Re: CORE::GLOBAL magic?

Here is some speculation:

$ perl -e '$x=\&fff;print $x->(2)'
Undefined subroutine &main::fff called at -e line 1.
[download]

Looks like when you reference \&<Something>, it pre-declares a coderef/subref named <Something>

However, since the Name <Something> is declared without any content, it will complain when called.

If you happen to provide the content of the name <Something>, it works:

$ perl -e '$x=\&fff; sub fff{44};print $x->(2)'
44
[download]

Memory fault -- brain fried

Comment on Re: CORE::GLOBAL magic? Select or Download Code

Replies are listed 'Best First'.
Re^2: CORE::GLOBAL magic? by Anonymous Monk on Jun 02, 2018 at 12:49 UTC
I believe that it is close to the notion of "auto-vivification." An object is created within the interpreter to reflect the fact that you have referred-to this thing, and you see the (dummy) object that it made, but it has no actual definition and therefore no content. CORE::GLOBAL like all of CORE is a pseudo-namespace that exists because a namespace is needed not because it is actually real in the sense of other namespaces which appear in source code.	[reply]
Re^3: CORE::GLOBAL magic? by LanX (Saint) on Jun 02, 2018 at 16:10 UTC
Yes auto-vivification is close, the analogy becomes clearer since there is a hash like syntax to access the CODE slot of a typeglob. (like `*CORE::sleep{CODE}` ) anyway the important question is to know if a real coderef exists. That's why you can check this with `exists &Func;` (sic) Cheers Rolf _{(addicted to the Perl Programming Language :) Wikisyntax for the Monastery}	[reply] [d/l] [select]