Re: stack overflow at compile time

> In the absence of recursion or non-trivial dynamic allocation, bounding
> stack size at compile time is obviously a straightforward problem:
> generate the call graph, label each arc with the stack size of the routine
> being called, and find the greatest "distance" from the root to a leaf.
> The maximum may occur on a path that's actually impossible, but at least
> you do get a bound.
>
> As a practical matter, you'll need either compiler-linker cooperation or a
> compiler that sees the whole program (including all libraries), but that's
> no big deal.
>
> I have a dim recollection that some commercial compiler systems for
> microcontrollers (some of which have quite small fixed-size stacks) do
> in fact do this.


The occam 2 compilers definitely did/do this. In their case, "stack" was/is
synonymous with the tranputer's "workspace". The compiler does consider
overlap properly on the call graph.


Jan

Return to the comp.compilers page.
Search the comp.compilers archives again.