This is added to the site via Epistle on my tablet. it goes through Dropbox, which Epistle syncs, and then I visit a webpage which gets an auth token (asking me to sign in to DB if it doesn’t have one), and pulls down files from the folder Epistle syncs to. these are then checked against the blog platform to see if a file is an update of and existing post, and then the post is added to version control amd published. The site runs on Jekyll / Heroku, setup was cribbed from a blog post called Blogging For Hackers.
Things it needs before Gihubbing :
- Some validation. At the moment if you’ve got a Dropbox UID in your cookies then the web-app will run. Admittedly the UID is salted and hashed, but still it makes me uncomfortable.
- The script needs to be refactored. Currently it is in the ‘make it work’ state.
- Configurables need to be abstracted. App keys, Epistle folders, Jekyll directory, token storage mechanism…
- Ideallyit would use Dropbox deltas to track deleted files.
Also I could use it as a thing to learn another language. Web requests with sessions / cookies, file IO, database connection… then either executing system commands or finding a git library, a rake library and a heroku library. Oh my.
All in all it’s been a lovely exercise in weiting something that stops me from writing blog posts.