Deprecating CoffeeScript and JavaScript Meta Parsers

Does anyone use the --- coffee or --- javascript meta data headers? These allow a document’s meta data to be parsed via coffeescript or javascript respectively.

These are undocumented it seems - as the meta-data docs doesn’t cover it - however it is supported, and tests for such a thing have been long-standing.

The latest major version of extendr only works with plain objects, and parsed CoffeeScript and JavaScript code is not a plain object due to the sandboxing (the object prototype of the sandbox isn’t the same as the parent) - Update Dependencies · Issue #1045 · docpad/docpad · GitHub - so as such, this feature is causing a delay in the upcoming DocPad release.

I’m considering removing this functionality, considering this complication, its undocumented status, and apparent lack of use in the real world (I couldn’t find a real-world project where I actually used it), as it was pretty advanced stuff, and mostly useful for dynamic documents, which is a use case we are moving away from.

While I can code a workaround for this use case, I’d rather remove this functionality altogether, in the process of streamlining, simplifying, and improving the performance of DocPad. The less complex edge cases, that can be accomplished by more streamlined ways, the better.