[m-rev.] fix review: fix bug with ctgc and lcmc in high-level C grades

Peter Wang novalazy at gmail.com
Mon Feb 4 14:56:09 AEDT 2008


On 2008-02-04, Julien Fischer <juliensf at csse.unimelb.edu.au> wrote:
>
> On Mon, 4 Feb 2008, Peter Wang wrote:
>
>> Estimated hours taken: 4
>> Branches: main
>>
>> Fix a bug with structure reuse and the last-call-modulo-cons optimisation on
>> MLDS backends.  When constructing a cell by reusing a dead cell, we didn't
>> take into account that the lcmc optimisation needs to assign the address of
>> some fields into variables.  Those variables remained unassigned and pointed
>> into random memory locations.
>>
>> Only tested on hlc grades with low-level data.
>
> LCMC is not currently compatible with --high-level-data or the non-C
> backends.

Deleted.

>> +    ->
>> +        Offset1 = Offset + 1,
>> +        ArgNum1 = ArgNum + 1,
>> +        ( TakeAddr = [ArgNum | TakeAddr1] /* ? */ ->
>
> Why is there a question mark there?

Just a reminder for myself. Deleted.

Peter

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



More information about the reviews mailing list