Re^5: perl inheritance

No, I wouldn't put it in an import module. A BEGIN block would be better. If you put it in import, you have no guarantee that the import method will ever actually be called. It's quite easy to load a module without ever calling the import method. Conversely, it's very easy for the import method to get called multiple times.

I actually do something pretty similar to the technique being discussed in XML::LibXML::Augment.

perl -E'sub Monkey::do{say$_,for@_,do{($monkey=[caller(0)]->[3])=~s{::}{ }and$monkey}}"Monkey say"->Monkey::do'

Comment on Re^5: perl inheritance Select or Download Code

Replies are listed 'Best First'.
Re^6: perl inheritance by JavaFan (Canon) on Mar 23, 2012 at 23:26 UTC
What's the advantage of using a BEGIN block for that? Do you really think there's something to gain putting those methods in place, before the other methods are compiled?	[reply]
Re^7: perl inheritance by tobyink (Canon) on Mar 24, 2012 at 13:12 UTC
Very little advantage in this case. It just so happens that in the same loop I'm also playing around with `@ISA` and past experience has taught me to alter `@ISA` as early as possible. Modern versions of Perl are pretty smart when it comes to invalidating method resolution caches, but force of habit makes me put any `@ISA` alteration in a `BEGIN` block unless there's a good reason not to. `perl -E'sub Monkey::do{say$_,for@_,do{($monkey=[caller(0)]->[3])=~s{::}{ }and$monkey}}"Monkey say"->Monkey::do'`	[reply] [d/l] [select]
Re^8: perl inheritance by JavaFan (Canon) on Mar 24, 2012 at 13:46 UTC
Well, I can imagine putting assignments to @ISA in a BEGIN block if you have the habit of putting other things in a BEGIN block. I typically have my assignment to @ISA near the top of the file, before any method calls are done. And that never is a problem. Perl has always tried to invalidate method caches on assignment to @ISA -- although in a dim past, there was a bug preventing this to happen. This was fixed in 5.004 or 5.005. Long enough ago to not care about anymore. Not that method cache has any reason to be relevant here; for that to be relevant, we need to have something like: Assign to @ISA. Call a method "X". Assign to @ISA so X resolves to a different sub. Call method "X" again.	[reply]