... it is theoretically impossible for Perl 5 to be tail-recursive.
I disagree with this, or at least with the sentiment I inferred from it. While tail-recursion optimisation could not be turned on by default for the reasons you cite, I see no reason that this and other optimisations could not be introduced optionally by means of lexically-scoped pragmas.
In principle, a more complex approach would also be possible: always to perform the optimisation, but to retain enough information to emulate the theoretical caller stack on demand. I wouldn't recommend this approach though - for a simple recursion, you'd probably need to retain only the number of levels of recursion to emulate, but it would be much more difficult to cope with even slightly more complex cases, such as a subroutine that called itself from two different places, or a pair of co-recursive subroutines.
HugoIn reply to Re: Re: Iterative vs Recursive Processes
by hv
in thread Iterative vs Recursive Processes
by mvaline
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |