Path: utzoo!utgpu!attcan!uunet!husc6!bloom-beacon!tut.cis.ohio-state.edu!cwjcc!hal!nic.MR.NET!tank!uxc!uxc.cso.uiuc.edu!m.cs.uiuc.edu!s.cs.uiuc.edu!hummel
From: hummel@s.cs.uiuc.edu
Newsgroups: comp.sys.amiga.tech
Subject: Re: Manx precompiled include files
Message-ID: <212300001@s.cs.uiuc.edu>
Date: 2 Nov 88 16:12:00 GMT
References: <8327@gryphon.CTS.COM>
Lines: 35
Nf-ID: #R:gryphon.CTS.COM:8327:s.cs.uiuc.edu:212300001:000:1518
Nf-From: s.cs.uiuc.edu!hummel    Nov  2 10:12:00 1988


Written  9:20 am  Oct 31, 1988 by collins@pnet02.cts.com in comp.sys.amiga.tech:
/* ---------- "Manx precompiled include files" ---------- */
> I am using Manx 3.6 a which supports pre-compiled include files.
> [ ... ] When I try to precompile a file containing somthing like:
>
> struct my_structure structure1= {
>                                        a bunch of structure intializers
>                                 };
> and then subsequently link it, I get a global called structure1, but
> it is uninitialized. 
/* End of text from comp.sys.amiga.tech */

Precompilation works by dumping the symbol table to a file and reading it
back in.  There's no initialization information contained in the symbol
table, hence none gets saved.  Paul and I wound up doing precompilation in
PDC 3.3 this way, and I'd imagine Jim Goodnow did the same when he implemented
Manx.  The bottom line is that you can't have initialized data in a pre-
compiled header.

In the same posting, he asks:
> What's the smart way to initialize a bunch of structures that contain
> 20-30 element arrays of LONGS?  I mostly want to avoid making the
> compiler chew on them every debugging cycle.

Compile them in their own .c file and link with the resulting .o file.
How's that?
				< Lionel
----------

Lionel Hummel					404 W. High St., #6
ldh90267@uiucuxa.cso.uiuc.edu			Urbana, IL  61801
{seismo,pur-ee,convex}!uiucuxa!ldh90267		(H)  (217)344-5303
Dept. of Computer Science			(W)  (217)333-7408
University of Illinois, Urbana-Champaign