[m-rev.] for post-commit review: fix mantis bug #367

[Julien Fischer]

Hi Zoltan,

On Wed, 3 Dec 2014, Zoltan Somogyi wrote:

> I think the fix to array.m is simple enough, but I would like
> feedback on the updated documentation.


> diff --git a/library/array.m b/library/array.m
> index f9cba24..be7fb8b 100644
> --- a/library/array.m
> +++ b/library/array.m

...

> +    % approx_binary_search(A, X, I) does a binary search for the element X
> +    % in the array A. If there is an element with that value in the array,
> +    % it returns its index I. If there is no element with that value in the
> +    % array, it returns an index whose slot contains the highest value in the
> +    % array that is less than X, as messured by the builtin Mercury order

s/messured/measured/

> +    % on terms for approx_binary_search/3, and as measured by the supplied
> +    % ordering for approx_binary_search/4. It will fail only if there is
> +    % no value smaller than X in the array.

...

> The array.bsearch predicate has been sort-of pre-deprecated
> since 2009, via a note that tells people to call array.binary_search
> instead. Any objections to adding an actual pragma obsolete on it?

No, go ahead and add the pragma.

Cheers,
Julien.