Re: Modifiers and variables

On Thu, Aug 19, 2004 at 12:18:50PM +0100, Fergal Daly wrote:
> So what you really need is a parse_one_arg which tries to use the start of
> the string as an expression, consuming as much as possible but also gives
> you back what it couldn't use. Then you could easily do

That's an interesting idea. Right now the dev release of
Petal::Utils::Base contains a fetch_arg() subroutine which I've posted
to the list which contains the following rules:

  if string contains single quote, remove quotes and return
  if string == number and/or contains a decimal point, return it
  else lookup in hash and return

Your rule to return just the string if item is not found in the hash may
be a useful addition and wouldn't be hard to implement. It would be made
even easier if Petal::Hash::fetch returned the original arg if it did
not exist in the hash.

However, this may not sit well with the current implementation of fetch
and could lead to unexpected results in the case of a typo in the
template or by the programmer filling out the hash being passed to
process(). Of course, I don't think Petal complains now if an item is
not found. But some existing tests (both conditional tests and test
scripts) may fail if fetch starting returning values when none were
expecting.

So far, my needs as a modifier author have been very simple. The simple
syntax rules above are sufficient. I suspect that they probably miss
several edge cases that Jean-Michel and you can dream up, though.  I'll
have to leave those details to the brains of this operation <g>.


Cheers,
William

-- 
Knowmad Services Inc.
http://www.knowmad.com