[m-dev.] Re: Problem with region frame slots

[Julien Fischer]

Hi Quan,

On Tue, 7 Aug 2007, Quan Phan wrote:

> I send this email to developer's list so that everyone can involve
> (especially Julien :) ).
>
> On Tue, Aug 07, 2007 at 05:44:29PM +1000, Zoltan Somogyi wrote:
>> On 06-Aug-2007, Quan Phan <quan.phan at cs.kuleuven.be> wrote:
>>>> However, I cannot reproduce the .c file and the HLDS dump you sent me.
>>>> Compiling test.m in grade asm_fast.gc.rbmm does not introduce any use of
>>>> regions for me. Do you know what could be causing that?
>>
>>> I compile as follows:
>>> $thismmc/lmc --use-regions --region-analysis --dump-hlds=240 test.m
>>>
>>> How did you compile it?
>>
>> In grade asm_fast.gc.rbmm, which sets --use-regions but not --region-analysis.
>> Is there any point in --use-regions without --region-analysis? I don't think
>> so,
> I also think no.
>
>> which means that --use-regions should imply --region-analysis, or that
>> --region-analysis should be REPLACED by --use-regions.
> One reason for the existence of the two may be to use --region-analysis
> without --use-regions to dump the hlds with region annotations.

That was the reason before the .rbmm grades were added.

> But as far as I understand, I think only --use-regions is enough. That
> means I am for the latter.
>
> What do you think?

We can probably get rid of `--region-analysis'.  It dates your (Quan's)
original addition of the region analysis.  I added --use-regions because
it correspond more closely to the macro name, MR_USE_REGIONS, used to
enable region support.

Julien.

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