[m-rev.] for review: fix mantis bug 351

[Peter Wang]

On Sat, 09 Aug 2014 13:55:26 +0200 (CEST), "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> I think most developers should have a look at the comment I added
> at the top of proc_info_init in hlds_pred.m, because it may you to find
> that it exposes a problem with some aspect of the compiler that
> you work with, whether it has been exposed as a bug or not.
> 
> For review by anyone.
> 

> Fix Mantis bug 351.

I suggest a descriptive title, e.g.

Fix has_parallel_conj field on generated procedures (Mantis bug #351).


> @@ -2661,6 +2674,9 @@ proc_info_init(MainContext, Arity, Types, DeclaredModes, Modes, MaybeArgLives,
>      MaybeHeadModesConstraint = no `with_type` maybe(mode_constraint),
>      % argument MaybeArgLives
>      % argument MaybeDeclaredDetism
> +        % Inferred determinism gets initialized to `erroneous'.
> +        % This is what `det_analysis.m' wants. det_analysis.m
> +        % will later provide the correct inferred determinism for it.
>      InferredDetism = detism_erroneous,
>      goal_info_init(GoalInfo),
>      ClauseBody = hlds_goal(conj(plain_conj, []), GoalInfo),

Indentation?

It looks good otherwise.  Thanks for fixing it so quickly.

Peter