[csw-maintainers] Code issue w/ different behaviour on sparc and x86

[Peter FELECAN]

Laurent Blume <laurent at opencsw.org> writes:

> All,
>
> I've just noticed an issue with glib: during the configure part, on
> x86, header inlining is enabled. On Sparc, it is not. They end up
> having slightly different headers installed.
> This can have some effect on dependencies: pidgin-sipe builds without
> a hitch on x86, fails on sparc, because of that.
>
> I've tracked the issue down to the example attached, straight from the
> configure script.
>
> It fails like that on sparc, but no error on x86:
>
> /opt/solarisstudio12.3/bin/cc -o conftest -xO3 -m64
> -features=extensions -xc99 -D_XPG6 -I/opt/csw/include -m64 -lsocket
> -R/opt/csw/lib/$ISALIST   -R/opt/csw/lib/64 -L/opt/csw/lib/64
> conftest.c -lintl
> Undefined                       first referenced
>  symbol                             in file
> glib_test_func2                     conftest.o
> ld: fatal: symbol referencing errors. No output written to conftest
>
> Does anybody know of a clean workaround?

Did you try using GCC? What purpose do you pursue using Sun Studio?
(Note that I'm not trying to start a compiler war, just trying to
understand).

> Also, can OpenCSW report the issue, as it seems a clean bug? Or should
> I do it on my side?

IMHO you do it on your side, i.e., OCSW is not doing it on your behalf;
dont't forget: you are OCSW...
-- 
Peter