Re: trouble with gtk--

[Havoc Pennington <rhp zirx pair com>]

On Wed, 23 Jun 1999, Dan Rosen wrote:
> 
> Boy, you don't miss a thing, Havoc. The error was from one of many
> compilation attempts (I did at one point swap GdkEvent* for GdkEventAny*).
> But if you close your eyes and pretend the GdkEvent* is a GdkEventAny*,
> it's the same exact error.
> 
> I guess I'm just looking for a little more tutelage in the zen of
> connecting signals (which are already provided by an ancestor in the
> inheritance structure) to member functions.
> 

Well, it's just a simple type error. You aren't passing the right type or
number of arguments to connect_to_method(). But the GdkEvent/GdkEventAny
type mismatch was the only one I saw.

It seems somewhat odd that it reports _GdkEventAny as the type in the
signal, but GdkEventAny elsewhere; maybe the GdkEventAny used in EccoGUI
is a forward declaration ("struct GdkEventAny;") instead of the typedef
from gdktypes.h? This is unlikely. I just don't know.

egcs often reports "No match for function blah blah" when it would be more
helpful to say "wrong type argument in position X," but since there are
many possible "connect_to_method()" functions it can only report the one
you asked for and say it's not there, it can't tell you which one you
meant. It needs to know which you meant to say you have the wrong type
rather than a nonexistent function. I hope that makes sense. :-) Anyway,
just usually assume "No match for" means you have the wrong argument
types. I guess it can also mean you forgot to include a header.

Havoc