-
-
Notifications
You must be signed in to change notification settings - Fork 88
Hosting
Currently the training docs are hosted on T.P.O and on RTD.
With a commit to master, travis will run some basic checks (Spell-check, Link-check, Test-build of Sphinx). If these tests succeed, travis will upload then the HTML to T.P.O.
Because of the way how travis works, this is of course not working with forks, it only works with branches. This is also not a bug, this is OK, otherwise there would serious something wrong with travis !
I am writing this because people were 'complaining' about this and that they are getting a failing travis from within their fork....
RTD is connected with a webhook to the docs repository, that means if you commit/merge this webhook is called and pushing to RTD.
Still a interesting challenge !
The new repository structure will not work with RTD out of the box, the questions here is, what are the reasons to have the docs on T.P.O and RTD ?
There are different ways/solutions IMHO we should not forget the 'user story', too!
What does it mean ?
Besides finding a good solution for deploying and hosting we should also take in that people need/want to have the docs locally running.
This is less of a 'problem' for 'more advanced' people who are following a 'developer training' but it can hard for people who normally do not use pip/buildout/python/.....
Currently we tell people 'setup Sphinx, clone the training repo and run make html'. This is hard and can be a 'stopper' for less experienced people.
If we tell people to install something it should be well documented and kind of easy, like installing an 'app on macOS or Windows'.
Have every training as container, this works well on Linux, Windows and macOS. Still what to do with *BSD ?
Have a vagrant box for every training documentation, Overkill on resources and work intensive, but works on all platforms