This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [Patch, fortran] PR32103 - Module with equivalence draws "unsatisfied reference"
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Toon Moene <toon at moene dot indiv dot nluug dot nl>
- Cc: Paul Richard Thomas <paul dot richard dot thomas at gmail dot com>, "fortran at gcc dot gnu dot org List" <fortran at gcc dot gnu dot org>, gcc-patches List <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 14 Jun 2007 11:00:47 -0700
- Subject: Re: [Patch, fortran] PR32103 - Module with equivalence draws "unsatisfied reference"
- References: <339c37f20705300611s49e35314kbe424f647b98e38f@mail.gmail.com> <467180A1.4040302@moene.indiv.nluug.nl>
On Thu, Jun 14, 2007 at 07:53:37PM +0200, Toon Moene wrote:
> Paul Richard Thomas wrote:
>
> >This problem came about because module equivelance groups with one or
> >more unused members were cleared. This being so, the remaining
> >members referenced variables in the module, rather than members of a
> >union representing the equivalence. Even if that were not the case,
> >the missing, unused members carry offset information that is essential
> >to obtain the correct alignment.
> >
> >The fix detects unused equivalence members, gives them a symtree with
> >a "unique name" that hides the symbol. Whilst completely unused
> >equivalence groups are now cleared, the hidden symbols remain. I was
> >unable to find a convenient, siple way of clearing the hidden symbols
> >and so decided to ignore them, since they get cleared up at the end of
> >compilation.
>
> Do you (Paul) or we (the Fortran group) think this patch is worthwhile
> to port back to the 4.2 branch, in spite of the fact that it isn't a
> regression (it has always been wrong) ?
>
(IMHO) If this is not a regression, then it should not be backported.
Of course, I'm only one voice in the choir.
--
Steve