> sub submenu_filter_factory {
> my ($context) = @_;
> return sub {
> my $text = shift;
> my $count = 1;
This should probably be my $count = 0;
> $text =~ s[<h2>(.+?)</h2>][replace_tags( $context, $1, $count++ )]gie
> ;
You might want to add "s" to "gie" to handle multi-line headings (but
your hrefs and tags will need to be fixed too). You might want to
ignore optional attributes too...
>
> return $text;
> }
> }
>
>
> sub replace_tags {
> my ( $context, $tag, $count ) = @_;
> my $stash = $context->stash;
>
> my $page = $stash->get( ['template', 0, 'page', 0 ] );
> my $jump = '#' . $page . $count;
>
> # I'm lost here
> $stash->set( ['sections', $count ], { href => $jump, text => $tag } );
> $stash->set( 'IMPORT', $stash->get('sections') );
>
> return qq[<a name="$jump">\n<h2>$tag</h2>];
> }
The first $stash->set() looks fine. I don't know what the second
$stash->set() is meant to do; it is likely not needed.
> So in my wrapper I PROCESS menu: (and I'm sure this can be written better!)
The remaining problem likely has to do with localization of the stash.
INCLUDE (and WRAPPER) localize the stash, meaning a local copy of the
top level is made, and then restored when the INCLUDEd template is
finished.
I'm not 100% sure of your hierarchy, but if your filter is below
(inside) an INCLUDE or WRAPPER, then your sections variable won't
exist after the INCLUDE or WRAPPER.
Here's a quick test: add something this at the top:
[% sections = []; %]
This pre-defines an empty array, whose entries (ie: sections.0, ...) will
now survive delocalization.
Also, I find the Dumper plugin is a great way to debug these kinds
of things since you can print out variables at various points to
make sure everything is ok.
Craig
