[omniORB] Problems shutting down OmniORB

[Duncan Grisby]

On Monday 29 September, Nickolay Samofatov wrote:

> Research shows that if DllMain with DLL_PROCESS_DETACH is invoked from
> TerminateProcess when all application threads including OmniORB
> service threads are already forcedly terminated by Windows.
> OmniORB is endlessly waiting for already terminated threads to exit
> normally in this case.

This problem has been discussed on the mailing list before. If I
recall correctly, nobody could come up with a good solution, because
Windows just doesn't give you the hooks you need to cleanly shut down
omniORB before Windows rips bits of it out from underneath it.

> Another question is where do I find rules about patches format to be
> accepted by OmniORB development group if I have to fix explained
> problem myself.

If you do manage to come up with something, and it doesn't
significantly impact other platforms, please submit a patch in unified
diff format.

Cheers,

Duncan.

-- 
 -- Duncan Grisby         --
  -- duncan at grisby.org     --
   -- http://www.grisby.org --