[m-rev.] diff:

[Zoltan Somogyi]

On 16-May-2003, Fergus Henderson <fjh at cs.mu.OZ.AU> wrote:
> To elaborate on my earlier comment: I think the right solution for this
> is probably to leave mlds.m unchanged, and to instead just modify mlds_to_c.m
> so that it outputs `MR_ArrayPtr' rather than `MR_Word' for arrays.

I agree that it is the right *long term* solution, but it won't work with
the old version of array.m; you'd get pages of error messages from the C
compiler. Without the change to mlds.m, compilation in high level C grades
works either only before the change to array.m or only after, depending on
whether mlds_to_c outputs MR_Word or MR_ArrayPtr. For bootstrapping on roy,
ceres and hg, the machines on which hlc.gc is the default grade, you need
hlc compilation to work *both* before and after. The change to mlds.m
makes the type output by mlds_to_c dependent on the contents of array.m,
making it work both before and after.

I will back out my change to mlds.m and replace it with the solution you
requested (which I am bootchecking on aral right now in both hlc.gc and hl.gc)
*after* the change to mlds.m has been installed on all three of roy, ceres and
hg, enabling bootstrapping of the change to array.m. If the diff to mlds.m
causes problems for you when you work on Java, you can simply back it out
right now in your workspace; it won't cause a CVS conflict.

Zoltan.
--------------------------------------------------------------------------
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
--------------------------------------------------------------------------