Beefy Boxes and Bandwidth Generously Provided by pair Networks
Clear questions and runnable code
get the best and fastest answer

Re: Refactor method calls or not?

by Aristotle (Chancellor)
on Jan 20, 2002 at 14:41 UTC ( #140198=note: print w/replies, xml ) Need Help??

in reply to Refactor method calls or not?

Of all the replies, ++tilly's seems both closest to the original goal as well cleanest. If you really have a lot of such methods, I'd add a hash that maps add_financial_diary => 'financialDiary' so that a foreach in BEGIN() could generate these, but of course that's an obvious detail.

On the other hand I wouldn't put AUTOLOAD aside so soon despite all the advocacy against it. Since you mention that you typically only need these methods once or twice in a program's runtime, it seems an obvious idea to skip installing the methods in the symbol table altogether, in which case I find it can be written pretty cleanly. Imagine japhy's code here, but with the innermost if block replaced with something like (where _generic_add() directly does the job rather than generating a closure of course):

my $table = $calls{$meth} || croak "Unknown table action '$meth'"; $self->_generic_add($table, @_);

It is easy to substitute this for a loop in BEGIN() that installs routines which call _generic_add() if you later need more performance - and the only place that changes is a shift of code from AUTOLOAD to BEGIN, the rest of the module remains untouched, including the generic add routine and the "dispatch" hash.

Makeshifts last the longest.

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://140198]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others scrutinizing the Monastery: (3)
As of 2021-04-23 00:16 GMT
Find Nodes?
    Voting Booth?

    No recent polls found