The begin doesn't work without the our before it...

our and my have compile time semantics, like pragmas (e.g. strict). They tell the compiler how to set up the variables declared. That impact might be the same as an implicit BEGIN.

And if it hits a use of a module it is already in the middle of processing, all it will do is call that modules import sub, and not try to compile it's code a second time (thus preventing loops).

Sounds right; but it will call the import sub only if it is defined at that point. At compilation, pragmas are evaluated, globs are set up, scopes are built, my variable slots are allocated in that scope's PAD - i.e. all structures needed for execution are created. If at that time an import() subroutine has yet been allocated in a typeglob slot, it will be called. I'm not sure about whether that import() call triggers a compile at that point, or if it has been compiled previously - but I guess the former. Anyways, some code that uses some other code which is just being compiled sees whatever has been defined in that other code's name space at that moment.

_($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                              /\_¯/(q    /
----------------------------  \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}