Does gdb/configure ignore --with-gmp?

[Stephen Casner]

On Mon, 15 Feb 2021, Maciej W. Rozycki wrote:
> On Wed, 3 Feb 2021, Stephen Casner wrote:
>
> > > > Thank you for the prompt reply.  That option is not listed by the
> > > > top-level configure -h which has only this to say about gmp:
> > > >
> > > >   --with-gmp-dir=PATH     this option has been REMOVED
> > > >   --with-gmp=PATH         specify prefix directory for the installed GMP
> > > >                           package. Equivalent to
> > > >                           --with-gmp-include=PATH/include plus
> > > >                           --with-gmp-lib=PATH/lib
> > > >   --with-gmp-include=PATH specify directory for installed GMP include files
> > > >   --with-gmp-lib=PATH     specify directory for the installed GMP library
> > > >
> > > > Is that a (minor) bug then?  Or is --with-gmp the proper form for gcc?
> > >
> > > I'll pile on to this one. I did manage to find the right option in
> > > gdb/configure after some digging into the error message. I don't
> > > think anyone not willing to dig would find it at all. Not being in
> > > sync with the top level configure help is quite confusing.
> >
> > Thanks for concurring.
>
>  FYI, there's the `--help=recursive' option to use for the very case like
> this where subdirectiories have their own `configure' scripts.  It's there
> for a reason.

Thanks for that hint.

> > > Overall, this is very unlike how GCC handles this one and awkward.
> >
> > Would it be reasonable to change the top-level configure to transfer
> > --with-libgmp-prefix to gdb/configure if --with-gmp is provided to the
> > top-level?  I think both have the same semantics.  That would avoid
> > the need to introduce another potentially confusing entry in the
> > top-level help.
>
>  The top-level option has been there since 2004 while the GDB addition is
> quite recent, so if any way, I would suggest changing GDB to use the same
> name and semantics for the option.

That would be a fine solution as well, but I anticipate that the GDB
group would be reluctant to change, or at least reluctant to remove
the --with-libgmp-prefix option.  Perhaps they would accept supporting
--with-gmp in parallel.  How should that question be directed?

                                                        -- Steve