Segmentation Fault

[Moinak Ghosh]

Hi Frey,

   I recompiled without the threading option and tried, but still got
that segmentation fault. I will get the latest gdb and try to locate 
the problem.

Regards,
Moinak.

Frey, Gregor Karl Dr. wrote:
> 
> Hi Tom,
> I tried to build my sample-programm statically with: "gcj -Xlinker -Bstatic
> -lgcj -Xlinker -Bdynamic --main=sample -g -ggdb -Wall -o sample
> -sample.java"
> But the result remains the same.
> Is it a threading-problem as Moinak Ghosh assumes? I did not configure with
> a threading-option (neither gcc nor libgcj).
> Ciao,
> Gregor
> 
> -----Original Message-----
> From: Tom Tromey [ mailto:tromey@cygnus.com ]
> Sent: Samstag, 7. August 1999 03:56
> To: Kresten Krab Thorup
> Cc: Frey, Gregor Karl Dr.; java-discuss@sourceware.cygnus.com
> Subject: Re: Segmentation Fault
> 
> >>>>> "Kresten" == Kresten Krab Thorup <krab@gnu.org> writes:
> 
> Kresten> What is happening here is, that the "module initializer" for
> Kresten> the module natClass.cc has not yet been run.  (I.e., the
> Kresten> logic that also does C++ file-scope initialization).  This
> Kresten> renders the variable `init_name' and `init_signature'
> Kresten> unititialized (which asre the null values you observe).  This
> Kresten> is strange, since you've reached main, and this stuff happens
> Kresten> in __main (which is secretly called as the first thing in
> Kresten> main).
> 
> This makes it sound like either a Solaris runtime (ld.so) bug, or
> perhaps a bug in the linker being used.  Gregor, what linker are you
> using?  Can you try a static link and tell me if it works for you?
> 
> Kresten> I had the same problem on Solaris a while back, linking
> Kresten> statically solved the problem, but that's a pain of cause.
> 
> Especially since the LGPL is really a pain when static linking.
> 
> Tom