Smalltalk › Squeak › Squeak - Dev

stable VM?

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

7 messages Options

Chris Muller-4

stable VM?

The production VM from the squeak.org download page is from last
August and seems to be crashing.

Does anyone know where to download the latest 64-bit cog spur linux vm?

Thanks.

Bert Freudenberg

Re: stable VM?

On Wed, Feb 1, 2017 at 7:41 PM, Chris Muller <[hidden email]> wrote:

The production VM from the squeak.org download page is from last
August and seems to be crashing.

Does anyone know where to download the latest 64-bit cog spur linux vm?

Thanks.

https://bintray.com/opensmalltalk/vm/cog/_latestVersion#files

which is linked from

https://github.com/OpenSmalltalk/opensmalltalk-vm/

(maybe we should link to that on squeak.org too)

- Bert -

Chris Muller-3

Re: stable VM?

Thanks. 5.0-201608171728 has been pretty stable, but I have a
software demo this Friday and would like to use the most recommended
VM for stability.

That link is to the daily builds. I was wary to automatically
consider the latest as the most-stable, since the VM development is
now an active collaboration of developers on github.

Is it safe to use the latest? Or, if not, is there a blessed version
since 5.0-201608171728?

On Wed, Feb 1, 2017 at 12:45 PM, Bert Freudenberg <[hidden email]> wrote:

> On Wed, Feb 1, 2017 at 7:41 PM, Chris Muller <[hidden email]> wrote:
>>
>> The production VM from the squeak.org download page is from last
>> August and seems to be crashing.
>>
>> Does anyone know where to download the latest 64-bit cog spur linux vm?
>>
>> Thanks.
>
>
> https://bintray.com/opensmalltalk/vm/cog/_latestVersion#files
>
> which is linked from
>
> https://github.com/OpenSmalltalk/opensmalltalk-vm/
>
> (maybe we should link to that on squeak.org too)
>
> - Bert -
>
>
>
>

Tobias Pape

Re: stable VM?

On 01.02.2017, at 21:25, Chris Muller <[hidden email]> wrote:

> Is it safe to use the latest? Or, if not, is there a blessed version
> since 5.0-201608171728?

nope :(

Nicolas Cellier

Re: stable VM?

Though, if you analyze the commits, they are essentially bugfixes.

You'll find not so many new features:

- some backports from Pharo VM (rather solid and/or protected by some #ifdef)

- the Lowcode from Ronie which is essentially separated from the rest of the VM

- the Sista stuff from Clement

The new garbage collector from Eliot is in a separate branch

So maybe the VM are not blessed, but I don't expect less stability than the "officially" blessed.

Anyway, do we care of "official" blessing?

What we need is more testing rather than more blessing, otherwise how could we certify anything?

Of course, it may seem cavalier to ask production guys to be the beta-tester, but you know that nothing help more...

2017-02-01 21:51 GMT+01:00 Tobias Pape <[hidden email]>:

On 01.02.2017, at 21:25, Chris Muller <[hidden email]> wrote:

> Is it safe to use the latest? Or, if not, is there a blessed version
> since 5.0-201608171728?

nope :(

Chris Muller-3

Re: stable VM?

> So maybe the VM are not blessed, but I don't expect less stability than the
> "officially" blessed.

Okay, that's what I really wanted to know. Thanks, I'll upgrade to
today's built version.

> Anyway, do we care of "official" blessing?

Like you said Eliot is in a separate branch with the new GC'er. I
assume he created a separate branch because the changes are bound to
cause breakage for a while. I wanted to know whether any similar
experiments in the main branch could impose similar risks.

> What we need is more testing rather than more blessing, otherwise how could
> we certify anything?

More testing of which version? New versions are popping out every
day, upgrading becomes something done when there's a reason to, not
because there's a new version. Within 24 hours of upgrading, it'll be
just another old version -- unless it was blessed.

Some situations need to emphasize stability, it seems like it would be
nice to have a baseline version that is known not to have have any
halts or debugging stuff or experimental stuff, and so it could be
more easily accepted into wide usage and testing in those production
situations.

>
> Of course, it may seem cavalier to ask production guys to be the
> beta-tester, but you know that nothing help more...
>
>
> 2017-02-01 21:51 GMT+01:00 Tobias Pape <[hidden email]>:
>>
>>
>> On 01.02.2017, at 21:25, Chris Muller <[hidden email]> wrote:
>>
>> > Is it safe to use the latest? Or, if not, is there a blessed version
>> > since 5.0-201608171728?
>>
>> nope :(
>>
>
>
>
>

Hannes Hirzel

Re: stable VM?

Chris,

My impression is that going for a 32-bit cog spur linux vm is safer in
your case.

--Hannes

On 2/2/17, Chris Muller <[hidden email]> wrote:

>> So maybe the VM are not blessed, but I don't expect less stability than
>> the
>> "officially" blessed.
>
> Okay, that's what I really wanted to know. Thanks, I'll upgrade to
> today's built version.
>
>> Anyway, do we care of "official" blessing?
>
> Like you said Eliot is in a separate branch with the new GC'er. I
> assume he created a separate branch because the changes are bound to
> cause breakage for a while. I wanted to know whether any similar
> experiments in the main branch could impose similar risks.
>
>> What we need is more testing rather than more blessing, otherwise how
>> could
>> we certify anything?
>
> More testing of which version? New versions are popping out every
> day, upgrading becomes something done when there's a reason to, not
> because there's a new version. Within 24 hours of upgrading, it'll be
> just another old version -- unless it was blessed.
>
> Some situations need to emphasize stability, it seems like it would be
> nice to have a baseline version that is known not to have have any
> halts or debugging stuff or experimental stuff, and so it could be
> more easily accepted into wide usage and testing in those production
> situations.
>
>
>
>
>>
>> Of course, it may seem cavalier to ask production guys to be the
>> beta-tester, but you know that nothing help more...
>>
>>
>> 2017-02-01 21:51 GMT+01:00 Tobias Pape <[hidden email]>:
>>>
>>>
>>> On 01.02.2017, at 21:25, Chris Muller <[hidden email]> wrote:
>>>
>>> > Is it safe to use the latest? Or, if not, is there a blessed version
>>> > since 5.0-201608171728?
>>>
>>> nope :(
>>>
>>
>>
>>
>>
>
>