[wg-camlp4] A new branch to experiment with extension points

[Alain Frisch]

On 03/01/2013 05:54 AM, Yaron Minsky wrote:
> Is there any hope of keeping the "with" syntax for type definitions
> that type-conv uses?  There's quite a bit of code that depends on it,
> and there seems at least some utility in keeping it.

There will be a need for attributes on individual type declarations (not 
only structure items, since a set of mutually recursive declarations in 
a single structure item), and "with" seems quite good.   However, I 
don't believe that it will be possible to find a syntax which will allow 
reusing existing code based on type-conv without any change.  Leo 
suggested that existing Camlp4 extensions could be turned into mapper 
from "old syntax" (the one they used to interpret) to the new one (based 
on attributes/extension nodes, to be interpreted by a -ppx mapper). 
This could help in the migration (at least as a way to document what's 
to be done).

-- Alain