Re: [open-cobol-list] gdb support from gcc-cobol

[Bernard Giroud]

Keisuke Nishida a écrit :

> At Tue, 03 Jun 2003 08:19:25 +0200,
> Bernard Giroud wrote:
> >
> > > Perhaps this may be naive, but shouldn't the above be converted into a
> > > structure ?
> > >
> > > struct elem_G
> > > {
> > >   unsigned char X[2];
> > >   unsigned char Y[2];
> > > };
> > >
> > > static struct elem_G f_G;
>
> I initially thought about it, but how would you represent
> REDEFINES and RENAMES?

REDEFINES only inherits the address of the redefined item
address; it does not induce a particular format (might be
a group or an elementary). RENAMES also, except
it can span over several items, IFAIR.

>  How would you implement reference
> modification such as G(2:2)?
>

This is an implicit X(2) starting at the address G+2;
so for gdb, only G should be known, and the work
of G+2 for 2 chars must by done by the language
front-end of gdb. For the descriptors, we could
find a f_static_refmod_nnnn field with the correct
content.

>
> COBOL assumes the data field is sequential and it is accessed
> in a variety of ways.  I think trying to translate COBOL data
> items into a C structure makes things complex.
>

I agree it is a little more complex, but it is much more
natural. Another aspect: this is probably a point of
divergence between OC and gcc-cobol (aka Gobol).

>
> Keisuke
>

--
Bernard Giroud
TinyCOBOL and Gobol Developer