Re: foo.inferred.mli and lazy typists

[Anil Madhavapeddy <anil@xxxxxxxxxx>]

On 25 Aug 2011, at 10:03, Richard Mortier wrote:

> trying to regenerate a skeleton mli for a largeish module that i don't really 
> want to have to type everything in for by hand.  aiui the best way to do this 
> is via the "ocamlbuild foo.inferred.mli" convenience target.
> 
> unfortunately this is a little less convenient than it might be - if the 
> module in question opens mirage related modules (Lwt, Bitstring) or uses 
> syntax extensions (certainly Bitstring, maybe more), the generation seems to 
> fail.

How does it fail? It certainly shouldn't.  If you open modules, they will 
appear in full in the inferred interface, but you just delete that from the MLI.

> 
> are there some other switches i should be passing on the command line, or is 
> there a better way of doing this?  at the moment i'm commenting out the bits 
> that cause failure just to force a partial skeleton to be built, but that 
> really does feel kludgy :)

No idea without seeing the actual failure.

-anil