Ján

@DavidSnider is probably the best person to decide if this makes sense, right now hierarchies are introduced fairly late (5th page) which is probably intentional? I don't see any examples...

What's the rationale for this? Do we expect the documentation to not build on different systems (with the same HHVM version)?

Oh, the Ruby stuff. Makes sense.

Agreed, this definitely needs improvement. > How to get values from arrays ([], idx, maybe nested example as well) This is currently documented in the [subscript](https://docs.hhvm.com/hack/expressions-and-operators/subscript) section but would probably...

There's also some documentation on the `idx` reference page: https://docs.hhvm.com/hack/reference/function/HH.idx/ That should probably just link to wherever we decide to document all this.

> much clearer for structures such as arrays and maps I can't even find any description of the file format anywhere. The only thing I could find is https://en.wikipedia.org/wiki/Hierarchical_Data_Format which...

Seems reasonable, that page is way too long anyway so probably worth splitting.

"Containers" seems fine to me but also we could just call it "vec, dict, keyset" or even "Containers (vec, dict, keyset)" if there's enough space.

"Containers" page could also briefly mention the interfaces (Container, Traversable, Keyed...); I think we previously covered that as part of the documentation about collections.

> open source the merge driver(s) I think the relevant code is here: https://github.com/facebookexperimental/eden/blob/master/eden/scm/edenscm/hgext/mergedriver.py Unfortunately I can't find any good documentation, but there are some tests that can be used...