NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/comp/lang/cplus/13052

Text File | 1992-08-29 | 2.0 KB | 49 lines

Newsgroups: comp.lang.c++
Path: sparky!uunet!usc!elroy.jpl.nasa.gov!decwrl!borland.com!pete
From: [e-mail] (Pete Becker)
Subject: Re: Callbacks - C++ needs an extension?
Message-ID: <[e-mail]>
Originator: [e-mail]
Keywords: callback event eventdriven extension
Sender: [e-mail] (News Admin)
Organization: Borland International
References: <[e-mail]>
Date: Sat, 29 Aug 1992 16:13:56 GMT
Lines: 35

In article <[e-mail]> [e-mail] writes:
>Hi guys!
>
>Maybe some of you have, like me, done some windowprogramming in
>C++. There are a lot of toolkits to choose from, but whatever
>you choose, wether it is a C++-package like TNT or a C-package
>like xview, you'll probably stumble upon the notion of call-
>backs.
>
>Callbacks are essencially a functionpointer you deliver to
>the windowing toolkit when creating a widget, for instance
>a button. This function is then supposed to be called when
>the user have taken some action, like pressing the button.
>
>Their existance is motified by the nature of event-driven
>programming and that it is most unpractical to have to edit
>sourcecode in the toolkit when you want a special function in
>your source to be called.
>
>Most often a button is connected to some sort of other
>object, like a window, and it is most often the window that
>knows how to response when the button is pressed.
>
>The problem is that it seems impossible in C++ to implement
>a callback strategy which calls this window-object directly...
>

    The problem is not in C++ but in libraries that constrain callback
functions to look like ordinary C functions.  Such libraries are old
technology, and making newer techniques work will always require a certain
amount of work.
    The most straightforward solution is to work with the library and
provide a callback function that examines the message being sent and forwards
it to the appropriate recipient.
    -- Pete