[m-dev.] argument packing

[Julien Fischer]

On Tue, 15 May 2018, Zoltan Somogyi wrote:

> So if you have compiled the Mercury programs you are working on
> with an installed compiler which has had its default values of
> allow_packing_ints and allow_packing_dummies set to `yes',
> and had those programs pass all their tests, please tell me.
> If you haven't, then please do so and tell me.
> And if you had any failures, please tell me as well.

I have tested the new options with the following:

    * my JSON library [passes in asm_fast.gc]
    * my CSV library [passes in asm_fast.gc]
    * the Zinc compiler [passes in hlc.gc]
    * Opturion's runtime system [passes in hlc.gc.trseg]
    * Two of our transport optimisers

At least one of the latter is now giving me different results that
differ from a system built with rotd-2018-05-10.  They're not incorrect
as such, just marginally different.  (There may be factors other than
the Mercury compiler in play, I'll have to investigate further.)

Note that none of the above use sub-word sized ints (yet), the
only packing I observed was for enumeration arguments.

Julien.