Re: [Evolution-hackers] Loading of PKCS#11 module

[Milan Crha <mcrha redhat com>]

On Tue, 2014-08-12 at 08:57 -0300, Watson Sato wrote:
> So, I think I'll stick with SECMOD_LoadUserModule() for now, is that 
> ok?
> Also, there is a segfault when unloading the module, high chances of

> being related to glib not being unloadable [0] (still need further
investigation). So there is a chance the module will be loaded all
>
the time while Evolution is running.

        Hi,
sure, it sounds reasonable to have the module loaded all the time.

> Maybe the place for the module to be loaded is in shell/main.c?
> Or in camel/camel-smime-context.c when actually needed to fetch a 
> certificate.

I would do changes only in evolution-data-server. Either load the 
module on demand (which sounds like a good plan), or load it during 
the camel_init() call (eds/camel/camel.c). There is a camel_shutdown() 
counterpart too, but if you cannot figure out why the segfault happens 
then do not unload it (maybe the NSS_Shutdown() unloads it for you).

By the way, is the code publicly available anywhere, thus others can 
look in it?
        Bye,
        Milan