Re: Re[2]: Re[2]: mixing of asign() and assign_by_ref() overwrites external vars

--- Dmitry Koteroff <[email protected]> wrote:
> B> Another example of breaking BC without adding anything new...
> It  is  not "breaking". It is adding new feature without changing
> existing behaviour (apropos, undocumented behaviour).

It is not a new feature, assigning by ref facilities are already
provided by Smarty.

> (What is BC? "Before Christ"? Or, maybe, "Bachelor of Commerce"?)

Hmmm. have you heard of the term Backwards Compatible? FWIW, "Bachelor
of Commerce" is usually denoted B.Com.

> Seems  I  know why are you so afraid to modify existing core. The
> reason  is  good and noble - you are afraid of new possible bugs.
> But everything has its measure.

You are right about that -- conversely, I'm getting the idea that you
are too ready to modify the core without considering how others might
be using it :) 

This breaks BC because it changes the way assign works. In particular,
what if someone is currently doing something like this:

  $a = "b";
  $b =& $a;
  $smarty->assign('b', $b);

So now in template {$b} == "b". However, if from a plugin:

  global $a;
  $a = "c";

With your patch, the in-template result is that now {$b} == "c" and
this is *not* the currently expected result in Smarty.

You can say what you want about the value of that usage pattern -- the
point is, you can't predict if people are using it or not (and there
are other ways to invoke that scenario).

Sorry if I sounded too abrupt before -- I realize you are trying to add
value and I should have taken the time to provide a more thorough
comment to your proposal instead of dashing out a one-liner as I did.

Best Regards,
xo boots


	
		
__________________________________
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/ 

-- 
Smarty Development Mailing List (http://smarty.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php