Hi Ben,

On Sunday 06 May 2007, Ben Goodrich wrote:
> Hope things are well. To replicate this crash, execute
>
> det(diag(3))
>
> I would send a backtrace but rkward does not seem to play nice with gdb on
> my system. Perhaps this is another bug?

yes, this is not easy, since there are actually two wrapper scripts being 
called, and gdb isn't smart enough to handle this. R has a "--debugger" 
option for this, perhaps I should add something like this, too, sooner or 
later.

> All I've got is the konsole message 
>
> rkward: symbol lookup error: /usr/lib/R/lib/libRlapack.so: undefined
> symbol: _gfortran_copy_string

This should be the relevant error message, though. It's a symbol lookup 
problem, so you wouldn't see too much in gdb. After upgrading several 
portions of my system, I can reproduce the problem, but not sure, where it 
really comes from in the first place (worked smoothly shortly before).

Either way, I have found a fix (explicitely linking against libgfortran). I'll 
try to prepare a bugfix release, soon. Meanwhile, you could try the SVN 
version (see http://rkward.sourceforge.net/wiki/index.php?title=RKWard_SVN ; 
you need the make -f Makefile.svn stage, even if you have done this, 
previously).

Regards
Thomas

Attachment: pgpguAVmKbRek.pgp
Description: PGP signature

Reply via email to