Peter,
On 2005 Jan 26 , at 17.24, Peter W. Draper wrote:
> Spoke too soon. Seems some applications didn't link correctly, but at
> the
> same time didn't produce an error result, just messages saying:
>
> ld: multiple definitions of symbol _dummy_
> /Users/pwd/starlink_cvs/build/lib/libsubpar_adam.dylib(single module)
> definition of _dummy_
> /Users/pwd/starlink_cvs/build/lib/libgks.dylib(single module)
> definition of _dummy_
> ld: warning suggest use of -bind_at_load, as lazy binding may result
> in
> errors or different symbols being used
> symbol _dummy_ used from dynamic library
> /Users/pwd/starlink_cvs/build/lib/libsubpar_adam.dylib(single module)
> not from earlier dynamic library
> /Users/pwd/starlink_cvs/build/lib/libgks.0.dylib(single module)
>
> and then no executable. Back to the drawing board it seems.
As you'll probably have seen by now, that's a relic of my push just
before ADASS to get things working on OSX. The presence of the dummy.f
file, with just a dummy routine in it, was enough to persuade automake
to use $FC to do the link of the corresponding executable. I have a
better mechanism on a branch, so I'll shortly move that on to the
mainline and get rid of these wartish dummy files.
Norman
--
----------------------------------------------------------------------
Norman Gray : Physics & Astronomy, Glasgow University, UK
http://www.astro.gla.ac.uk/users/norman/ : www.starlink.ac.uk
|