The idea is that some sort of declaration can
change the meaning of subsequent code.
Pragmas do it:
use integer;
print (0.3 + 1.2);
The + sign takes on a different meaning
when use integer is in scope. It does
integer addition rather than floating point addition.
use strict is a pragma which restricts
the meaning of your code, rather than actually changing it.
If the strict 'refs' pragma is in scope, then when
we see an expression like $$foo we can be sure
that $foo is a reference rather than a string.
An ambiguity has been eliminated by prohibiting one
of the possible meanings: symbolic reference.
Pragmas can only work by co-operating with the perl core,
but...
It's possible, to some extent, to write pure modules which
can change the meaning of code. I'll call them Mutagenic
Modules.
There's a hotchpotch of
mechanisms in Perl, each of which allows us to change the
meaning of a particular type of operation. We'll look at
some of these in the next section.
Dave Cross's Symbol::Approx::Sub is, of course, a
mutagenic module. So is Abigail's Pythonesque.
Brian Ingerson's Inline and Damian Conway's
Switch are both examples of modules which add
extra syntax to the language, but couldn't ever change
the meaning of existing programs. So I'd say they aren't truly
mutagenic, and I'll call them language extensions. We'll
see that some of the techniques that are useful for writing
mutagenic modules can also be used to extend the language.
There is also a family of modules which change the code syntax,
often very radically. Damian's Lingua::Romana::Perligata,
Evil Dave's Pony and so on. I'll call them
translator modules. They're not the same as mutagenic
modules.
>>