[Pmwiki-users] A (perhaps) small desire for track authorship:

[Jonathan Scott Duff]

On Thu, Aug 14, 2003 at 08:06:49AM -0600, Patrick R. Michaud wrote:
> On Thu, Aug 14, 2003 at 08:40:25AM -0500, Jonathan Scott Duff wrote:
> > On Thu, Aug 14, 2003 at 09:54:24AM +0200, Bernhard.Weichel at t-online.de wrote:
> > > I have another hint/propsal/desired for "Track authorship":
> > > 
> > > in wikipedia, the markup ~~~~ is replaced by a signature on *save time*.
> > > 
> > > This allows in particular on discussion pages, to sign contributions
> > > very easily.
> > 
> > That's kinda cool.
> 
> Yeah, but Scott may remember that he and I discussed something like this
> about a year ago and I felt it was somehow "unclean" to have PmWiki
> making modifications to the markup text as it saved it and wanted to
> avoid this.  

Yep, I remember clearly.  In fact, I don't like the proposal in
general, because the markup disappears when the document is saved.
But it's still kinda cool and something that I see no reason to bar
local administrators from doing.

It's too bad you can't apply the "all is fair if you predeclare"
principle from perl to pmwiki markup because writers can't necessarily
see the predeclarations of their administrators.

Random idea for the group: Wouldn't it be interesting if WikiAuthors
could decide which customizations pertain to a particular
WikiPage/WikiGroup by simply including some markup in the document?

> > We don't currently have any hooks into save-time actions, but you could
> > always create a local customization that overrides the HandlePost()
> > function to add such a hook, and then another local customization to
> > implement the above feature. 
> 
> Actually, HandlePost() is likely to provide hooks that allow customizations
> to specify an array of pre-save and post-save functions to be called when
> a post occurs.  This would give local customizations the ability to perform
> operations like this (and others) without having to chain HandlePost.

Yep, that'd be the way to do it.  But in lieu of that, I've given
Bernhard a solution he can implement now  :-)

> The diff routines would then make use of this hook instead of being
> hardcorded, and could eventually be replaced by a PHP-based diff if/when 
> one is ever created.

Yeah, and you can handle the RecentChanges updates that way too.

-Scott
-- 
Jonathan Scott Duff
duff at cbi.tamucc.edu