Re: Dynamic runtime failure.

[Kevin Ryde]

[All bug matters to bug-gmp, not to me.]

"Carlos D. Garza" <address@hidden> writes:
>
> Has any other linux users noticed this?

It's unlikely anybody uses the linux kernel alone, but if you mean a
GNU/Linux system, then no.

> Considering the sysmptems I described do you think this is a linker problem , 
> a loader problem or what?

The loader would be the likely candidate, nothing in libgmp.so should
be run if there's no references to it.

> I'm not getting a core dump or anything just a 
> segfault very early. Considering I only have control over my own code I have 
> no idea what could be going on.

You can put it under gdb to see where it dies.

You can also try with some other shared library, link against it but
don't call it, to see if the problem still happens.  I'd be surprised
if whatever is happening is specific to gmp.

(The only potentially hairy possibility with gmp might be if you've
got an old libgmp.so from gmp 2.  We never created a shared library in
gmp 2, but some gnu/linux distributions did.  The soname ought to
ensure it's not hit though.  "ldd a.out" can tell you what's used.)

-- 
All followups to address@hidden only.