In message <450488.31672...@web33002.mail.mud.yahoo.com>,
Risto Vaarandi writes:
>hi John,

Hi Risto:

>after giving it some thought, I realized why I the parser currently
>allows for multiword context names -- although context actions only
>accept a single word name (largely for parsing purposes), the name
>could become a multiword name after variable substitution (as you
>have mentioned). In order to allow for matching multiword names with
>context expressions, the parser currently supports them (for example,
>someone might want to create a constant for matching a particular
>multiword name).

Right, but the constant would be a single word from the parser's
perspective. It expands to multiple words, but the specification is a
single word.

>In order to recall that, I had to look at the code
>for some time :) I have to admit, though, that I have never used
>multiword names myself.

I claim it's a historical artifact. How about this:

  have the parser recognize a multi-whitespace-delimited word
  and emit a "WARNING: Multi-word contexts are depricated and
  will be removed in a future release".

This should allow SEC upgrades to occur seamlessly and will allow
people using multiword context specification to convert to single word
names by adding _'s or whatever. Like you, I have never used
multi-word context names because a context named that way can't be
used with any of the context actions.

Once the detection/warning is in place, the decision on what to do next:

  remove all support for multiword contexts in context expressions
    (since multiword contexts are invalid in all other places)

  create a newer parser and keep the current parser, selecting the
     context parser from the command line with the newer more strict
     parser as the default

  ...

can be deferred until we see how many people complain about the
deprication warning 8-). Just having the warning would have saved me
hours of effort.

For the purpose of my class, a context specification will always be a
single word 8-).

--
                                -- rouilj
John Rouillard
===========================================================================
My employers don't acknowledge my existence much less my opinions.



------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Simple-evcorr-users mailing list
Simple-evcorr-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/simple-evcorr-users

Reply via email to