Re: [Boost-users] [signals] Automatic connection management and MT

From: Frank Mori Hess (frank.hess_at_[hidden])
Date: 2007-02-03 11:17:48

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Friday 02 February 2007 19:12 pm, Timmo Stange wrote:
> Boost.Signals already solves this problem sufficiently well, by
> only delaying the connection removal if the disconnect() call
> originates from the signal call context 

My code will deadlock if the slot tries to disconnect() or otherwise 
manipulate the signal it is connected to.  The present semantics are that 
when disconnect() returns, the slot is disconnected and furthermore not 
running.  I suppose that could be weakened to "the slot is disconnected 
and not running in any other thread" by using recursive locks.

- -- 
Frank
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFFxLWw5vihyNWuA4URAhPDAKCU1IS+rvUx+F4N9F21vxhEuaAddwCgiGZZ
XkBsD0INMhFQoAtNoMlDz+4=
=j5LE
-----END PGP SIGNATURE-----