Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!sunybcs!boulder!hao!oddjob!gargoyle!ihnp4!alberta!ers!teletron!andrew
From: andrew@teletron.UUCP
Newsgroups: comp.lang.c
Subject: Initialization of Externals (was Re: documentation standards...)
Message-ID: <123@teletron.UUCP>
Date: Tue, 27-Oct-87 19:08:23 EST
Article-I.D.: teletron.123
Posted: Tue Oct 27 19:08:23 1987
Date-Received: Sat, 31-Oct-87 09:15:08 EST
References: <9897@brl-adm.ARPA> <6917@prls.UUCP>
Organization: TeleTronic Communications Ltd., Edmonton, Alta.
Lines: 30
Summary: variable initialization

In article <9897@brl-adm.ARPA> tom@vax1.acs.udel.EDU (Tom Uffner) writes:
>all externals and statics can be
>initialized "for free" to nearly anything desired at compile time,
>(cf. constant expressions) but an 'initialize()' proc would generate
>code and take time to execute.

Sometimes you have no choice..  (see below)

In article <6917@prls.UUCP>, gardner@prls.UUCP (Robert Gardner) writes:
> I prefer compile-time initialization when possible, but it's not
> necessarily "for free".

This is true when dealing with ROM-able code.  You are forced to write
explicit 'initialize()' procedures.

Our compiler (like most others) generates different segments for different
portions of code:
	.text - program instructions
	.data - initialized externals
	.bss -  uninitialized externals
We bind the .text and .data sections to ROM addresses and .bss sections to
RAM addresses at link time.  Thus, we are forced to allow initialized externals
for static data structures only.  Variable data must be initialized by an
'initialize()' routine called from the startup code.

I realize this is a digression from the original thread, and Tom's original
point about documentation of variable initialization is well taken, but one
can't always use compile-time initialization.

	Andrew