Re^2: Perl Cannot Be Parsed: A Formal Proof

[Jeffrey Kegler]

in reply to Re: Perl Cannot Be Parsed: A Formal Proof
in thread Perl Cannot Be Parsed: A Formal Proof

In a series of three articles in The Perl Review, now available online, I laid this proof out much more carefully, in three different versions. The most bullet-proof version derives the result directly from Rice's Theorem.

You are right that the proof would not go through if the only way of establishing a prototype for a Perl function was via a function definition. As I said in Perl Review (4.3, p. 28):

I must show that I can use Turing-complete Perl code to determine the prototype of the dunno subroutine at compile time. A function definition won’t work.

Summarizing from that article, there are at least two ways that the prototype of a subroutine can be established using Turing-complete Perl code. One is with symbol table manipulation, for example:

BEGIN { *dunno = sub () { 3 } }

A second way is to put the function definition into a string which is eval'd in a BEGIN block. I believe clever monks will be able to think of others.