Re: [Nautilus-list] Re: nautilus & signals ...

[Alex Larsson <alexl redhat com>]

On 10 Apr 2002, Michael Meeks wrote:

> Hi Alex,
> 
> On Tue, 2002-04-09 at 16:46, Alex Larsson wrote:
> > Nice. But its sad that we have to fuck around with our design because 
> > GSignal is so slow. And it seems tim has almost stopped working on it, 
> > which is bad because he claimed to have lots of optimizations in store.
> 
> 	Well it's certainly a shame that signal emissions are not faster,
> presumably 99% of the notify emissions are not actually listened to by
> anyone, tens of thousands of them - it seems that handling the case
> where there is no-one listening to or implementing a signal is not
> overly fast - we get to do a chunk of lock/unlocks, unwind the argument
> va_list, etc. etc. looks like it certainly could be optimized. Anyway,
> there's no reason why Tim should be the only one that can optimize ?

I once tried to optimize the "null" signal emission. But it is really hard 
to figure that nobody is gonna recieve it in a cheap way, so it never got 
anywhere.

Of course tim is not the only one who can do it, but I was hoping he would 
implement the stuff he had planned.
 
> 	The other thing is that I must say I'm not convinced about the choice
> of overloading shed loads of signals on objects to implement complex
> functionality instead of just sub-classing it in the normal way. It
> seems to me a fairly obscure and complex way of doing something that is
> more legible, standard (and faster) to do with virtual methods and
> inheritance.

Sure.

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Alexander Larsson                                            Red Hat, Inc 
                   alexl redhat com    alla lysator liu se 
He's a short-sighted ninja rock star gone bad. She's a brilliant tempestuous 
single mother trying to make a difference in a man's world. They fight crime!