Re^2: Entry Widget - validatecommand

sub print { print $value unless $value eq '-'; }
[download]
...

Incidentally, why does this print sub [not] simply call itself recursively??

Its a namespace problem (feature? artifact?). Your new print is defined in main and the built-in print is defined in CORE and the CORE namespace has search precedence. Try the following:

>perl -wMstrict -le
"sub print { CORE::print('foo') }
 ::print;
"
foo

>perl -wMstrict -le
"sub print { ::print('foo') }
 ::print;
"
Deep recursion on subroutine "main::print" at -e line 1.
Terminating on signal SIGINT(2)
[download]

Comment on Re^2: Entry Widget - validatecommand Select or Download Code

Replies are listed 'Best First'.
Re^3: Entry Widget - validatecommand by Athanasius (Archbishop) on Jul 24, 2012 at 08:47 UTC
Thanks, AnomalousMonk, for the explanation. Your statement: the `CORE` namespace has search precedence. raises a new question: How, then, does the original code ever find the `print` sub defined in namespace `main` if the sub isn’t explicitly qualified with its namespace (i.e., `main::print`)? Why doesn’t Perl always find `CORE::print`? I think I found the answer: the original code works only because it uses a reference to the sub: `-command=> \&print` [download] and in this case the reference is disambiguated using lexical scope. I found an explanation of the latter in the Camel Book, 4th Edition, “Name Lookups,” pages 62–65. Is there any documentation on how Perl looks up subroutine names when a subroutine is actually called (as opposed to being referenced)? Thanks, Athanasius <°(((>< contra mundum	[reply] [d/l]

Replies are listed 'Best First'.

Re^3: Entry Widget - validatecommand
by Athanasius (Archbishop) on Jul 24, 2012 at 08:47 UTC

Thanks, AnomalousMonk, for the explanation. Your statement:

the CORE namespace has search precedence.

raises a new question: How, then, does the original code ever find the print sub defined in namespace main if the sub isn’t explicitly qualified with its namespace (i.e., main::print)? Why doesn’t Perl always find CORE::print?

I think I found the answer: the original code works only because it uses a reference to the sub:

-command=> \&print
[download]

and in this case the reference is disambiguated using lexical scope. I found an explanation of the latter in the Camel Book, 4th Edition, “Name Lookups,” pages 62–65. Is there any documentation on how Perl looks up subroutine names when a subroutine is actually called (as opposed to being referenced)?

Thanks,

Athanasius <°(((>< contra mundum

[reply]
[d/l]