At 10:28 AM 5/12/2007, Tobias Schlüter wrote:
My inclination would be to do the cosmetic and small fixes to this
and go ahead and commit it and then work on the
alternate-character-set hypotheticality and the derived-type spacing
stuff in follow-up patches, since AFAIK this doesn't cause anything
to break that worked previously. But that's not a strong
inclination, and I'll let Paul and FX make the call on that. :)
I think we should assert that we're not dealing with derived types,
since otherwise we would leave the user with a potentially
hard-to-find error. Apart from that, I agree with this course of action.
That sounds like it might be a good idea.
If we do that, my suggestion would be to take the old code at the head
of gfc_simplify_transfer that returns NULL unless gfc_init_expr, and
make that conditional on SOURCE or MOLD being a derived type. That way,
instead of ICEing on a derived type argument that could be handled at
runtime, we'll leave it unsimplified and let the runtime handle it.
Something like (in semi-pseudocode):
/* Derived types are not known to be handled correctly yet. */
if (type(source) is BT_DERIVED || type(mold) is BT_DERIVED)
{
if (gfc_init_expr)
gfc_error ("TRANSFER intrinsic not implemented for initialization"
" at %L", &source->where);
return NULL;
}