Linux VM 2.8 howto?

Stephan Rudlof sr at evolgo.de
Fri Jun 23 16:58:13 UTC 2000


Hans-Martin,

Hans-Martin Mosner wrote:
> 
> Hi,
> I'm currently trying to compile a Linux 2.8 VM, and am failing
> miserably. Since I guess others have done it, could someone please
> either point me to a precompiled binary or a detailed recipe?
> I've started out with Tim's source tree, generated everything from the
> 2.8 image (PL 2345), placed the stuff into the source tree, executed
> configure and make.
> The result is a long list of undefined references, starting out with
> "LargeIntegers_primDigitBitOr" which is referenced from sqNamedPrims.c.
> Who knows what's going on here?
> 
> Hans-Martin

Some time ago
Tim Rowledge wrote:
> 
> danielv at netvision.net.il wrote:
> >
> > I looked at the usual locations, just checking I'm not missing something
> > -
> > Is there somewhere an updated unix/linux vm with fast primitive failure?
> 
> If you look on http://sumeru.stanford.edu/tim/pooters/SqFiles/plugs you
> will find a close-to-up-to-date source tree (a couple of weeks old)
> derived from Ian's last announced release and with FilePlugin,
> SocketPlugin, SoundPlugin, LargeIntegerPlugin setup as external plugins.
> If you generate new sources, support files and plugins code you will
> find it fairly obvious where to move the file in order to compile them.
> 
> tim

He writes *Plugin setup as *external* plugins. This *could* be a hint.

Because much has changed to supply modularity I'm waiting for an actual
port, too.

I was successful in compiling the VM2.8(alpha) from Ian's port with his
sources, but not with actual from ST newly generated sources. And this
VM 2.8(alpha) hasn't worked with an *external* LargeIntegers module (I
think it was for both 2.7 and 2.8 port plugin compilations).
So I'm compiling my LargeIntegerPlugin with Ians 2.7 port and use the
VM2.7, because it doesn't work with Ian's 2.8(alpha) port onto which
Tim's seems to rely. Just overwriting in Ian's 2.8(alpha) port all
generated *.c and *.h files from an actual image has resulted in similar
problems for me (just before ECOOP).
I know, that this has disadvantages, because the VM2.7 fails slowly for
missing plugin primitives, but the 2.8 image awaits many of them, which
are not there!

I hope, that the porters coordinate their work (a common sources
repository has been introduced, but I don't know its state; Andreas Raab
has introduced gcc for Windows) well to ease future changes. Then it
should be clear at least, what platform specific stuff has to be changed
and what not.

I also think that currently it'd be better to have an actual port and a
coordinated source tree instead of a complicated and soon out of date
description of how to do that...

Porters?

I know this *actually* helps not so much,

Greetings,

Stephan

-- 
Stephan Rudlof (sr at evolgo.de)
   "Genius doesn't work on an assembly line basis.
    You can't simply say, 'Today I will be brilliant.'"
    -- Kirk, "The Ultimate Computer", stardate 4731.3





More information about the Squeak-dev mailing list