DidiWiki 0.5

DidiWiki is a small and simple WikiWikiWeb implementation written in C. Its intended for personal use for notes, Todo's etc. It includes its own webserver and hopefully its as simple as just compiling, running and pointing your browser at it.

Installation

To build do the usage ./configure and make. You can then make install or run DidiWiki right from the src dir. On startup it'll tell you the address to point your browser at. You should then be ready to go.

If upgrading from an earlier version its a good idea to delete ~/.didiwiki/DidiHelp before running so you get the latest version of the help file.

Remote API

As from 0.5 DidiWiki has a very simple 'REST' like http api. Hopefully this will make it possible to write alternate front ends to didiwiki.

The api looks like this with parameter passable via GET or POST;

http://didiwiki/api/page/get?page=XXX

Get the raw wiki text for specified page. Returns http 500 code if page does not exist.

http://didiwiki/api/page/set?page=XXX&text=XXX

Set the wikitext of a page. Page is created if it doesn't exist.

http://didiwiki/api/page/exists?page=XXX

Returns 'success' if page exists, 500 error if not.

http://didiwiki/api/page/delete?page=XXX

Deletes a page.

http://didiwiki/api/pages

Returns a seperated list ( one entry per line containing page title,TAB,modified date) of wiki pages with the most recently modified first.

http://didiwiki/api/search?expr=XXX

Like the above but lists only pages which contain 'expr'

Notes & Excuses

• If you want to contribute, please check the TODO !

• Pages are stored in ~/.didiwiki ( you can override this by setting 'DIDIWIKIHOME' env var to a valid path. ). You can add a styles.css to this directory to override the 'inbuilt' stylesheet.

• The webserver part is heavily based on cvstrac's internal server, following the same kind of process;

  • Its a pretty simple lightweight forking server. A child process is forked to handle each http request.

  • These forked children are assumed to not be around for very long ( less than a second ), thus no real attempt is made to free up memory in these child processes.

  • Its probably not very secure at all. !!PLEASE DONT RUN AS ROOT!!

• The code uses asprintf in a couple of places. I dont know how portable this is to non Linux systems. If its a problem to you Feel free to send patches.

• You can debug segv's by running like 'didiwiki debug' which will not start the http server, it existing in a single process but expecting the http request on stdin.

• The formatting style is very similar to that of kwiki's - I just like kwiki :-)

• I wrote it in C as I've mostly forgotten Python, Perl etc ( and I wanted it to fit easily on a PDA ).