[Pmwiki-users] Password for URL-approve?

[chr@home.se]

On Wed, 29 Dec 2004, Patrick R. Michaud wrote:

> > behaviour surprised me a bit to be honest, although I'm not sure it's 
> > necessarily bad.
> 
> In general I suspect that there are three possibilities:
>   1. all of the new links on a page are good, in which case approving 
>      the entire set at once is exactly what we want,
>   2. all of the new links on a page are bad (come from a spammer),
>      in which case we'll restore a previous version of a page and not
>      approve any links, or
>   3. some links are good and some are bad, in which case someone will
>      probably first edit the page to remove the bad links and put us
>      into case #1 above.  :-)
> 
> I couldn't come up with any situations where I would want to approve links
> one-at-a-time, although the code does support that.

I'm not sure, but what if you have a page that no one has approved the 
links for in quite a while, so it has lots of links?

(I'm actually not thinking only of spam here, but imagine a wiki for kids, 
where you don't want kids inserting links to port etc?)

Anwya, so assuming there are lots of links, I guess the important thing is 
being able to extract a list of them so that you can check them out one by 
one. See below..

> > In addition, I think that the page being shown when you do
> > 	action=approvesites
> > should list what sites are being approved, so you don't accidentally 
> > approve links.
> 
> Yeah, perhaps thats true -- maybe they can go in a footnote, 
> a popup list, or in a title of some sort.

How about a list of bullets, where each bullet is the link (in a working 
form), and clicking on it opens it up in a new window (and a snippet of 
text saying that the links are opened in a new window)? This would make it 
easy to go through lots of links at once.

Is 'action=approvesites' based on e.g. Main.ApproveSiteTemplate or 
something? That together with a directive to output the list of unapproved 
uris from a page should do it... of course, any page containing that 
directive should then have a 'noindex' thingy in the header - or we're 
back to square one.

caveat emptor or something

> > If I understand this correctly (and my brain is fuzzy because 
> > of a cold), we are actually approving domains rather than URIs?
> > Maybe the name should reflect this? (i.e. Main.ApprovedDomains)
> 
> Nope, the code actually checks for valid uri "prefixes" -- i.e.,
> "http://www.pmichaud.com" approves everything on www.pmichaud.com,
> while "http://www.example.com/somepath" approves only those uris 
> on example.com that begin with "/somepath".  In addition to
> "?action=approvesites" there's an "?action=approvelinks" that adds
> the individual uris to the list rather than just the domains.  I
> started out using ?action=approvelinks but quickly found that
> ?action=approvesites seems to be more appropriate.

Ah.. I see (action=approveuriprefix is more literal but kind of 
unreadable... :-)

> > I also wonder if this module is documented somewhere?
> 
> Not yet.  One of the many things that still needs documenting.  :-)

It'd probably be a good idea if you at least decided on an intial place 
for the documentation to. Then I (and others) could've added some text 
and/or comments there.  Writing this, I realize there's probably a PITS 
for this, but I'm thinking more of a page with a "proper name". If nothing 
else, that page could refer to the PITS page?

/Christian

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