in reply to Re: RFC: Idiom for named read-write arguments (aliases) instead of using $_[0] etc
in thread RFC: Idiom for named read-write arguments (aliases) instead of using $_[0] etc
Can't say much about the literals and the missing error, I think swapping literals doesn't make much sense ... (well maybe in LISP).
Anyway much of your reasoning is about manipulating arguments is bad practice.
Actually perl has plenty of bultins which do that and many lamda-patterns use this feature (i.e. map/grep {BLOCK} like commands where at least $_ is aliased) and if you look closely into perl6-signature you will find the attribute "rw" for exactly this purpose.
So I can understand if manipulating aliases is not OK for _your_ coding practice, but perl is a multi-paradigm language, where other want to implement their own patters.
For me using two $-sigils like in "$$a" is indication enough that this is no normal assignment, and certainly better to read and more obvious than spreading $_[n] around a 50 line long sub.
update:
I can show you old CORE modules which are programmed w/o unpacking variables and are very hard to maintain. Automatically refactoring $_[n] into $$name would instantly work, even on LHS of assignments and improve readability.
Cheers Rolf
|
|---|