[wg-camlp4] About Ptyp_poly nodes in the Parsetree

[Hendrik Tews]

Alain Frisch <alain.frisch at lexifi.com> writes:

   I'd like to get your opinion about the representation of "polymorphic
   types" (Ptyp_poly) nodes in the Parsetree.

Personally, I don't think that I or II is a problem. Well, II is
a bit inconvenient, and it is nice that it is fixed now.

Besides I and II there are probably more 
side conditions / invariants for the ast that are not enforced by
the type of the ast. I would view I and II as a specific instance
of how to deal with incorrect ast's. 

>From the extension developers point of view it would be nice if
the type checker would enforce all invariants. Alternatively, one
could have an ast-checker extension, that checks all invariants.
One could put the ast-checker extension into the extension pipe
whenever something is strange.


Bye,

Hendrik