I've started upgrading APIDump to aid us in generating the wiki contents semi-automatically. The first part's there - it outputs the functions and constants in tables ready-to-copy into the wiki. One has to fill in the args, returns and notes, though. I'd prefer the tool do that for us, too - so we'd keep the documentation somewhere in the git source and generate all the wiki pages from it.
Hmm, an idea - do we need to make it a wiki? It might be easier (read: more manageable) for the plugin to just output a set of HTML pages, something like what Doxygen does with C++ sources. It'd ask MCServer for the API, we could feed it a file containing the descriptions of all (some of) the API, the tool will output a folder full of HTML files ready to browse or upload to a website. It could also output a file listing all the API that is present but not documented, and the API that is documented but not present. Then we could replace the API part of the wiki with a link to (a nightbuild of) the HTML API docs.
After all, wiki is meant for human editting, we want the API update process to be automated, so a wiki is not really the best platform.
What do people think about this?
Hmm, an idea - do we need to make it a wiki? It might be easier (read: more manageable) for the plugin to just output a set of HTML pages, something like what Doxygen does with C++ sources. It'd ask MCServer for the API, we could feed it a file containing the descriptions of all (some of) the API, the tool will output a folder full of HTML files ready to browse or upload to a website. It could also output a file listing all the API that is present but not documented, and the API that is documented but not present. Then we could replace the API part of the wiki with a link to (a nightbuild of) the HTML API docs.
After all, wiki is meant for human editting, we want the API update process to be automated, so a wiki is not really the best platform.
What do people think about this?