On Mon, Feb 19, 2001 at 08:17:17AM -0600, Jarkko Hietaniemi wrote:
> I would define a relatively strict and standard way to do this so that
> the documentation can be extracted out.
I'd like to see Perl 6 written as a literate program, but that's probably
the axe I need to grind. :) Seriously, the apidoc is a really good thing,
but it needs to be made more extensible; at the moment it only really works
for documenting functions. I think we'll need a more general "structural"
documentation on the source; RFC281 was an initial set of thoughts on this
idea.
> A related matter is that currently we have a central .pl file that is
> used to generate the appropriate prototype headers and various cpp
> magic incantations e.g. creating the threaded versions. I find this
> to be curious direction: I would expect the .c files to be the master
> files for function signatures, from which the prototype headers are
> generated.
If we're dealing with, say, vtable functions, I think we'll need to
generate a lot of information from the function signature itself; see
the dummy vtable generation code I posted the other day. (Incidentally,
I recently modified that so that you can put POD in the vtable source. :)
--
Morton's Law:
If rats are experimented upon, they will develop cancer.
