Re: [Fortran-Experiments]: patch

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Mon, Feb 26, 2007 at 11:24:39AM -0700, Christopher D. Rickett wrote:
> 
> >>also, i wanted comments about binding labels in general.  right now, i
> >>only consider the binding label field if the symbol is BIND(C) or from
> >>the iso_c_binding module.  however, if the compiler wants to assign a
> >>binding label w/o being either of these, then i can simply consider
> >>the binding label if it's non-null.  this should easily allow setting
> >>of binding labels for library routines.  does anyone have objections
> >>or comments?
> >I'm not sure whether I understood this. What do you mean by "if the
> >compiler wants to assign a binding label" and "consider the binding
> >label if it's non-null". How do you get a binding label without using
> >bind(c,name="label") ?
> 
> the idea is to allow for binding labels for internal compiler use without 
> having to make the symbol bind(c).  right now i do that for some of the 
> iso_c_binding procedures.  i thought it could be useful for the compiler 
> to be able to set a binding label if it wanted to.
> 
> >In any case I'm slightly opposed to adding compiler-specific extensions,
> >but maybe it makes sense after I understood the proposal.
> >
> 
> i don't think this is really an extension since it's not available to the 
> user; i was just thinking that the compiler could do it for itself 
> internally.  is that more clear?
> 

I vaguely recall that this might be something that FX wanted to
use with the IEEE754 module.  Unfortunately, FX is unavailable
for another week or two.

-- 
Steve