[gtkmm] I/O Connection::disconnect(): solved

[Eric Buddington <eric ma-northadams1b-39 bur adelphia net>]

One cause of my problems was a double-call of connect(), one of which
did not store the return value. I never did not determine why I got
poll()s but no callbacks, but my drive to debug the hell out of it is
less then my urge to go biking right now.

I have submitted a bugzilla report to mention the effect of the
callback's return value.

Sorry for distracting the list with my internally-generated bug...

-Eric

On Thu, Sep 26, 2002 at 10:18:32AM +0200, Martin Schulze wrote:
> Am 26.09.2002 06:39 schrieb(en) Eric Buddington:
> > On Wed, Sep 25, 2002 at 09:56:55PM +0200, Martin Schulze wrote:
> > > connection.disconnect() and returning 'true' in the callback function
> > > should have the same effect, i.e. stop further calls of the callback
> > > function, remove the file descriptor from the poll table and internally
> > > destroy the 'connection node' and the 'source' structure. No recovery
> > > is possible as all information about the filedescriptor to poll and
> > > the callback function to call is deleted.
> > 
> > Are you certain that a return value of 'true' acts as disconnect()? I
> > had previously assumed that returning 'false' would do this (guessing
> > from old GTK+ timeout docs), and it mostly worked. Swapping logic per
> > your description above (true => disconnect) seems to break things more.
> 
> I didn't look it up but assumed that from your previous mail. I'll look
> into the docs now... Done...
> 
> "It should return false if the source should be removed."
>                   ^^^^^^^
> 
> Sorry for the confusion! Please mentally replace 'true' with 'false' in
> my reply and check whether it is working for you. Otherwise please file
> a bug report and supply a test case.
> 
> Thanks,
> 
>    Martin