Re: [Fresco-devel] kit destruction not informing the server?

On Fri, Sep 13, 2002 at 09:34:33PM -0400, Nick Lewycky wrote:
> I'm tracking down the problem where clicking the Print button twice 
> causes the server to crash. The problem is that the server context 
> thinks it has a pointer to a valid postscript drawing kit when it doesn't.
> 
> KitImpl::decrement calls ~KitImpl when it detects that its own usage 
> count is zero. Unfortunately, nobody ever bothers to tell the server 
> context about that, so it never removes it from 
> ServerContextImpl::_kits. The only time it ever removes an entry from 
> _kits is in its own destructor, which means that the server is being 
> shut down.

Ah, interesting.

> I've heard that this lifecycle stuff can be complicated so I'm asking 
> what the correct fix is. I see two options:
> 
> a) make the ServerContextImpl test that the kit is still valid before 
> trying to make calls on it. (Not certain myself how to do that.)
> 
> b) make the KitImpl hold on to a ServerContext_var and add a function to 
> the API that allows the Kit to inform the server context that the kit is 
> now gone
> 
> Or is there supposed to be another mechanism that I simply failed to 
> notice? I have a hard time believing that kit destruction never worked.

I'm not an expert on this, so Stefan can thwack me if I spread
incorrect information, but:

I wouldn't be too surprised if kit destruction never worked --
normally, the kits should only be destructed when a client exits, at
which point it's unlikely that that client's kits will be used again.
So there could well be no crash.

Why don't you just have the ClientContext hold a reference to the
kits?

(In general, I suppose we can't count on a malicious/buggy client
double-dereferencing, though...)

-- Nathaniel

-- 
"But in Middle-earth, the distinct accusative case disappeared from
the speech of the Noldor (such things happen when you are busy
fighting Orcs, Balrogs, and Dragons)."

This email may be read aloud.