Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!uwm.edu!rpi!zaphod.mps.ohio-state.edu!usc!snorkelwacker!bloom-beacon!ATHENA.MIT.EDU!swick
From: swick@ATHENA.MIT.EDU (Ralph Swick)
Newsgroups: comp.windows.x
Subject: Re: SYSV shared Xaw and XtInheritTranslations
Message-ID: <9005081303.AA17670@lyre.MIT.EDU>
Date: 8 May 90 13:03:23 GMT
References: <10118@ingr.com>
Sender: daemon@athena.mit.edu (Mr Background)
Organization: DEC/MIT Project Athena
Lines: 18

> externalref int _XtInheritTranslations;
> #define XtInheritTranslations  ((String) &_XtInheritTranslations)
> ...
> Is the reason that is done as opposed to something like
>     #define XtInheritTranslations  -1
> so that the item is truly a data pointer to prevent problems
> for certain architectures?

Yes, and also because according to ANSI C there seems to be only
one portable (integer) pointer constant with a defined value (0)
and we need a second unique constant.

> does anyone see any problem
> with me changing it for my libraries and binaries?

No problem; you're certainly allowed to use any architecture-specific
knowledge you may have.  Other small positive integers are probably
preferable to -1 though.