[OpenSmalltalk/opensmalltalk-vm] Support for OpenBSD (#413)

Is it possible to build OpenSmalltalk VM for OpenBSD?

If yes, how? If no, is there any other solution available?

OpenBSD ports for 6.4 for AMD64 has only the 3.10-1 version available.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

I haven't worked with building the VM on OpenBSD, but I see this past discussions...
http://forum.world.st/Squeak-on-OpenBSD-td4945634.html
http://forum.world.st/Why-the-hell-directives-for-compiling-with-clang-were-not-taken-into-account-td5072444i20.html#a5073221

So there seem two places to review...
c1790b3...krono/openbsd
https://github.com/aryler/ostoobsd

Now it would be really great if you could create folders build.openbsd32x86 and build.openbsd64x64 to work in to merge the stuff the two urls above - to make it easier for the next person.

If files/folders need to be organised into a particularly way for OpenBAD Ports, it would be good to have a script to do that to convert from the standard opensmalltalk-vm folder arrangement.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Hey

the Cog branch should build on OpenBSD.
If not, ping me.

No need to make a dedicated openbsd folder, just use the linux one(s).

Note that the VM is not w^x compatible, nor mapstack, neither unveil.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

I wont push this point, but just to share a thought... One benefit of a dedicated folder is that it "advertises" the availability of a OpenBSD. It helps potential users "feel" more confident about trying it. There will be a percentage of people who came looking, didn't see any reference and just moved on without asking.

If you don't want separate folders for Free/Net/Open-BSD, perhaps a build.bsd64x64 folder would be appropriate?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

well, it even could be unix bla
but that is a whole different story.

Maybe have a different issue for that?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Can you do the build scripts less Bash? Or even better, can use Makefiles for that?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

I understand your point.

We've had a lot of discussing about the whole build infrastructure and as it stands, it currently works for all the major dists we can support; moreover at the moment we have to focus on other parts of the VM.

That means that the build infra (directories, bash, etc) will stay as is until we have enough manpower and consesus to find a more universal solution.

That being said, opensmalltalk-vm works and compiles on OpenBSD.
Yes it needs ports, but I think especiallt @ckeen can give some pointers on how to do things efficiently around that parts.

If anyone steps up making an upgrade to the port, all PRs are welcome.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Take a look:

output.log

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

@asarch I have re-run your build on OpenBSD 6.4 and cannot reproduce.

Btw, do you really want to build a newspeak VM? I suggest starting with a plain squeak vm, which would mean starting ins

opensmalltalk-vm/build.linux64x64/squeak.cog.spur/build

Also, make sure that you run squeak on a w^x-disabled mount, otherwise it won't start (that's what's happening, I think):

# mount
/dev/sd0a on / type ffs (local, wxallowed)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Does it need to be root for compilation?

output.log

Here is my /etc/fstab:

63a90b947a6cabb5.b none swap sw
#63a90b947a6cabb5.a / ffs rw,softdep,wxallowed 1 1
63a90b947a6cabb5.a / ffs rw,softdep 1 1
63a90b947a6cabb5.d /home ffs rw,softdep,nodev,nosuid 1 2

The mounted partitions:

$ mount
/dev/sd0a on / type ffs (local, softdep)
/dev/sd0d on /home type ffs (local, nodev, nosuid, softdep)

What's wrong?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

It is missing wxallowed as in the commented line.

OpenSmalltalk-VM needs wxallowed to run, at the moment.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Ok, after doing a dd if=/dev/zero of=/dev/rsd0c bs=1m count=1 by mistake in order to create a vnconfig device for compilation and restoring the entire system, I have:

output.log

So, If used the Squeak5.2-18229-64bit.image, Squeak5.2-18229-64bit.changes and SqueakV50.sources from http://squeak.org/ and this is what I have now:

One last question: how do you install it?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

Thank you!

Thank you very much for all guys.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

@asarch Great to see it working for you.
Sorry for your borked installation, tho :/

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

On the contrary, thank you, thank you very much for your patience and your time. Thank you.

P.S.

Just one last issue if I may, how would you compile the image from sources?:

http://files.squeak.org/sources_files/SqueakV50.sources.gz

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

On 2019-08-13, at 11:11 AM, asarch <[hidden email]> wrote:

On the contrary, thank you, thank you very much for your patience and your time. Thank you.

P.S.

Just one last issue if I may, how would you compile the image from sources?:

That's just not how it works. It *is* possible to build an image from a prescriptive definition - some folk on the Pharo list did a version last year I think, and Alejandro did a related project around '03 and I did some stuff and Craig has done a bunch of related stuff. But as a normal thing it simply doesn't mean anything. We have an image, we modify it and save it and start it, rinse and repeat. Dave Ungar pointed out many years ago that Smalltalk is saved but not born again.

Dave Ungar pointed out many years ago that Smalltalk is saved but not born again.

Nice quote.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

Norbert writes:

Pharo since version 7.0 is bootstrapped on every build. And it is surely the best way to get a
reliable artifact.

The approach you cited uses a plethora of external tools processing source code. That's composition, but not bootstrapping (something building itself, or "pulling itself up by its own bootstraps").

Whether you consider the essence of the system to reside in that source code or in machine instructions, reliability comes from being able to explain why each bit in the resulting artifact is there. I think that process is most understandable, and most effectively verified, when the space of those external tools is smallest. This is why I prefer systems that actually do bootstrap.

-C

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.