From: Aleksey Gurtovoy (alexy_at_[hidden])
Date: 2000-04-10 13:15:50
Karl Nelson <kenelson_at_[hidden]> wrote:
> Hmm, did you ask about this on the mailing list?
No, I didn't. Mostly because there were so many references to the necessity
of derivation from SigC::Object in the documentation that I considered it to
be true :).
> It might
> have saved you a lot of coding.
Not really. Our home-grown signals/slots system is about 700 lines of code.
The missed (in comparison with libsigc++) functionality is a support for
signals with 2 and more arguments. That probably makes the main difference
in sizes. Signals with only one argument are sufficient for us because we
use event objects anyway :).
As you noted before, there are many issues on the topic which need to be
resolved and which can be resolved in many ways. Our particular
implementation went in a different way (and were built at all) because the
points we considered to be important were different.
And I do agree that libsigc++ is a very good piece of work and can be the
source of many ideas, knowledge and experience which we can hardly get from
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk