cases there should be another parameter to the directive that allows linking to switched on. Just switching
it on universally at this point might break a number of people's pagespecs. -- [[Will]]
+> There's a simple reason why these directives don't generate a record of a
+> wikilink between them and the pages they include: Semantically, inlining
+> a page is not the same as writing a link to it. Nor is generating a map that
+> lists a page the same as linking to it. I don't think this is a bug.
+> --[[Joey]]
+
+>> Fair enough. I guess we can mark this as [[done]] then.
+>>
+>> Just a bit of background on where I was going here... I was looking for
+>> a simpler way of attacking [[todo/tracking_bugs_with_dependencies]].
+>> In particular, rather than introducing changes to the pagespec definition,
+>> I wondered if you could use wiki pages as the defined pagespec and
+>> introduce a 'match_mutual' function which matches whenever two pages
+>> link to the same third page, then you don't need to alter the pagespec
+>> handling code.
+>>
+>> But that requires being able use use a pagespec to decide what pages
+>> are linked to. e.g. I want to make an 'openbugs' page that links to all
+>> open bugs. Then I could make a 'readybugs' page that links to
+>> `backlink(openbugs) and !mutualLink(openbugs)`. That is, all bugs
+>> that are open and do not themselves link to an open bug.
+>>
+>> The problem with all this is that it introduces an ordering dependency,
+>> as I noted below. I think the original proposal is better, because it
+>> handles that ordering dependency in the definition of the pagespecs.
+>> --[[Will]]
+
Here is a patch to make map link to its linked pages (when passed `link="yes"`). It is a bit problematic in that it uses a pagespec
to decide what to link to (which is why I wanted it). However, at the time the pagespec is used the links
for each page haven't finished being calculated (we're using the pagespec to figure out those links,