Re: 1.7.2: scmsigs.c,signal_delivery_thread

[Werner Scheinast <W.Scheinast@web.de>]

Am 07.05.05 schrieb KevinRyde:

> Werner Scheinast <W.Scheinast@web.de> writes:
> >
> > -Wall
> 
> Did you add that?  It's probably not a good idea.

No, it was done automatically. (Or can there be a system-wide default 
option string?) I only added some processor optimization.

> > scmsigs.c: In function `signal_delivery_thread':
> > scmsigs.c:152: warning: no return statement in function returning non-void
> 
> I don't think that function ever returns, gcc probably shouldn't
> complain about such cases.

I wonder why he reports a mere "warning" and then stops with a fatal 
error.

> I guess it could be void, but the function pointer is used in a spot
> where the prototype should be an SCM return.

When I say "void", he exits with another message:

 gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I.. -pthread -O3 -march=pentium4 
-mfpmath=sse -msse2 -Wall -Wmissing-prototypes -Werror -MT 
libguile_la-scmsigs.lo -MD -MP -MF .deps/libguile_la-scmsigs.Tpo -c 
scmsigs.c  -fPIC -DPIC -o .libs/libguile_la-scmsigs.o
scmsigs.c:125: error: two or more data types in declaration of 
`signal_delivery_thread'
scmsigs.c: In function `start_signal_delivery_thread':
scmsigs.c:160: warning: passing arg 1 of `scm_spawn_thread' from 
incompatible pointer type
make[2]: *** [libguile_la-scmsigs.lo] Fehler 1
make[2]: Leaving directory `/home/werner/Archiv/guile-1.7.2/libguile'
make[1]: *** [all-recursive] Fehler 1
make[1]: Leaving directory `/home/werner/Archiv/guile-1.7.2'
make: *** [all] Fehler 2

When I follow Paul's suggestion and add "__attribute__((noreturn))" before 
the function name (right?), it doesn't change anything.

Perhaps the mistake is somewhere else ... in my SuSE version of the gcc 
compiler or so ...?

Thanks anyway for the tips
Werner


_______________________________________________
Bug-guile mailing list
Bug-guile@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-guile