ikiwiki/doc/todo/wikitrails.mdwn

41 lines
2.4 KiB
Plaintext
Raw Normal View History

## summary
at times it is useful to have a guided tour or trail through a subset of the pages of a wiki; in pmwiki, this is implemented as [wikitrails](http://www.pmwiki.org/wiki/PmWiki/WikiTrails).
i'm working on a python xmlrpc plugin for ikiwiki to support wikitrails, both as a navigation feature (have "forward" and "back" links based on a sequence) and a modified inline that includes all pages in the trail with appropriate modifications (suitable for printing if necessary).
the current status is published on `git://github.com/github076986099/ikiwiki-plugins.git`; as of now, i don't have a public demo of it.
feedback on both the concept and the code is very much appreciated by [[discussion]] or [email](mailto:chrysn@fsfe.org).
## usage
two preprocessor commands are provided:
### \[[!trail index="my_indexpage"]]
embeds a navigation object with forward and back links as well as an indicator of the current position in the trail.
if index is not specified, a suitable page up the path is used.
this works very well together with the [[sidebar|plugins/sidebar]] plugin if the pages in a directory are roughly the same as the pages in the trail and the `index` is directory index page; just put the \[[!trail]] in the sidebar.
### \[[!trailinclude index="my_indexpage"]]
all pages linked from the index page are included in the same way as \[[!inline]] does, but in the proper sequence, with headings according to the indent in the source page and adoptions for the headings inside the page (a level 2 heading in a page that is a sub-sub-chapter in the whole trail will become a level 5 heading when trailincluded).
## the index page
the index page is parsed as markdown; numbered lists and "`*`" bulleted lists are discovered.
## current issues
* rebuilding --- currently, there is no propper rebuilding of pages (will use `will_render` and `add_depends`). care has to be taken of how not yet created pages play into this.
* inline recursion --- there is simply no guard yet
* navigation layout --- has to be both flexible and usable-by-default
* heading shifting
* currently only works for markdown
* can break the limit of html's six heading levels
* search for index page is currently next to hardcoded
* reading the index --- markdown syntax parsing is currently on a it-can-use-what-i-produce level; maybe integrate with existing mdwn parser
* uses undocumented titlepage command