Re: Plugins redux (was Re: [PATCHES] PL instrumentation

> > Also, should we create an on_proc_exit() handler that would unload all
> > dynamic libraries (specifically to call the _PG_fini() functions)?
>
> Yeah, I thought about that too, but I'm inclined not to do it;
> it seems like just excess cycles. The process is quitting anyway,
> so the only reason this would be useful is if the library thinks it's
> going to update external or shared state during _PG_fini ... and on
> the whole that sounds like a bad idea. Besides, if a library really
> needs this it can add its own on_proc_exit handler.

It seems a little dangerous for a dynamic library to register an
on_proc_exit() handler. If we ever add support for unloading a dynamic
library, we'll have to add an unregister_on_proc_exit() too. Otherwise,
a dynamic library might register a function pointer with on_proc_exit()
and then leave a dangling pointer when it gets unloaded.

Given that, I assume you don't feel the need to unload old shared
libraries if the user (a superuser) removes an entry from
backend_load_libraries, right?

In fact, it looks _PG_fini() is only called if you *reload* a library,
unless I'm missing something somwhere.

--
Korry Douglas korryd(at)enterprisedb(dot)com
EnterpriseDB http://www.enterprisedb.com