want to abandon the code I have already written and rewrite it in pascal, the FPSpreadsheet is of no use to me.