On Fri, Apr 11, 2003 at 05:17:23PM +0200, Martin Pitt wrote: > Hi Sven and others, > > Am 2003-04-11 17:09 +0200 schrieb Sven Luther: > > > > and is not possible or good if you have C bindings included, like > > > > the advi package does for example. The correct way is to either say > > > > you don't care and package a -custom bytecode or a native code, both > > > > arch: any, or to separate the binding library in its own package, > > > > which will be arch: any and provide both the bytecode and nativecode > > > > libraries, and build the app depending on this library. > > > > > > I do not fully understand that but I will investigate what C-Bindings in > > > ocaml programs are. > > > > Do you have any .c source, and does the result of the makefile run > > produce any file of the kind dll***.so ? > > > > C bindings are in reality composed of two pieces, one the so called > > stublib is a small source file which wrapps calls to an external > > C library to functions palatable by the ocaml-C interface, and the other > > is the ocaml source file that makes calls to the stublib. > > Ah! No, it's really just plain ml code, no other languages involved. > I think that settles this issue. I also checked with ldd, nothing > unusual.
There is a sqrt.c file, but i guess it is not built, that said, it is just an executable and not a library like any other uses of rpath that where detected. I suppose the : planets: binary-or-shlib-defines-rpath ./usr/bin/planets /usr/lib:/usr/X11R6/lib got pulled in by the use of labltk. Anyway, nice little program, but the planets have a tendency of going offscreen :((( Friendly, Sven Luther
