Pandoc's Markdown (was: automate "Elements")

[boisgera]

Think of a way to automate the generation of the Elements section :

• fetch the markdown source automatically,
• use content tabs to display the markdown on one hand and the python Pandoc element on the other,
• make a custom test runner behavior for this stuff the kind of elements to find could even be guessed from the display once it has been generated ; it does not however says anything about what to pick during the generation. Mmm we could just put a placeholder in the results with the types names to pick, that could work.

Try this under "Pandoc's Markdown" section.

[boisgera]

Mostly done in c227496 (including tests).

What remains the need to (manually ?) tweak the doc in some place so that the tests are relevant. For example, sometimes only fragments are displayed that do not make sense unless some additional context is present in the document (e.g. link which is never used).