Squeak
  links to this page:    
View this PageEdit this PageUploads to this PageHistory of this PageTop of the SwikiRecent ChangesSearch the SwikiHelp Guide
Ian Piumatra: Squeak 2.8. Unix port status report - 26 Aug 2000
Last updated at 1:34 pm UTC on 16 January 2006

Ian Piumatra consideres the Unix port of Squeak version 2.8 to be in beta state and 2.9 in alpha state


Date: Sat, 26 Aug 2000 01:18:06 +0200 (MET DST)
From: Ian Piumarta
Subject: Unix 2.8/2.9a

Folks,

If you look here:

ftp://ftp.inria.fr/INRIA/Projects/SOR/users/piumarta/squeak/

you will find the second PRE-release of 2.8 for Unix:

Squeak-2.8pre2-img.tar - Dan's 2.8 image from earlier this week
Squeak-2.8pre2-src.tar - sources

plus the usual:

Squeak-2.8pre2-alphaev56-dec-osf4.0e.tar
Squeak-2.8pre2-i686-pc-linux-gnu.tar
Squeak-2.8pre2-powerpc-unknown-linux-gnu.tar
Squeak-2.8pre2-sparc-sun-solaris2.5.1.tar
Squeak-2.8pre2-sparc-unknown-linux-gnu.tar

The above VMs are statically-linked against the FFI library (no need
to install libffi to use them).

2.8 still isn't quite ready for prime time. I've not integrated the
Netscape plugin stuff yet, and there are new problems:


Squeak3D.

called sqFooPrims.c, but now the SoundCodec plugin is called
"SoundCodecPrims.c" and its associated primitives are called
"SoundCodecPlugin.c". Similarly for SqueakFFIPrims.c, which
contains a plugin. Confused? "configure" certainly was (is).

simply omitting it breaks the build because sqNamedPrims is
expecting the primitives to be there. I have put a temporary
solution in place, but I'm not proud of it.

We should fix all of the above. Much more important than the
"secondary effect" of it causing a configuration nightmare on Unix,
it's just plain sloppy – and confusing for anybody trying to make
sense of the sources for the first time.

In spite of all that I wanted to get the new 2.8 out ASAP because I
spent some time making the socket code work again. Hopefully there
will be far fewer problems with it now, but I still consider it "beta"
quality: hammer on it hard and let me know what breaks./b (While I was
at it I implemented the UDP specific primitives and get/setOption too.
[Who's responsible for the socket get/setOption code in the image? We
need to talk.])

I also want to get feedback from people who compile on platforms to
which I don't have access: the "tricky" stuff is driven by libtool
now, and I'd like to know how that works out.

For space cadets only, the above directory also contains:

Squeak-2.9a-2524.pre1-img.tar - 2.9a with the latest updates
Squeak-2.9a.pre1-src.tar - my current sources for 2.9a

plus the traditional:

Squeak-2.9a.pre1-alphaev56-dec-osf4.0e.tar
Squeak-2.9a.pre1-i686-pc-linux-gnu.tar
Squeak-2.9a.pre1-powerpc-unknown-linux-gnu.tar
Squeak-2.9a.pre1-sparc-sun-solaris2.5.1.tar
Squeak-2.9a.pre1-sparc-unknown-linux-gnu.tar

Coming soon: proper event support, asynchronous file i/o, and a
completely re-worked socket implementation that will share all the
NBIO and ASYNC stuff with the async file prims.

No guarantees that any of the above are anything but collections of
compressed line noise: they were all built in rather a hurry.
(Remember: "It might blow up, but it won't go `pop!'" ;-)

Have fun.

Ian

PS: Does anyone have a Unix box thatdoesn't have the /usr/bin/test
program?
[JohnMcIntosh]

For fun I compiled Squeak 2.8 on a Quadra 950 running linux-m68k.
Thats
Squeak-2.8-m68k-unknown-linux2.0.38

I had to upgrade my gnu autoconf. I did not find an FFI library I could use so I deleted all references to FFI from the scripts and headers which enabled the linking to work.

It's hosting a Swiki for my website. So now we'll see how stable it is.

For more fun we'll try a OpenBSD 2.7 68K box, and an OpenBSD 2.7 Intel box later this weekend.