Path: utzoo!utgpu!attcan!uunet!auspex!guy
From: guy@auspex.UUCP (Guy Harris)
Newsgroups: comp.windows.x
Subject: Re: X 10 for the Sun 4: Does it exist? (clarification)
Message-ID: <396@auspex.UUCP>
Date: 3 Nov 88 18:07:19 GMT
References: <21943@cornell.UUCP> <21944@cornell.UUCP> <22278@cornell.UUCP>
Reply-To: guy@auspex.UUCP (Guy Harris)
Distribution: comp
Organization: Auspex Systems, Santa Clara
Lines: 18

>Finally, one person suggested that there is a problem compiling the X
>server under SUN OS 4.0 --- Something about the code assuming that
>locals are initialized to 0, and the compiler not so initializing them.

I know of no compilers that so initialize them.  It just so happens that
on UNIX implementations, when stack pages are created they are filled
with zeroes, which means that the *first* time you use a stack page, the
locals are likely to be zero.

Unfortunately, in SunOS 4.0, "main()" and its descendents are usually
*not* the first guys to use the first stack page or so; the run-time
loader gets there first, and the result is that they aren't filled with
zeroes when "main()" and procedures it calls use them.

Neither the C language nor UNIX guarantee that locals will be
initialized to any particular value - and, if some routine is called
more than once, there is a good chance that they will *not* be
initialized to zero on calls after the first.