Re: [ANN] lisp/ob-tangle-sync.el

[Mehmet Tekman]

Ihor Radchenko <yantar92@posteo.net> writes:

> No. :result-params is not an actual header argument. It is
> implementation detail - parsing :results header arg internally produces
> (:results . "all the values concatenated") and _also_ (:result-params
> "val 1" "val 2" ...).

Hmm... what is the benefit if they encode the same
information? (e.g. :tangle "one two \"three and four\"" vs
:tangle-params "one" "two "three and four").

Edit: Oh, I see -- Org parses the action from the `:tangle-params'
and Users can extend tangle as much as they want via `:tangle'

> > Why can't we just go back to that?
>
> We can. I did not expect that we would need to go this far into the
> rabbit hole. Although, I still think that unifying header argument
> parsing is something worth doing, even if I need to implement it myself.

Hah, me either. At the same time, it doesn't make sense to have
`:tangle-sync' present in one release of Org-mode and then deprecated in
the next (after your header rewrite). It would just confuse users.

I'm happy to have another go at a unified header approach, I will
just need to explore the code base a bit more.