another DL problem

[Ulrich Drepper]

I'm currently reviewing the next draft for the next POSIX/XPG
standard.  And the dlclose description says that the implementation
has to track dependencies established by the relocation and that this
must prevent an object from actually being unloaded until the objects
with the dependencies are also gone.  Another triumph of people doing
stupid things.

The test case currently fails but the make will not abort.  Take a
look at the test in dlfcn.

-- 
---------------.      drepper at gnu.org  ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------