Re: [gtkmm] retype() and retype_return()

[MHL Schulze t-online de (Martin Schulze)]

Am 27.07.2002 11:01 schrieb(en) Murray Cumming:
> On Fri, 2002-07-26 at 20:23, Martin Schulze wrote:
> > Am 25.07.2002 18:36 schrieb(en) Murray Cumming:
> > > What functionality do you mean? Bear in mind that I intend to keep
> some
> > > kind of hide_return() and bind_return() functionality.
> >
> > I may sound like a broken record but:
> > There are only bind, bind_return, hide, retype, retype_return, convert.
> > It's really not that much! I consider these adaptors all of equal
> > importance. Adaptor one to five in my enumeration form some kind of
> > group
>
> bind and bind_return are about dictating _values_.
> hide and hide_return are about _ignoring_ parameters
> retype and retype_return are about changing the _types_ of parameters.
>
> These are 3 distinct groups, though some might be implemented internally
> in terms of others.
>
> > - if you remove one of them there is a lack in functionality.
> > So you may either remove them _all_ of them or none - it would be very
> > inconsistent to leave two of them that are chosen by mere chance
>
> bind() is clearly the most popular. It is used more and asked about
> more.

At least hide() must be used nearly as often - I introduced it for sigc++ 
1.1
and up to now there were some 10-20 people who wanted my implementation for
the stable sigc++ 1.0 (doesn't seem much but if you look at the traffic on
the sigc++ ml ...)
As for retype we can't say how much it _will_ be used because it is new
in libsigc++ 1.1.

Regards,

  Martin