[pmwiki-users] Re: Upgrade question - tracking down wiki words

[chr at home.se]

On Sun, 12 Jun 2005, Patrick R. Michaud wrote:

> > In the process of upgrading to pmwiki-2, I'm planning to stop using camel 
> > case for wiki words. Is there a way to find all such links, so that I 
> > (manually?) can fix the links?
> 
> In general this can be a bit tricky, because we have to find all "bare"
> wikiwords and ignore any that might appear in the middle of directives,
> urls, [[...]] links, etc.

Ok

> One temporary solution might be to continue to support WikiWords, but have
> them generate a link to a page only if the page exists

That sounds like a wortwhile function in itself. From a practical point of
view, I suspect the strategy would have to be combined with disabling
certain words (e.g. PmWiki) from being treated as wiki word links.

> There could even be an option (e.g., in a URL or cookie) that causes
> WikiWord links to display differently, so that someone browsing the site
> could easily identify where the wikiword links are and manually convert
> them to non-WikiWord links.  I may do this for the documentation section
> of pmwiki.org...

How about calling it a 'debug-skin'? It might be convenient to have
something like this shipped with default pmwiki, and be able to use it for
things like that.

> Short of that, the possibilities I can imagine...
>    1. Build a pagelist-filter to find pages containing bare wikiwords.
>    2. Create a script that goes through a list of pages, processes them
>       for wikiwords (ignoring wikiwords that appear in the middle of other
>       link constructs) and modifies them accordingly.

N:o 1 seems useful, but I think 2 might be overkill. I guess it depends on
the size of your site. Putting more than a little effort in this only
makes sense if there are many sites out there that'd like to convert from
wiki word links to [[wiki word]]?  (This might be the case though...)

/Christian

-- 
Christian Ridderström, +46-8-768 39 44               http://www.md.kth.se/~chr