Interesting solution indeed!

Thinking out loud, my first hope was that we could fix this in a very isolated part of the source code, instead of involving several functions and new kinds of data/tokens being created. E.g. only in the Writer.prototype.indentPartial() function would have been ideal, as that function sounds very much in charge for indentation of partials.. Might be easier said than done though.

Thanks a lot for your quick response. I'll dive in as well to see if we end up with the same solution or maybe a combination of both.

I agree. I tired to find the quickest solution that satisfies our requirements. It emphasises the complexity of having too much coupling between the parse function and the indentPartials function like you said. I will try to figure out a better way, now that we have all tests in place, it should be easy to verify the solution.

It emphasises the complexity of having too much coupling between the parse function and the indentPartials function

Well said!

I've also tried looking for different ways to fix it, but couldn't find anything else that introduced less coupling between the different functions & tokens passed.

Therefore leaning towards merging this as is and pushing a fix, as that would at least be valuable for end-users, and postponing improvements to reduce coupling for later as that's a more general concern.

Thanks a lot for finding a fix to this indentation bug! 👍

Published in v3.1.0.