[1218] in Coldmud discussion meeting

root meeting help first first in chain previous in chain previous next last

Re: [COLD] WIN32s

daemon@ATHENA.MIT.EDU (Sun Jan 26 15:40:58 1997 )

Date: Sun, 26 Jan 1997 13:35:16 -0700 (MST)
From: Brandon Gillespie <brandon@cold.org>
To: Stephen Smoogen <smooge@shout.net>
cc: "Jon A. Lambert" <jlsysinc@ix.netcom.com>, coldstuff@cold.org
In-Reply-To: <Pine.LNX.3.95.970126135639.9777C-100000@duracef.shout.net>

On Sun, 26 Jan 1997, Stephen Smoogen wrote:
> On the gdbm port you may have problems due to it having been ported with
> 16 bit access in mind versus 32 bit access.I ran into something similar 4
> years ago... and it took forever to track down. I would not be sure that
> the 2 binary databse outputs will be identical. I am not sure that 2
> Unixes (Say BSD and Linux) would output the same binary databses on the
> same machine.

The binary databases will NOT be the same, unless you can guarantee that
the same ndbm package is used.  Genesis will use ndbm by default, over
gdbm.  ndbm is (by default) an OS-provided package.  Some OS's use gdbm
for their ndbm implementation, but this is not a rule.

So yeah, what he said :)


> > I attempted to get p22 running on Win95 using the BorlandC 5.01 compiler
> > and a DOS port of GDBM.  The Coldcc program appears to work and the 
> > Genesis driver actually loads the data base and begins to start execution.
> > However I am encountering errors in method sizes read from the database
> > in the emalloc routine.  I compared the binary output of Linux and Win95
> > generation of the MinimalCore4.0 and there appear to be differences due
> > to uninitialized memory(?).  I know very little Unix and much less about
> > Cold design.  Here's my driver log just to show you how far I've gotten
> > with it.  I suspect that once I get an exact match on the DB with Linux's
> > DB most all my problems will be solved.  Any information on the
> > format of the binary would be very helpful.  Note that I am not using CYGWIN
> > 17.1 because it crashes quite frequently for me.     
> > Is there anyone interested in helping with this? 

Good luck :)

The work I've done so far on the Win32 port were actually geared towards
compiling in the GNU Win32 environment, and/or Visual C++ (I was compiling
it in VC++ primarily to gather errors and warnings for cleaning).