[m-rev.] for post-commit review: don't create redundant type_infos

Zoltan Somogyi zs at csse.unimelb.edu.au
Wed Sep 16 12:23:35 AEST 2009


On 15-Sep-2009, Mark Brown <mark at csse.unimelb.edu.au> wrote:
> The number of variables in the procedure that was causing problems has
> roughly halved.  Compilation time has significantly improved but is still
> likely to be a problem for some models.  E.g., compiling it in asm_fast
> there is still most of a minute spent computing liveness for that procedure,
> where it was more like 10 minutes previously.

I will look into that next.

> The remaining variables are largely redundant typeclass_infos.  Note that
> most type_ctor_infos are indirectly reused via the type_info they are built
> for, so there are almost no redundant type_ctor_infos in this case.

I am just now posting a diff that fixes this.

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