Re: Once more: -static linking

[rudolf]

Hi Gaius,

a few days ago I sent some 'example files' to you which show
my 'static linking problem'. Did the files reach you or were
you able to reproduce the problem?


Thanks

Rudolf

--
Rudolf Schubert                 \
Kirchstr. 18a                    \  mailto:rudolf@muc.de
82054 Sauerlach                   > http://www.dose.muc.de
Deutschland                      /   (alpine archkiste)
Tel. 08104/908311               /


On Thu, 19 Oct 2023, Gaius Mulley wrote:

> rudolf@muc.de writes:
>
> > Hi Gaius,
> >
> > quite some time ago (20 May 2023) I reported problems with static linking
> > for my 'big squash1 program' (which I sent to you back in April). But I did
> > not find out how to work around these problems until today.
> >
> > I played a bit with command line parameters and found that if (additionally)
> > using '-fm2-whole-program' static linking is working!
> >
> > As reported, for dynamic linking I use this command:
> >
> > gm2 -fiso -fsoft-check-all squash1.mod Ctv2.o unix.o -o squash1 -lc -lcrypt -lm
> >
> > which is working fine.
> >
> > Then I added '-static' and  got the reported errors:
> >
> > gm2 -fiso -fsoft-check-all squash1.mod Ctv2.o unix.o -o squash1 -lc -lcrypt -lm 
> > -static
> >
> > Now I use:
> >
> > gm2 -fm2-whole-program -fiso -fsoft-check-all squash1.mod unix.o -o squash1 -lc 
> > -lcrypt -lm -static
> >
> > which in fact produces a static binary!
> >
> > Note that Ctv2.o is missing from the command line. Otherwise I get
> > 'multiple definition' of certain symbols.
> >
> > My question now is: is this a 'valid workaround' for my problem or should
> > '-static' also work without '-fm2-whole-program'?
>
> Hi Rudolf,
>
> It should produce a static binary without -fm2-whole-program.  What
> commands were used to generate objects Ctvt.o unix.o ?
>
> If you are able to email me the objects or source - I'll investigate,
>
> regards,
> Gaius