On Thu, Nov 11, 2021 at 3:39 PM Jean Abou Samra <j...@abou-samra.fr> wrote:
> Le 11/11/2021 à 15:12, Paolo Prete a écrit : > > I see, but my advice included your Scheme example, which is formatted > > as plain text as well. > > IMHO, a tool that formats on the fly, as a tree or list on the > > standard output in plain text, a sum of cross links, would be actually > > unusable in practical cases, even if it initially appears helpful. > > If you need an autodoc on the fly, you should consider how IDEs > > interface to the autodoc directory/files, giving hints through popups, > > beyond the scope of LilyPond (and huge effort to implement for any > > editor). > > Therefore I encouraged you in spending time in observing what is the > > "de facto standard" way of managing the API doc that is already part > > of the LilyPond project. > > > I am not managing to convey my meaning across. The > Scheme example was _not_ intended to become something > outputting documentation for on-the-fly usage in editors. > It was meant to be a proof-of-concept for retrieving > and structuring the information (rather than formatting > it) in a way that could be reused in the autogeneration > process that outputs the official Internals Reference. > Our autogeneration script is written in Scheme. > > I see. Is the autogeneration script made from scratch, or does it feed some preexisting tool, that inspects a set of files, for autodoc? I mean: in case of C++ code, Doxygen is (more or less) simply fed by a set of options and then, voila, the autodoc is generated. If not, I wonder as well if is it possible to use an alternative to the previous script and generate the autodoc by pointing to the src dir with some tool. best, P >