Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!sun-barr!newstop!sun!gazooch!brianw
From: brianw@gazooch.Sun.COM (Brian Warkentine)
Newsgroups: comp.windows.x
Subject: Re: Defining Callbacks from resource files
Message-ID: <129431@sun.Eng.Sun.COM>
Date: 19 Dec 89 20:18:28 GMT
References: <38@guug.UUCP> <8912191502.AA22891@LYRE.MIT.EDU> <129425@sun.Eng.Sun.COM> <15040@boulder.Colorado.EDU>
Sender: news@sun.Eng.Sun.COM
Reply-To: brianw@sun.UUCP (Brian Warkentine)
Organization: Sun Microsystems, Mountain View
Lines: 17

In article <15040@boulder.Colorado.EDU> grunwald@foobar.colorado.edu writes:
>>>>>> On 19 Dec 89 18:45:23 GMT, brianw@gazooch.Sun.COM (Brian Warkentine) said:
>
>Action procs are bound at compile time; a string -> proc converter
>would need to load the symbol table to determine the function address.
>This is O/S dependent. Also, if the symbol table has been striped, it
>won't work.

	You missed the point:  binding functions (aka "actions procs") to 
	function names is already being done with action procs.  What makes 
	doing it with callback procs any more system dependent?  You're 
	reading "arbitrary function binding" into what I wrote.  I agree,
	it's not portable.  Furthermore, I don't believe this is necessary or 
	even desirable.

-brian warkentine
brianw@sun.com