[m-dev.] another interface file?

Peter Wang novalazy at gmail.com
Wed Jun 10 16:30:39 AEST 2009


2009/6/4 Zoltan Somogyi <zs at csse.unimelb.edu.au>:
> On 04-Jun-2009, Julien Fischer <juliensf at csse.unimelb.edu.au> wrote:
>> One alternative would be to always hoist abstract equivalance type
>> definitions (plus any supporting definitions) into the implementation
>> sections of the existing interface files.  Obviously, this will affect
>> separate
>> compilation to some extent.  On the upside, it would also make detecting
>> overlapping type class instances easier.
>
> Any such forced export should be ignored by the compiler until *after*
> the front end. We do not yet have a mechanism for this.

I believe Julien's idea was already implemented by Simon all this time:

http://www.cs.mu.oz.au/research/mercury/mailing-lists/mercury-reviews/mercury-reviews.0311/0154.html

i.e. stage 115.

However, when intermodule optimisation is switched off, some abstract
equivalence types still aren't expanded properly.  Probably just bugs.

Peter
--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at csse.unimelb.edu.au
Administrative Queries: owner-mercury-developers at csse.unimelb.edu.au
Subscriptions:          mercury-developers-request at csse.unimelb.edu.au
--------------------------------------------------------------------------



More information about the developers mailing list