[Vm-dev] [squeak-dev] Looking for a new VM release candidate (was: Broken 5.3 downloads need fixing)

Marcel Taeumel marcel.taeumel at hpi.de
Wed Apr 22 08:31:29 UTC 2020

... followed by some bad news: No, 202004210425 won't do it:
https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/488 [https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/488]

Am 22.04.2020 09:09:23 schrieb Marcel Taeumel <marcel.taeumel at hpi.de>:
Hi Eliot,

thank you. That's good news. :-)

Am 22.04.2020 02:12:27 schrieb Eliot Miranda <eliot.miranda at gmail.com>:
Hi Marcel,

On Tue, Apr 21, 2020 at 6:10 AM Marcel Taeumel <marcel.taeumel at hpi.de [mailto:marcel.taeumel at hpi.de]> wrote:

Well, there are no builds for that version. Are you referring to the build 202004210425?
https://bintray.com/opensmalltalk/vm/cog/202004210425#files [https://bintray.com/opensmalltalk/vm/cog/202004210425#files]

So, that would be like bleeding edge, right? Are there any experiments going in these days? Could someone sum up what happened since mid March there?

Things are pretty stable. There was one major change to how the VM moves from machine code into the interpreter, replacing a setjmp/longjmp pair with a much lighter-weight "jump-call with substituted return address". But this "just works".  We're able to perform experments without disturbinf=g trunk.  SO for example, adding makefiles to allow building of the VM under clang-cl.exe on Windows was done without affecting the other builds at all.

Am 21.04.2020 15:04:39 schrieb Marcel Taeumel <marcel.taeumel at hpi.de [mailto:marcel.taeumel at hpi.de]>:
Hi Eliot.

> Can we please use  ... 202004192151 ...

Let's start running some tests with on this one.

Am 21.04.2020 06:44:36 schrieb Eliot Miranda <eliot.miranda at gmail.com [mailto:eliot.miranda at gmail.com]>:

On Mon, Apr 20, 2020 at 8:39 AM Marcel Taeumel <marcel.taeumel at hpi.de [mailto:marcel.taeumel at hpi.de]> wrote:

Hi Tim!

Yes, let's make a new VM release to be used in Squeak 5.3 and 6.0alpha. I think we should start this dicussion on vm-dev, so I cross-post this.

Once we find a new candidate that was build after March 24 to fix the ARM bug you mentioned, we can update the CI and the bundles. (https://github.com/OpenSmalltalk/opensmalltalk-vm/releases [https://github.com/OpenSmalltalk/opensmalltalk-vm/releases])

That new release will then be copied over to the CI's base files:
http://files.squeak.org/base/ [http://files.squeak.org/base/]

Once the new bundles are produced...
http://files.squeak.org/5.3/ [http://files.squeak.org/5.3/]

http://files.squeak.org/6.0alpha/ [http://files.squeak.org/6.0alpha/]
...we can update squeak.org [http://squeak.org].

But first things first: We need to find a new release candidate for the VM to be tested at least a week. Tim proposed the following version:


Eliot? Nicolas? Any thoughts on this?

Can we please use 

CoInterpreter VMMaker.oscog-eem.2741 uuid: 809e8b1f-fe07-4d33-bfc1-db9e5ab13a99 Apr 15 2020

StackToRegisterMappingCogit VMMaker.oscog-eem.2739 uuid: b15f5d3d-0493-4edc-a9d3-4775bb9a21e5 Apr 15 2020
VM: 202004192151 CommitHash: 6e5241e

This has this one change that I'd like to be in the VM:

    Spur: Make sure that becomeForward: answers a useful error code for read-only targets (when copyHash is true).

Am 18.04.2020 01:43:32 schrieb tim Rowledge <tim at rowledge.org [mailto:tim at rowledge.org]>:
The VM included in the 5.3 package for the ARM linux systems is fatally broken due to a code generator bug; we had various discussions about the issue in mid-March. We also noticed a problem with the bitBLT code for all platforms that was fixed by VMMaker.oscog-nice.2732

We really need to push the magic buttons to replace the ARM package in order for it to run at all; updating the other platforms to fix bitBLT would be nice. A quick test of the next available VM on bintray suggests that the https://bintray.com/opensmalltalk/vm/download_file?file_path=squeak.cog.spur_linux32ARMv6_202003291024.tar.gz [https://bintray.com/opensmalltalk/vm/download_file?file_path=squeak.cog.spur_linux32ARMv6_202003291024.tar.gz] version is functional - but maybe we want to use the latest?

What needs doing to solve this for the download page?

tim Rowledge; tim at rowledge.org [mailto:tim at rowledge.org]; http://www.rowledge.org/tim [http://www.rowledge.org/tim]
Useful random insult:- On permanent leave of absence from his senses.



best, Eliot



best, Eliot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/vm-dev/attachments/20200422/38d9ee52/attachment-0001.html>

More information about the Vm-dev mailing list