[m-rev.] diff: fix mantis bug 113

Julien Fischer juliensf at csse.unimelb.edu.au
Fri Oct 16 18:01:31 AEDT 2009

On Fri, 16 Oct 2009, Zoltan Somogyi wrote:

> Fix Mantis bug 113. Unfortunately, even without bug 113, the program
> has an error, so there is no simple way to add it as a test case.
> Getting rid of the error would take away the raison d'etre of the test.
> compiler/mercury_compile_front_end.m:
> 	Call the part of post_typecheck that fixes up the interface of
> 	predicates directly from here, not from purity.m. purity.m was
> 	calling this part first thing anyway, so there was no point in
> 	doing this from purity.m in the first place. Doing it from
> 	the top level allows us to stop doing purity checking (and thus
> 	post_typecheck) on the definitions of predicates if the program
> 	had a type error. Bug 113 was caused by an operation of post_typecheck
> 	on a goal that (previously incorrectly, now correctly) assumes
> 	the absence of type errors.
> 	Move the main (-v) progress messages here from purity.m as well,
> 	since here is where they are printed for other compiler passes.
> compiler/purity.m:
> 	Conform to the changes above, and fix some bitrot in old documentation.
> compiler/post_typecheck.m:
> 	Delete an unused predicate, and fix some bad programming style.

It is possible that this is also a fix for bug #110 - I believe Peter
Wang has uploaded a test case for that one to Mantis.

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

More information about the reviews mailing list