aboutsummaryrefslogtreecommitdiff
path: root/doc/todo/wikitrails.mdwn
blob: 6168579667c07b5b5c52954d9f43e1c5fbe4342f (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
## 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).

### smcv's implementation

... is the out-of-tree [[plugins/contrib/trail]] plugin, see there for details.

### chrysn's implementation

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
   > Don't worry about that, titlepage isn't going anywhere, and will probably before a formal part of the api next time I consider api changes. --[[Joey]]