Re^3: requiring pm modules

Perhaps your code sample is too simplified and out of context, but I see absolutely no reason for using symbolic references. Heck, even within context I doubt it. I've written dozens of Perl modules and programs and haven't really needed symbolic references, except within the very limited context of exporting symbols or creating methods automatically (which doesn't seem to be the case here).

Added 3 min later: if you don't know the name of the subroutine beforehand (and that's why you have it in $x), you can achieve a similar result without symbolic references by using object-oriented notation:

   require Value;
   $x = 'getValue';
   Value->$x();
[download]

Note that you'd have to change the getValue subroutine so that it shifts its first argument, which will hold the package name.

Comment on Re^3: requiring pm modules Download Code

Replies are listed 'Best First'.
Re^4: requiring pm modules by revdiablo (Prior) on Nov 03, 2004 at 23:08 UTC
Using the class method syntax will allow the symbolic reference even with strict, but it's still a symbolic reference. Instead, I would use `can`, as follows: `require Value; my $x = 'getValue'; if (my $coderef = Value->can($x)) { $coderef->(); }` [download] This has several benefits. It does not add "Value" to the argument list, it does not use a symbolic reference, and it checks that the subroutine actually exists before trying to run it. PS: for the curious, `can` is documented in UNIVERSAL.	[reply] [d/l] [select]
Re^4: requiring pm modules by Anonymous Monk on Nov 03, 2004 at 21:15 UTC
The name of the function 'getValue' comes from the configuration file too. I guess I didn't have no any other way to implement it without using the symbolic reference. Michael	[reply]
Re^4: requiring pm modules by Anonymous Monk on Nov 03, 2004 at 21:17 UTC
The name of the function 'getValue' comes from the configuration file too and I guess I didn't know any other way to implement this. thanks, Michael	[reply]