[m-rev.] diff: IL back-end: fix another bug with nested classes

[Tyson Dowd]

On 14-Aug-2001, Fergus Henderson <fjh at cs.mu.OZ.AU> wrote:
> On 13-Aug-2001, Tyson Dowd <trd at cs.mu.OZ.AU> wrote:
> > On 13-Aug-2001, Fergus Henderson <fjh at cs.mu.OZ.AU> wrote:
> > > compiler/mlds_to_il.m:
> > > 	When generating field references, fix up the class name for nested
> > > 	types so that the nested part gets put in the nested classes
> > > 	field of structured_name; this is needed to ensure that ilasm.m
> > > 	outputs the class qualifiers using the proper "/" qualifiers
> > > 	rather than using "." (which is for namespace qualifiers).
> > > 
> > > 	This also required changing get_ilds_type_class_name to handle
> > > 	value classes.
> > 
> > This is fine although it feels a little hacky to do it this way.
> 
> Agreed.  Very hacky.  I think we really need to modify the MLDS
> so that it keeps the distinction, rather than trying to recreate
> lost information at the ILDS level.  
 
An XXX to this effect near the code would be good.

> That was a bigger task than I was willing to handle in my haste to
> get some benchmark results for grade `il' in time to meet the BABEL'01
> paper deadline, which is why I went for the hacky approach above.
> I committed it on the main branch because even this hacky approach
> is better than the not-working code that was there before.
> But in the long term I think we should use a different approach.

Yes, I thought this was the case.

Tyson.
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------