[pmwiki-users] random musings

[Joachim Durchholz]

Patrick R. Michaud schrieb:
> One option is to just allow wildcards directly on "group" and "name"
> values in conditionals, so that
> 
>     (:if group PmWiki* :)        # any group begining with "PmWiki"
>     (:pagelist group=PmWiki* :)  # all pages in groups starting with "PmWiki"
>     (:if name A*:)               # any page starting with "A"
> 
> But how to match both group and name simultaneously?  For consistency 
> with other markups, do we just use "fullname"?  (Somehow I think it's
> a big ugly.)
> 
>     (:if fullname PmWiki.A* :)   # pages starting with "A" in the PmWiki group
>     (:pagelist fullname=2005.Blog* :)  # "Blog*" pages in the 2005 group
> 
> Hmm, I guess "fullname" isn't so bad.

How about making the left "fullname" side of the comparison a series of 
PmWiki variables and delimiters?
This would open up the possibility to compare more than just the group 
name; one could do things like
   (:if group.user PmWiki.admin :)
I'm not sure what characters should

(Now what's really missing is a way to OR and negate conditions... but I 
haven't followed conditional markup lately, it may be already there.)

 > Lists of wildcards will also be allowed
> 
>     (:pagelist fullname=2005.Blog*,2006.Blog*:)
> 
> as well as negations
> 
>     (:pagelist fullname=-*.HomePage,-*.RecentChanges:)  

... oops, already catered for :-)
Quite expressive, though not full boolean expressions. Hopefully that's 
enough anyway.

> 3. Pagelist refactoring
> Pagelist has evolved *very* nicely from where I started over a year
> ago, but the code is becoming pretty crufty in the interest of 
> preserving backwards compatibility.  It may be time to do a major 
> refactoring for pagelist (we are in beta, after all), and provide 
> the current pagelist code as an optional module or recipe for 
> sites with custom pagelist settings and need time to migrate
> those into the new framework.  However, custom FPL formatting 
> functions would continue to work in the new version w/o modification 
> -- the refactoring would affect only those people who have 
> customizations that rely heavily on the existence of current 
> variables and their values.

No serious problems with that.
As long as there's a viable migration path :-)

Regards,
Jo