3 ms·
I don’t understand that argument. All function arguments still would be on a single stack (just as in Forth). That would only be a valid argument if some FORTRA
by Someone 6y ago
I don’t understand that argument. All function arguments still would be on a single stack (just as in Forth). That would only be a valid argument if some FORTRAN code inspected the return address, for example to behave differently, depending on the caller.
I would guess it’s more because early systems had small address spaces. If your heap grows up from the bottom of memory, and one stack grows down from top of your address space, how do you know where to place that second stack, especially in a system with, say, a 64kB address space?
- JoeAltmaier 6y agoYes, exactly. Some FORTRAN code did exactly that. And the stack could go in a page not addressable by general purpose instructions, save call and return instructions (and perhaps some kernel debug). Some limited-RAM embedded devices have dedicated stack RAM. But my point was about Intel x86