[WM] sitemap grep attribute

[Justin Mason]

alex canady said:

> i think this may be possible with the grep attribute
> of the sitemap function [...]

> what i can't figure out is where the $_ is in
> relationship to the overall processing of the sitemap
> rendering code - confusing i know.  i need to know
> what i can compare $_ (the current content name) to ??
> 
> if the node is for the current content ($_) then draw
> it.
> if the node is a parent for the current content then
> draw the node.
> if the node is a top level node then draw the node.
> but all i have to work with is $_ (the content name,
> not the node).  
> is this the right logic or am i just going down the
> wrong path?

Hmm, not sure...

The "current content item's name", as mentioned in the grep documentation,
is not the name of the current *page* -- it's the name of the content item
that's being looked at in the sitemap.

The "grep" code is evaluated once for every node in the sitemap, and $_ is
the name of that node; you can then decide to display/not display it.

Problem is, at the moment, the sitemap is only computed once... I hadn't
anticipated what you're planning to do :(

What you want is quite similar to the "breadcrumbs" concept, check out the
<breadcrumbs> tag for details on that.

Here's how to do it. First, generate a full sitemap, then *not use it*;
instead you just use the knowledge of the site layout from your own code.
When generating the page, include your own content item which does
something like this, in Perl:

  <{perlout
    # get the "main" content item on the page (A)
    my $cont = get_content_object("__MainContentName");

    # get the top-level "root" content item (B)
    my $root = $cont->{main}->getmapper()->get_root();

    my @family = ($cont);
    while (1) {
      $cont = $cont->get_up_content();
      if ($cont == $root) { last; }
      unshift (@family, $cont);
    }
    # now @family contains the list of this node, and the parent nodes,
    # like: ($parent1, $parent2, $thisnode)

    my @toplevels = $root->get_sorted_kids();
    # now @toplevels contains the list of all top-level nodes (ie.
    # nodes directly underneath the "root node".

    # now display them. Links and good-looking HTML are left
    # as an exercise for the reader ;)
    #
    my $topfamily = $family[0];
    foreach my $node (@toplevels) {
      print "+ <em>$node</em><br>\n";
      if ($node == $topfamily) {
	foreach my $famnode (@family) {
	  print "-   <em>$node</em><br>\n";
	}
      }
    }
  }>

There's a few undocumented things in there, I must make those visible
in the Perl code library to make it easier to do this stuff. Here's
what they do.

(A): "__MainContentName" is a magic content item; it's always set to the
name of the "main" content item on the page.  This is worked out as the
last non-generated content item displayed, which has "map=true" set.

(B): There's currently no decent API to get the root content item from
Perl code, so I've used the one I use internally. Sorry about this; both
of these will get proper APIs in the next version. :)

Another key point: when you refer to this content item you've just
defined, be sure to use a $[deferred_reference], otherwise it may be
expanded before __MainContentName is set.

In fact, this functionality sounds like a good feature for a new tag.
What do you think?


BTW I know the documentation is pretty skimpy.  If you run into particular
bits that are definitely poorly docced, send in suggestions/requests for
more doc and I'll know exactly which bits need work...

(Also: did the previous problem, with the .../*.txt not finding certain
files, work out OK, or is it still an issue?)

--j.