This is the mail archive of the
mailing list for the GNU Fortran project.
Re: [patch, fortran] PR 29393: Ranks of PARAMETER-lhs in initializations
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Erik Edelmann <erik dot edelmann at iki dot fi>
- Cc: gfortran <fortran at gcc dot gnu dot org>, gcc-patches <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 23 Oct 2006 19:05:04 -0700
- Subject: Re: [patch, fortran] PR 29393: Ranks of PARAMETER-lhs in initializations
- References: <20061023193501.GA6047@acclab.helsinki.fi>
On Mon, Oct 23, 2006 at 10:35:01PM +0300, Erik Edelmann wrote:
> This patch fixes PR 29393. Problem: In expr.c
> (simplify_parameter_variable), we replace the original expression p with
> a simplified experssion derived from p->symtree->n.sym->value. Now, if
> p->symtree->n.sym->value has different rank than p, the rank will become
> wrong. This oneliner will restore the rank to the correct value.
> Tested on trunk, Linux/x86. I intend to commit it tomorrow as obvious
> to trunk and, if it passes regressiontesting, 4.2 and 4.1, unless there
> are objections.
The patch is OK (as you say it is obvious).
As to backporting to 4.2, my own effort will be limited
by the effort required. If the effort is trivially
applying the patch, bootstrapping, and regression testing,
then I'm likely going to backport. If I need to add other
patches and rewrite some code, I won't bother. I will
not patch 4.1.
If someone supplies 4.1 and 4.2 backports, I won't object.
But, I think we should still concentrate our efforts on