[m-rev.] for review: fix dummy type bug in construct.construct

[Ian MacLarty]

On Mon, Nov 21, 2005 at 09:01:17PM +1100, Zoltan Somogyi wrote:
> On 19-Nov-2005, Zoltan Somogyi <zs at cs.mu.OZ.AU> wrote:
> > The fix is fine, and you can commit it. I have a more extensive fix
> > that also tries to prevent such problems from happening again; I'll
> > commit that after yours and after my bootcheck finishes.
> 
> Here it is.
> 
> Try to reduce the probability of any bug such as the one reported by
> Peter Ross happening again.
> 
> library/construct.m:
> 	Make sure that the relevant switches list all the values of the enum
> 	type the switch is on, since this way gcc will generate a warning
> 	message if the enum has a value added to it and the switch isn't
> 	updated.
> 
> 	Include the defaults only if MR_INCLUDE_SWITCH_DEFAULTS is defined;
> 	the intention is that this may be useful with non-gcc compilers.
> 
> 	Generate separate, more specific error messages for each type_ctor_rep.
> 
> library/deconstruct.m:
> 	Delete some old code that we haven't used in a while, and which we
> 	won't want to use in the future either. (The deleted predicate worked
> 	on the data representation scheme of the running compiler, whereas
> 	we wanted it to work on the data representation scheme of the generated
> 	program, which may be different.)
> 
> 	I found this when inspecting all uses of switch defaults in the
> 	library.
> 
> runtime/mercury_conf_params.h:
> 	Document MR_INCLUDE_SWITCH_DEFAULTS.
> 
> Zoltan.
> 

That looks good.

Ian.
--------------------------------------------------------------------------
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
--------------------------------------------------------------------------