[squeak-dev] [ANN] Changing the Trunk version from 6.0alpha to 5.2alpha

+1
This 5.2 should be a freeze and polish what we have now.
I propose
No red test if possible
Open mantis show I do not could create a new category named 5.2 , who could do this ?
Also shows 94 reported crashed , some very old.

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to "5.2alpha" to denote our plans for the upcoming Squeak release. After several discussions within the board and with our release manager Edgar, we came to conclusion that it makes sense to postpone the "6.0" a little bit to finish some bigger concerns such as the use of ephemerons and Etoys clean-up.
Best,
Marcel

Marcel Taeumel

2018-05-24 11:51:48 UTC

I propose to leave "60Deprecated" as the package where we would put deprecations. No nned for 52Deprecated, I suppose...

A "feature freeze" would be "5.2beta". Do we want to go there directly, Edgar?

Best,
Marcel
Am 24.05.2018 12:34:14 schrieb Edgar De Cleene <***@gmail.com>:
+1
This 5.2 should be a freeze and polish what we have now.
I propose
No red test if possible
Open mantis show I do not could Â create a new category named 5.2 , who could do this ?
Also shows 94 reported crashed , some very old.

On 24 May 2018, at 05:44, Marcel Taeumel <***@hpi.de [mailto:***@hpi.de]> wrote:

Hi, there.

We are going to change the current version number of the Trunk to "5.2alpha" to denote our plans for the upcoming Squeak release. After several discussions within the board and with our release manager Edgar, we came to conclusion that it makes sense to postpone the "6.0" a little bit to finish some bigger concerns such as the use of ephemerons and Etoys clean-up.

Best,
Marcel

Edgar J. De Cleene

2018-05-24 12:47:59 UTC

Maybe not.
A separate Directory in ftp and a new category in Mantis should be better
for now.

Post by Marcel Taeumel
A "feature freeze" would be "5.2beta". Do we want to go there directly, Edgar?

Bernhard Pieber

2018-05-26 10:45:03 UTC

Hi,

I wanted to get a current trunk image to help testing the new release. Normally, I go to http://squeak.org/downloads/ and click the trunk link http://files.squeak.org/trunk/. This leads to an empty directory. I guess this might be an unintended consquence of the rename.

Are you going to rename the folder and files in http://files.squeak.org/6.0alpha/ as well?

Best,
Bernhard

H. Hirzel

2018-05-26 11:20:38 UTC

Hi Bernard

A solution which works well for the meantime is to take early May download from

http://files.squeak.org/6.0alpha/

And update from within the image.

One of the updates is the rename to 5.2alpha. The result is attached.

Regards
--Hannes

Post by Bernhard Pieber
Hi,
I wanted to get a current trunk image to help testing the new release.
Normally, I go to http://squeak.org/downloads/ and click the trunk link
http://files.squeak.org/trunk/. This leads to an empty directory. I guess
this might be an unintended consquence of the rename.
Are you going to rename the folder and files in
http://files.squeak.org/6.0alpha/ as well?
Best,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Fabio Niephaus

2018-05-26 14:58:50 UTC

Dear all,

TravisCI just generated the first set of 5.2alpha artifacts including
updated VMs which are now available at [1]. This URL actually links to the
same folder you can find at [2].
Although recent TravisCI builds have passed, this folder was empty because
of an issue with the updater which Marcel and I just resolved.
Therefore, all new artifacts are now uploaded to the 5.2alpha directory and
should show up at [1]/[2].

Happy testing!

Fabio

[1] http://files.squeak.org/trunk/
[2] http://files.squeak.org/5.2alpha/

Post by H. Hirzel
Hi Bernard
A solution which works well for the meantime is to take early May download from
http://files.squeak.org/6.0alpha/
And update from within the image.
One of the updates is the rename to 5.2alpha. The result is attached.
Regards
--Hannes

guess

Post by Bernhard Pieber
this might be an unintended consquence of the rename.
Are you going to rename the folder and files in
http://files.squeak.org/6.0alpha/ as well?
Best,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and

Etoys

Post by Marcel Taeumel
clean-up.
Best,
Marcel

Bernhard Pieber

2018-05-26 16:40:47 UTC

Hi,

So it was just race condition. :-)

Thank you for fixing the upload!

I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on https://bintray.com/opensmalltalk/vm/cog).

Should testing happen with this VM or with the latest from Bintray?

Cheers,
Bernhard

Post by Fabio Niephaus
Dear all,
TravisCI just generated the first set of 5.2alpha artifacts including updated VMs which are now available at [1]. This URL actually links to the same folder you can find at [2].
Although recent TravisCI builds have passed, this folder was empty because of an issue with the updater which Marcel and I just resolved.
Therefore, all new artifacts are now uploaded to the 5.2alpha directory and should show up at [1]/[2].
Happy testing!
Fabio
[1] http://files.squeak.org/trunk/
[2] http://files.squeak.org/5.2alpha/
Hi Bernard
A solution which works well for the meantime is to take early May download from
http://files.squeak.org/6.0alpha/
And update from within the image.
One of the updates is the rename to 5.2alpha. The result is attached.
Regards
--Hannes

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Fabio Niephaus

2018-05-26 21:02:20 UTC

The VMs on Bintray are removed after a few weeks to save storage costs.
More importantly, these VMs are considered bleeding edge and I don't think
it's a good idea to test a bleeding edge Squeak image with a bleeding edge
VM. If something is broken, it can be really hard to tell if the problem is
in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM
pre-release which you can find at [1]. Unfortunately, the latest stable VM
release [2] is from 2016.

Fabio

[1]
https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Fabio Niephaus
Dear all,
TravisCI just generated the first set of 5.2alpha artifacts including

updated VMs which are now available at [1]. This URL actually links to the
same folder you can find at [2].

Post by Fabio Niephaus
Although recent TravisCI builds have passed, this folder was empty

because of an issue with the updater which Marcel and I just resolved.

Post by Fabio Niephaus
Therefore, all new artifacts are now uploaded to the 5.2alpha directory

and should show up at [1]/[2].

Post by Fabio Niephaus
Happy testing!
Fabio
[1] http://files.squeak.org/trunk/
[2] http://files.squeak.org/5.2alpha/
Hi Bernard
A solution which works well for the meantime is to take early May

download from

Post by Fabio Niephaus
http://files.squeak.org/6.0alpha/
And update from within the image.
One of the updates is the rename to 5.2alpha. The result is attached.
Regards
--Hannes

Post by Bernhard Pieber
Hi,
I wanted to get a current trunk image to help testing the new release.
Normally, I go to http://squeak.org/downloads/ and click the trunk

link

Post by Bernhard Pieber
http://files.squeak.org/trunk/. This leads to an empty directory. I

guess

Post by Bernhard Pieber
this might be an unintended consquence of the rename.
Are you going to rename the folder and files in
http://files.squeak.org/6.0alpha/ as well?
Best,
Bernhard

Edgar,

Post by Marcel Taeumel
we came to conclusion that it makes sense to postpone the "6.0" a

little

Post by Marcel Taeumel
bit to finish some bigger concerns such as the use of ephemerons and

Etoys

Post by Marcel Taeumel
clean-up.
Best,
Marcel

Eliot Miranda

2018-05-27 08:06:46 UTC

Hi,

The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.

This is so depressing. What's the process by which a stable VM release is made? That process is broken. We have to fix it. 2016?!?!

Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Marcel Taeumel

2018-05-27 08:45:17 UTC

Hi Eliot,

the latest pre-release seems okay. I have been using it for several weeks now. I suppose that the "last stable" from 2016 is not any different from our recent pre-release in terms of stability. Nothing changed with regard to some stability metrics in the last years. If it builds and lasts a few weeks, it should be fine. :-) Yet, the frequent seg-faults in the CI during test runs are kind of unsettling. :-/

Best,
Marcel
Am 27.05.2018 10:07:01 schrieb Eliot Miranda <***@gmail.com>:
Hi,

On May 26, 2018, at 2:02 PM, Fabio Niephaus <***@fniephaus.com [mailto:***@fniephaus.com]> wrote:

On Sat, May 26, 2018 at 6:40 PM Bernhard Pieber <***@pieber.com [mailto:***@pieber.com]> wrote:

Hi,

So it was just race condition. :-)

Thank you for fixing the upload!

I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on https://bintray.com/opensmalltalk/vm/cog [https://bintray.com/opensmalltalk/vm/cog]).

The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.

This is so depressing. Â What's the process by which a stable VM release is made? Â That process is broken. Â We have to fix it. Â 2016?!?!

Fabio

[1]Â https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952 [https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952]
[2]Â https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest [https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest]

Â

Should testing happen with this VM or with the latest from Bintray?

Cheers,
Bernhard

Post by Fabio Niephaus
Dear all,
TravisCI just generated the first set of 5.2alpha artifacts including updated VMs which are now available at [1]. This URL actually links to the same folder you can find at [2].
Although recent TravisCI builds have passed, this folder was empty because of an issue with the updater which Marcel and I just resolved.
Therefore, all new artifacts are now uploaded to the 5.2alpha directory and should show up at [1]/[2].
Happy testing!
Fabio
[1] http://files.squeak.org/trunk/ [http://files.squeak.org/trunk/]
[2] http://files.squeak.org/5.2alpha/ [http://files.squeak.org/5.2alpha/]
Hi Bernard
A solution which works well for the meantime is to take early May download from
http://files.squeak.org/6.0alpha/ [http://files.squeak.org/6.0alpha/]
And update from within the image.
One of the updates is the rename to 5.2alpha. The result is attached.
Regards
--Hannes

Post by Bernhard Pieber
Hi,
I wanted to get a current trunk image to help testing the new release.
Normally, I go to http://squeak.org/downloads/ [http://squeak.org/downloads/] and click the trunk link
http://files.squeak.org/trunk/ [http://files.squeak.org/trunk/]. This leads to an empty directory. I guess
this might be an unintended consquence of the rename.
Are you going to rename the folder and files in
http://files.squeak.org/6.0alpha/ [http://files.squeak.org/6.0alpha/] as well?
Best,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Eliot Miranda

2018-05-28 20:03:13 UTC

Hi Marcel,

Post by Marcel Taeumel
Hi Eliot,
the latest pre-release seems okay. I have been using it for several weeks now. I suppose that the "last stable" from 2016 is not any different from our recent pre-release in terms of stability. Nothing changed with regard to some stability metrics in the last years. If it builds and lasts a few weeks, it should be fine. :-) Yet, the frequent seg-faults in the CI during test runs are kind of unsettling. :-/

There have to be major differences in circumstances we do t have reliable tests for as a number of serious vm bugs have been fixed.

What I find depressing is that we don't have a process that promotes stable VMs to a Pam e where they are easily downloaded. People are still using the 2016 VMs. This is broken.

Post by Marcel Taeumel
Best,
Marcel

The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.

This is so depressing. What's the process by which a stable VM release is made? That process is broken. We have to fix it. 2016?!?!

Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Bernhard Pieber

2018-05-27 10:42:01 UTC

Hi Eliot,

I feel your pain. My I ask, what was your process for deciding that a VM is a stable one on http://www.mirandabanda.org/files/Cog/VM/ before the switch to GitHub? Could this old process be used on GitHub as well and if not, why not?

Thanks for your terrific work on the VMs!

Best,
Bernhard

Post by Bernhard Pieber
Hi,
So it was just race condition. :-)
Thank you for fixing the upload!
I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on https://bintray.com/opensmalltalk/vm/cog).
The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.

This is so depressing. What's the process by which a stable VM release is made? That process is broken. We have to fix it. 2016?!?!

Post by Bernhard Pieber
Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,

Eliot Miranda

2018-05-28 20:07:54 UTC

Hi Bernhard,

Post by Marcel Taeumel
Hi Eliot,
I feel your pain. My I ask, what was your process for deciding that a VM is a stable one on http://www.mirandabanda.org/files/Cog/VM/ before the switch to GitHub? Could this old process be used on GitHub as well and if not, why not?

We never really got to the point of having a process as just as we were talking about adding stable links we switched over to GitHub. And we had no CI infrastructure back then.

Given that tests for bugs like the compaction snapshot issue are hard to write and bite only with large images I think something semiautomatic, where the CI proposes candidates and then we decide based on longevity and bug reports which ones to promote might work. At least we can automate the packaging and deployment of a candidate once identified.

Post by Marcel Taeumel
Thanks for your terrific work on the VMs!
Best,
Bernhard

This is so depressing. What's the process by which a stable VM release is made? That process is broken. We have to fix it. 2016?!?!

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Bernhard Pieber

2018-05-27 11:05:09 UTC

This is so depressing. What's the process by which a stable VM release is made? That process is broken. We have to fix it. 2016?!?!

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,

Fabio Niephaus

2018-05-27 12:58:48 UTC

Hi Eliot,

Post by Fabio Niephaus
Hi,

The VMs on Bintray are removed after a few weeks to save storage costs.
More importantly, these VMs are considered bleeding edge and I don't think
it's a good idea to test a bleeding edge Squeak image with a bleeding edge
VM. If something is broken, it can be really hard to tell if the problem is
in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM
pre-release which you can find at [1]. Unfortunately, the latest stable VM
release [2] is from 2016.
This is so depressing. What's the process by which a stable VM release is
made? That process is broken. We have to fix it. 2016?!?!

I recently tried to do a release because of this, but got some pushback.
Instead, I did a pre-release, so there's at least a checkpoint we can use
for Squeak. We can easily promote that pre-release to be a proper release,
we can even do a new full release today or in a few days. But I don't want
to make that decision on my own, especially because I'm not update to date
on everything that's been worked on.

Fabio

Post by Fabio Niephaus
Fabio
[1]
https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Fabio Niephaus
Dear all,
TravisCI just generated the first set of 5.2alpha artifacts including

updated VMs which are now available at [1]. This URL actually links to the
same folder you can find at [2].

Post by Fabio Niephaus
Although recent TravisCI builds have passed, this folder was empty

because of an issue with the updater which Marcel and I just resolved.

Post by Fabio Niephaus
Therefore, all new artifacts are now uploaded to the 5.2alpha directory

and should show up at [1]/[2].

Post by Fabio Niephaus
Happy testing!
Fabio
[1] http://files.squeak.org/trunk/
[2] http://files.squeak.org/5.2alpha/
Hi Bernard
A solution which works well for the meantime is to take early May

download from

Post by Fabio Niephaus
http://files.squeak.org/6.0alpha/
And update from within the image.
One of the updates is the rename to 5.2alpha. The result is attached.
Regards
--Hannes

Post by Bernhard Pieber
Hi,
I wanted to get a current trunk image to help testing the new release.
Normally, I go to http://squeak.org/downloads/ and click the trunk

link

Post by Bernhard Pieber
http://files.squeak.org/trunk/. This leads to an empty directory. I

guess

Post by Bernhard Pieber
this might be an unintended consquence of the rename.
Are you going to rename the folder and files in
http://files.squeak.org/6.0alpha/ as well?
Best,
Bernhard

Edgar,

Post by Marcel Taeumel
we came to conclusion that it makes sense to postpone the "6.0" a

little

Post by Marcel Taeumel
bit to finish some bigger concerns such as the use of ephemerons and

Etoys

Post by Marcel Taeumel
clean-up.
Best,
Marcel

Bernhard Pieber

2018-05-27 10:57:13 UTC

Hi Fabio!

Thanks for the information. I will test using the pre-release version.

What is strange then is that far older versions than 201804030952 are still available on Bintray.

It would be good to have this descriptions on Bintray maybe with a link to the GitHub releases page. On the main GitHub page there are links to the stable release and to the bleeding edge versions. Maybe a link to the latest pre-release should be added?

Bernhard

Post by Bernhard Pieber
Hi,
I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on https://bintray.com/opensmalltalk/vm/cog).

The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.
Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?

Fabio Niephaus

2018-05-27 12:52:30 UTC

Post by Bernhard Pieber
Hi Fabio!
Thanks for the information. I will test using the pre-release version.
What is strange then is that far older versions than 201804030952 are
still available on Bintray.
It would be good to have this descriptions on Bintray maybe with a link to
the GitHub releases page. On the main GitHub page there are links to the
stable release and to the bleeding edge versions. Maybe a link to the
latest pre-release should be added?

Yes, we should extend the docs. Up until recently, I've always seen the
Bintray artifacts as something vm-dev people need, but it turns out that
there are some users downloading these artifacts as well.

We could promote the pre-release to be a proper release, then the link
would automatically point to it. Instead of manually adding a link to
pre-releases, I'd much rather like us to fix the release process.

Fabio

Post by Bernhard Pieber
Bernhard

Post by Bernhard Pieber
Hi,
I saw that the 5.2alpha build contains the VM with the version

201804030952, which is not the latest version on Bintray (and interestingly
is not even in the versions list on
https://bintray.com/opensmalltalk/vm/cog).

Post by Fabio Niephaus
The VMs on Bintray are removed after a few weeks to save storage costs.

More importantly, these VMs are considered bleeding edge and I don't think
it's a good idea to test a bleeding edge Squeak image with a bleeding edge
VM. If something is broken, it can be really hard to tell if the problem is
in the VM or in the image.

Post by Fabio Niephaus
The VMs that ship with the current 5.2alpha builds are from the latest

VM pre-release which you can find at [1]. Unfortunately, the latest stable
VM release [2] is from 2016.

Post by Fabio Niephaus
Fabio
[1]

https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952

Post by Fabio Niephaus
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

karl ramberg

2018-05-29 18:42:14 UTC

Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

Best,
Karl

Post by Bernhard Pieber
Hi Fabio!
Thanks for the information. I will test using the pre-release version.
What is strange then is that far older versions than 201804030952 are
still available on Bintray.
It would be good to have this descriptions on Bintray maybe with a link
to the GitHub releases page. On the main GitHub page there are links to the
stable release and to the bleeding edge versions. Maybe a link to the
latest pre-release should be added?

Yes, we should extend the docs. Up until recently, I've always seen the
Bintray artifacts as something vm-dev people need, but it turns out that
there are some users downloading these artifacts as well.
We could promote the pre-release to be a proper release, then the link
would automatically point to it. Instead of manually adding a link to
pre-releases, I'd much rather like us to fix the release process.
Fabio

Post by Bernhard Pieber
Bernhard

Post by Bernhard Pieber
Hi,
I saw that the 5.2alpha build contains the VM with the version

201804030952, which is not the latest version on Bintray (and interestingly
is not even in the versions list on https://bintray.com/
opensmalltalk/vm/cog).

Post by Fabio Niephaus
The VMs on Bintray are removed after a few weeks to save storage costs.

Post by Fabio Niephaus
The VMs that ship with the current 5.2alpha builds are from the latest

VM pre-release which you can find at [1]. Unfortunately, the latest stable
VM release [2] is from 2016.

Post by Fabio Niephaus
Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-

vm/releases/tag/201804030952

Post by Fabio Niephaus
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Levente Uzonyi

2018-05-29 20:57:07 UTC

Is it? The Downloads page[1] has a Virtual Machines section with all the
links you need. I think it would help if there were some information given
about what VM you need to download for your platform, but if you know
that, then those links should be sufficient.

Levente

[1] https://squeak.org/downloads/

Making it easier to find and download new VM builds should be a priority.It's quite hard to find a recent VM following links from squeak.org
Best,
Karl
Hi Fabio!
Thanks for the information. I will test using the pre-release version.
What is strange then is that far older versions than 201804030952 are still available on Bintray.
It would be good to have this descriptions on Bintray maybe with a link to the GitHub releases page. On the main GitHub page there are links to the stable release and to the bleeding edge
versions. Maybe a link to the latest pre-release should be added?
Yes, we should extend the docs. Up until recently, I've always seen the Bintray artifacts as something vm-dev people need, but it turns out that there are some users downloading these artifacts as well.
We could promote the pre-release to be a proper release, then the link would automatically point to it. Instead of manually adding a link to pre-releases, I'd much rather like us to fix the release
process.
Fabio
Bernhard

https://bintray.com/opensmalltalk/vm/cog).

The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge

Squeak image with a bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.

The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.
Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest

Post by Bernhard Pieber
Should testing happen with this VM or with the latest from Bintray?
Cheers,

Edgar De Cleene

2018-05-29 23:40:20 UTC

Sent from my iPhone

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org
Best,
Karl

This scared beginners
All in one should have the most stable and recent

Edgar

Tobias Pape

2018-05-30 05:22:28 UTC

Post by Edgar De Cleene
Sent from my iPhone

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org
Best,
Karl

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start building platform packages

David T. Lewis

2018-05-30 12:54:36 UTC

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start building platform packages long ago.

This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.

My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads

H. Hirzel

2018-05-30 13:30:12 UTC

Chris Muller
Tue, Feb 28, 2017 at 3:14 AM
AIO provides a compact example file of everything needed to deploy an application on each of the top platforms. For no more than its instructional value, it is something worth keeping, IMO.

The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.

It does not need to be built regularily.
Just for the release is fine.

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dav

Tobias Pape

2018-05-30 14:59:20 UTC

The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.

But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before starting it the first time, else things just do not work. The only viable Way I see for that is building read-only disk images, so that people must move the app.

However, nobody else can use DMGs, so we have two things already:
- ZIP
- DMG

And now that snaps (https://snapcraft.io/) enter the stage for linux, we have at least to _consider_ that, too.

And when signing comes into play, this is getting too complex for me.

I completely the understand the desirability of a portable app, but given we;re not a near-stateless browser and given our "workforce", I don't see this happen reliably in the forthcoming time.

Best regards
-Tobias

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release

Bert Freudenberg

2018-05-31 04:45:37 UTC

Chris Muller
Tue, Feb 28, 2017 at 3:14 AM
AIO provides a compact example file of everything needed to deploy an

application on each of the top platforms. For no more than its
instructional value, it is something worth keeping, IMO.

Post by H. Hirzel
The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.

But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before
starting it the first time, else things just do not work. The only viable
Way I see for that is building read-only disk images, so that people must
move the app.
- ZIP
- DMG
And now that snaps (https://snapcraft.io/) enter the stage for linux, we
have at least to _consider_ that, too.
And when signing comes into play, this is getting too complex for me.
I completely the understand the desirability of a portable app, but given
we;re not a near-stateless browser and given our "workforce", I don't see
this happen reliably in the forthcoming time.
Best regards
-Tobias

âWe don't necessarily need the All-in-One, meaning one download that works
on all major platforms. It would still be cool if we can pull it off,
though. E.g. users in schools used to love Etoys-to-Go installed on a USB
stick, which would work the same plugged into a Mac at school or a PC at
home.

âWe do need a VM+image bundle that works "out of the box" though. One
download, double-click, have a running image. Could be separate downloads
per platform, if that is more convenient.

âFor macOS that means it should be an App bundle that includes the image,
and the whole bundle needs to be read-only. On first run it should copy
image and changes to the writeable sandbox folder and work from there.â
This should satisfy Apple's security requirements going forward.

â- Bert -

â

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get

some

more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dave

John Pfersich

2018-05-31 07:19:57 UTC

I think the AIO causes more problems than itâs worth. Does it work in Windows better than it does in Mac OS and Linux? I hope so, because it only works for rank beginners on the latter platforms.

/ââââââââââââââââââââ/
Encrypted email at ***@protonmail.com
Web: www.objectnets.net and www.objectnets.org

Post by Tobias Pape
HI

The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.

But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before starting it the first time, else things just do not work. The only viable Way I see for that is building read-only disk images, so that people must move the app.
- ZIP
- DMG
And now that snaps (https://snapcraft.io/) enter the stage for linux, we have at least to _consider_ that, too.
And when signing comes into play, this is getting too complex for me.
I completely the understand the desirability of a portable app, but given we;re not a near-stateless browser and given our "workforce", I don't see this happen reliably in the forthcoming time.
Best regards
-Tobias

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dave

H. Hirzel

2018-05-31 07:24:08 UTC

I think the AIO causes more problems than it’s worth. Does it work in
Windows better than it does in Mac OS and Linux? I hope so, because it only
works for rank beginners on the latter platforms.

This is the question of the target user group.

As Bert mentions Etoys-to-go on a pen drive (use your environment in
school, e.g. Mac and at home Windows-PC) _is_ an issue. But as I just
write in the previous mail probably it has to be put on the
back-burner at the moment ....

The user group of "beginners" is also the one which is the largest
group of users!

/————————————————————/
Web: www.objectnets.net and www.objectnets.org

Post by Tobias Pape
HI

Chris Muller
Tue, Feb 28, 2017 at 3:14 AM
AIO provides a compact example file of everything needed to deploy an
application on each of the top platforms. For no more than its
instructional value, it is something worth keeping, IMO.

The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.

But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before
starting it the first time, else things just do not work. The only viable
Way I see for that is building read-only disk images, so that people must
move the app.
- ZIP
- DMG
And now that snaps (https://snapcraft.io/) enter the stage for linux, we
have at least to _consider_ that, too.
And when signing comes into play, this is getting too complex for me.
I completely the understand the desirability of a portable app, but given
we;re not a near-stateless browser and given our "workforce", I don't see
this happen reliably in the forthcoming time.
Best regards
-Tobias

Post by karl ramberg
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dave

Tim Johnson

2018-05-31 05:32:30 UTC

Hi David,

Post by David T. Lewis
This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.

Thanks for the request for comments. :)

I think the AIO is neat on its own. I sometimes use the AIO and/or app bundle for one-off experiments which require a quick "download-n-go" experience where I expect to throw away any changes.

But for Mac users, I'm afraid the AIO and especially the .app bundle take too prominent a place on the website compared to the more traditional VM & sources + image & changes combination, which takes a few extra clicks to discover. I say this because there is AFAIK hardly any in-image or in-VM support for the "work style" imposed by the AIO / .app bundle on OS X / macOS.

For example: new users working with the AIO / app bundle may try "save as..." or "save as new version..." from the world menu and be surprised when any changes they thought they'd saved are gone the next time they run the AIO. This is because, AFAIK, the AIO / .app bundle work style has no in-image support (or in-VM support) for loading any image other than the one bundled within an AIO / app bundle. IIRC there have been discussions on the mailing list about adding this support (say, making the .app bundle load the highest successive .1 .2 .3 image inside the bundle, as created by "save as new version...") but I don't know if any changes ever were implemented. Forgive me if they have, and I missed it!

When I really set an OS X system up for Squeak, I create a folder called "Squeak" in the /Applications folder. Inside that folder, I put VMs and .sources files. (The VM, when used in this way, already contains the facilities to pop up a file requester asking which image to load, which the AIO/.app wouldn't do. It usually even remembers the last-used directory.) I'll usually throw my current preferred VM onto the Dock. My image and changes files live elsewhere, either on my Desktop or in my Documents folder. To load an image, I either double-click, right-click and "open with," or drag its icon onto a VM icon on my Dock.

If I had my druthers, the default OS X / macOS download on squeak.org would no longer be the .app bundle, since this is too close to the AIO experience and shares its drawbacks. It would instead be an archive or dmg with some combination of VM, sources, image, and changes files. It could potentially even contain an installer .pkg which sets up the environment similar to what I described above.

Post by David T. Lewis
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.

Hear hear.

best,
{ 'tim'. 'tim'. 'tim'. 'tim' } anyOne

H. Hirzel

2018-05-31 06:46:07 UTC

Hello

I think a compromise could be to

a) go for a smooth Mac installation experience at the moment; but
still maintain on the to do list to later on go for an AIO experience
again [1] (Mac/Windows/Linux)

b) Thus remove the inclusion of Mac support from the All-In-One at the
moment. The All-In-One would only support Linux and Windows as it is
currently.

c) Focus on getting loading legacy image segments to work fully[2].
Then there is less pressure on having different VMs around in order to
open *.morph and *.pr files done in earlier versions. [3]

d) Work on updating the compile and installation instructions for
doing different VMs for different variants of Linux. Tobias Pape
mentions limited resources. In terms of priority it is not so much the
availability of an actual Linux VM which counts but to have a good
process descriptions which allow other developers to produce Linux
VMs. [4][5]

--Hannes

[1] Bert Freudenberg writes in this thread: 'E.g. users in schools
used to love Etoys-to-Go installed on a USB stick, which would work
the same plugged into a Mac at school or a PC at home.'

[2] As of now it is possible to load a limited set of 6502 interpreter
*.pr and *.morph files.
http://wiki.squeak.org/squeak/6502
Work is all done in Smalltalk - no longer in C on the VM level.
See class LegacyImageSegment and also
http://forum.world.st/How-does-ImageSegmentLoader-readObject-work-tp5077513.html

This allows people to contribute to finalize this without any other
tool support than just the regular Squeak enviroment.

[3] It is probably mainly the loading of MethodProperties as Eliot
Miranda writes

http://forum.world.st/How-does-ImageSegmentLoader-readObject-work-tp5077513p5077578.html

<citation>
3.10 had an instance of MethodProperties as the penultimate literal in
all methods. As of Cog (but this has nothing to do with the VM)
methods either have a selector in the penultimate literal, or an
instance of AdditionalMethodState (if the method has a pragma or
properties), hence saving considerable space.
</citation>

[4] Updated compiling and installation instructions for VMs on Linux.
This includes people reading the OpenSmalltalk lists (Cuis, Pharo,
Squeak) who do not have in-depth knowledge about VMs but can easily
follow well-written instructions.

This as well has to include more in-depth explanations about the
different types on VMs

- Virtual machine http://wiki.squeak.org/squeak/676
- Updated README file on
https://github.com/OpenSmalltalk/opensmalltalk-vm pointing to compile
and installation instructions

[5] See the thread on this list from 14th May 2018
'[squeak-dev] Squeak 4.6 on 32-Bits FreeBSD 11.1'
Questions about compiling a VM for FreeBSD. Then knowledge gained
there should be put in a place where it can be later on found easily.
(Help file - list of links to instructions?, Wiki on OpenSmalltalk?)

Post by Tim Johnson
Hi David,

Thanks for the request for comments. :)
I think the AIO is neat on its own. I sometimes use the AIO and/or app
bundle for one-off experiments which require a quick "download-n-go"
experience where I expect to throw away any changes.
But for Mac users, I'm afraid the AIO and especially the .app bundle take
too prominent a place on the website compared to the more traditional VM &
sources + image & changes combination, which takes a few extra clicks to
discover. I say this because there is AFAIK hardly any in-image or in-VM
support for the "work style" imposed by the AIO / .app bundle on OS X /
macOS.
For example: new users working with the AIO / app bundle may try "save
as..." or "save as new version..." from the world menu and be surprised when
any changes they thought they'd saved are gone the next time they run the
AIO. This is because, AFAIK, the AIO / .app bundle work style has no
in-image support (or in-VM support) for loading any image other than the one
bundled within an AIO / app bundle. IIRC there have been discussions on the
mailing list about adding this support (say, making the .app bundle load the
highest successive .1 .2 .3 image inside the bundle, as created by "save as
new version...") but I don't know if any changes ever were implemented.
Forgive me if they have, and I missed it!
When I really set an OS X system up for Squeak, I create a folder called
"Squeak" in the /Applications folder. Inside that folder, I put VMs and
.sources files. (The VM, when used in this way, already contains the
facilities to pop up a file requester asking which image to load, which the
AIO/.app wouldn't do. It usually even remembers the last-used directory.)
I'll usually throw my current preferred VM onto the Dock. My image and
changes files live elsewhere, either on my Desktop or in my Documents
folder. To load an image, I either double-click, right-click and "open
with," or drag its icon onto a VM icon on my Dock.
If I had my druthers, the default OS X / macOS download on squeak.org would
no longer be the .app bundle, since this is too close to the AIO experience
and shares its drawbacks. It would instead be an archive or dmg with some
combination of VM, sources, image, and changes files. It could potentially
even contain an installer .pkg which sets up the environment similar to what
I described above.

Hear hear.
best,
{ 'tim'.

Fabio Niephaus

2018-05-31 11:51:45 UTC

Post by Tim Johnson
Hi David,
This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
Thanks for the request for comments. :)
I think the AIO is neat on its own. I sometimes use the AIO and/or app
bundle for one-off experiments which require a quick "download-n-go"
experience where I expect to throw away any changes.
But for Mac users, I'm afraid the AIO and especially the .app bundle take
too prominent a place on the website compared to the more traditional VM &
sources + image & changes combination, which takes a few extra clicks to
discover. I say this because there is AFAIK hardly any in-image or in-VM
support for the "work style" imposed by the AIO / .app bundle on OS X /
macOS.
For example: new users working with the AIO / app bundle may try "save
as..." or "save as new version..." from the world menu and be surprised
when any changes they thought they'd saved are gone the next time they run
the AIO. This is because, AFAIK, the AIO / .app bundle work style has no
in-image support (or in-VM support) for loading any image other than the
one bundled within an AIO / app bundle. IIRC there have been discussions
on the mailing list about adding this support (say, making the .app bundle
load the highest successive .1 .2 .3 image inside the bundle, as created by
"save as new version...") but I don't know if any changes ever were
implemented. Forgive me if they have, and I missed it!

We probably need a list of images that are available in an AIO bundle and a
manager that allows users to remove certain images. Another way of fixing
this would be to remove the "Save as" option in AIO bundles all together.
Or we could replace it with an "Export as" option that saves the
image/changes outside the AIO to avoid any confusion, but then users might
not know what to do with those files.

Fabio

Post by Tim Johnson
When I really set an OS X system up for Squeak, I create a folder called
"Squeak" in the /Applications folder. Inside that folder, I put VMs and
.sources files. (The VM, when used in this way, already contains the
facilities to pop up a file requester asking which image to load, which the
AIO/.app wouldn't do. It usually even remembers the last-used directory.)
I'll usually throw my current preferred VM onto the Dock. My image and
changes files live elsewhere, either on my Desktop or in my Documents
folder. To load an image, I either double-click, right-click and "open
with," or drag its icon onto a VM icon on my Dock.
If I had my druthers, the default OS X / macOS download on squeak.org
would no longer be the .app bundle, since this is too close to the AIO
experience and shares its drawbacks. It would instead be an archive or dmg
with some combination of VM, sources, image, and changes files. It could
potentially even contain an installer .pkg which sets up the environment
similar to what I described above.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Hear hear.
best,
{ 'tim'. 'tim'. 'tim'. 'tim' } anyOne

Levente Uzonyi

2018-05-27 18:16:40 UTC

Unfortunately, the pre-release VM has the bug described in issue #260:
https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .

Levente

Post by Bernhard Pieber
Hi,
So it was just race condition. :-)
Thank you for fixing the upload!
I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on
https://bintray.com/opensmalltalk/vm/cog).
The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a
bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.
Fabio
[1]Â https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2]Â https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest
Â
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

Tobias Pape

2018-05-27 18:36:36 UTC

Unfortunately, the pre-release VM has the bug described in issue #260: https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per se…

-t

Levente

Post by Bernhard Pieber
Hi,
So it was just race condition. :-)
Thank you for fixing the upload!
I saw that the 5.2alpha build contains the VM with the version 201804030952, which is not the latest version on Bintray (and interestingly is not even in the versions list on
https://bintray.com/opensmalltalk/vm/cog).
The VMs on Bintray are removed after a few weeks to save storage costs. More importantly, these VMs are considered bleeding edge and I don't think it's a good idea to test a bleeding edge Squeak image with a
bleeding edge VM. If something is broken, it can be really hard to tell if the problem is in the VM or in the image.
The VMs that ship with the current 5.2alpha builds are from the latest VM pre-release which you can find at [1]. Unfortunately, the latest stable VM release [2] is from 2016.
Fabio
[1] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/tag/201804030952
[2] https://github.com/OpenSmalltalk/opensmalltalk-vm/releases/latest
Should testing happen with this VM or with the latest from Bintray?
Cheers,
Bernhard

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.

Chris Muller

2018-05-27 22:50:30 UTC

Wow, I'm so glad we're having this discussion!

For clarification, the current stable VM is 5.0-201608171728. And,
it IS stable. It's one of the last VM's before Eliot made the new
garbage collector, so while it can encounter very long GC pauses, I
have rarely encountered any crashes whatsoever.

I've tried upgrading my VM several times over the last 18 months since
then but it would always frequent-enough crashes that it would become
too disruptive to project work. With a new VM coming out every day,
at first I would wonder, "did that crash get fixed?" and I would try
again every other week or so. But, it's time consuming to switch
VM's, bear the instability, and then swtich back, so after too many
failures I eventually stopped and just stayed with 5.0-201608171728.

Then, with the announcement of the 5.2 release a few weeks ago, I
decided to try again, so ventured back out to bintray try the "latest"
VM. In this case, it was, 201804162229. (See, we should not refer to
"latest" in any context of searching for a stable release, but only
specific version #'s please)...

Unfortunately, it didn't last very long before crashing with a
WebClient SSL access, so I was once again forced to revert back to
5.0-201608171728.

It would be so nice if we could include a newer VM one that is just as
stable for the release! But...

Unfortunately, the pre-release VM has the bug described in issue #260: https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .>

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per se…

... IMO, it's a non-starter to include a newer VM in 5.2 release if it
means this plugin will be broken.

Thanks, Eliot, for Cog and Spur, they are masterpieces of mechanation!

Best Regar

Tobias Pape

2018-05-28 07:26:04 UTC

Unfortunately, the pre-release VM has the bug described in issue #260: https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .>

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per se…

... IMO, it's a non-starter to include a newer VM in 5.2 release if it
means this plugin will be broken.

No, It means that the plugin (apparently, I have not reproduced)
- cannot be built on ubuntu 14.04
- cannot be use on 14.04 with the plugin from travis.

Although being an LTS, 14.04 is four years old (and goes out of maitenance in a year…)

its a special config, debians around that time still used 0.9-whatever.
I think its a fix that takes around 2 hours, +/- time for settiung up the OS.
I don't have those 2 hours at

Levente Uzonyi

2018-05-30 17:50:58 UTC

Unfortunately, the pre-release VM has the bug described in issue #260: https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .>

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per seâŠ

... IMO, it's a non-starter to include a newer VM in 5.2 release if it
means this plugin will be broken.

No, It means that the plugin (apparently, I have not reproduced)
- cannot be built on ubuntu 14.04
- cannot be use on 14.04 with the plugin from travis.

The same thing happens on Debian 8.

Levente

Post by Tobias Pape
Although being an LTS, 14.04 is four years old (and goes out of maitenance in a yearâŠ)
its a special config, debians around that time still used 0.9-whatever.
I think its a fix that takes around 2 hours, +/- time for settiung up the OS.
I don't have those 2 hours atm.
Most users will be fineâŠ
-t

Tobias Pape

2018-07-24 12:14:42 UTC

Post by Levente Uzonyi

Unfortunately, the pre-release VM has the bug described in issue #260: https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .>

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per se…

... IMO, it's a non-starter to include a newer VM in 5.2 release if it
means this plugin will be broken.

No, It means that the plugin (apparently, I have not reproduced) - cannot be built on ubuntu 14.04 - cannot be use on 14.04 with the plugin from travis.

The same thing happens on Debian 8.

should work

Post by Levente Uzonyi
Levente

Although being an LTS, 14.04 is four years old (and goes out of maitenance in a year…)
its a special config, debians around that time still used 0.9-whatever.
I think its a fix that takes around 2 hours, +/- time for settiung up the OS.
I don't have those 2 hours atm.

H. Hirzel

2018-07-24 16:28:42 UTC

Should be verifiyed by a test and thus confirmed

Post by Levente Uzonyi

https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/260 .>

Yea, i haven't had the time to fix that one.
It's actually a plugin not the vm per se…

... IMO, it's a non-starter to include a newer VM in 5.2 release if it
means this plugin will be broken.

No, It means that the plugin (apparently, I have not reproduced) - cannot
be built on ubuntu 14.04 - cannot be use on 14.04 with the plugin from
travis.

The same thing happens on Debian 8.

should work

Post by Levente Uzonyi
Levente

Although being an LTS, 14.04 is four years old (and goes out of
maitenance in a year…)
its a special config, debians around that time still used 0.9-whatever.
I think its a fix that takes around 2 hours, +/- time for settiung up the OS.
I don't have those 2 hours atm.
Most users will be fine…

Edgar De Cleene

2018-05-26 22:31:29 UTC

To Fabio,Bernhard,Marcel and surely many other Great Job.
No odds on Linux this time.
Moving a link to my Mac create .image works

Post by Marcel Taeumel
Hi, there.
We are going to change the current version number of the Trunk to
"5.2alpha" to denote our plans for the upcoming Squeak release. After
several discussions within the board and with our release manager Edgar,
we came to conclusion that it makes sense to postpone the "6.0" a little
bit to finish some bigger concerns such as the use of ephemerons and Etoys
clean-up.
Best,
Marcel

John Pfersich

2018-05-31 08:18:25 UTC

Well, at least on Mac OS, the fact that Squeak doesnât work right off the bat is probably a turnoff for most users, especially since there donât seem to be any instructions on what to after you get this (I just tested this):

I mean, it's a 32 bit app (running on Mac OS 10.13.4) and you can't save the image. What good is it? It's worse than non-functional. It seems to me that a lot of people would just give up on it. I'm still running 4.6 because at least it works on Mac OS without having to screw around, trying to figure out what does work. And doing things not related to programming is what the small Squeak community obviously wants beginners to do. If Squeak was a so hard to use back in 2000, I would have abandoned it immediately.

On May 31, 2018, at 12:24 AM, "H. Hirzel" <***@gmail.com> wrote:

On 5/31/18, John Pfersich <***@mac.com> wrote:
I think the AIO causes more problems than itâs worth. Does it work in
Windows better than it does in Mac OS and Linux? I hope so, because it only
works for rank beginners on the latter platforms.

This is the question of the target user group.

As Bert mentions Etoys-to-go on a pen drive (use your environment in
school, e.g. Mac and at home Windows-PC) _is_ an issue. But as I just
write in the previous mail probably it has to be put on the
back-burner at the moment ....

The user group of "beginners" is also the one which is the largest
group of users!

/ââââââââââââââââââââ/
Encrypted email at ***@protonmail.com
Web: http://www.objectnets.net and http://www.objectnets.org

On May 30, 2018, at 07:59, Tobias Pape <***@gmx.de> wrote:

HI

On 30.05.2018, at 15:30, H. Hirzel <***@gmail.com> wrote:

Chris Muller
Tue, Feb 28, 2017 at 3:14 AM

AIO provides a compact example file of everything needed to deploy an
application on each of the top platforms. For no more than its
instructional value, it is something worth keeping, IMO.

The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.

It does not need to be built regularily.
Just for the release is fine.

But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before
starting it the first time, else things just do not work. The only viable
Way I see for that is building read-only disk images, so that people must
move the app.

However, nobody else can use DMGs, so we have two things already:
- ZIP
- DMG

And now that snaps (https://snapcraft.io/) enter the stage for linux, we
have at least to _consider_ that, too.

And when signing comes into play, this is getting too complex for me.

I completely the understand the desirability of a portable app, but given
we;re not a near-stateless browser and given our "workforce", I don't see
this happen reliably in the forthcoming time.

Best regards
-Tobias

On 5/30/18, David T. Lewis <***@mail.msen.com> wrote:
On Wed, May 30, 2018 at 07:22:28AM +0200, Tobias Pape wrote:

On 30.05.2018, at 01:40, Edgar De Cleene <***@gmail.com>
wrote:

On 29 May 2018, at 15:42, karl ramberg <***@gmail.com> wrote:

Making it easier to find and download new VM builds should be a
priority.
It's quite hard to find a recent VM following links from squeak.org

This scared beginners
All in one should have the most stable and recent

Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.

This topic deserves a new subject line, and it would be great to get
some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.

My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.

Dave

H. Hirzel

2018-05-31 08:57:02 UTC

John, I agree with you.

Before thinking about a All-In-One which might cover the Mac as well

there needs to be an stand-alone installation package for 5.2 (VM,
sources, changes, image)

for the Mac which actually runs ......

H.

Well, at least on Mac OS, the fact that Squeak doesn’t work right off the
bat is probably a turnoff for most users, especially since there don’t seem
I mean, it's a 32 bit app (running on Mac OS 10.13.4) and you can't save the
image. What good is it? It's worse than non-functional. It seems to me that
a lot of people would just give up on it. I'm still running 4.6 because at
least it works on Mac OS without having to screw around, trying to figure
out what does work. And doing things not related to programming is what the
small Squeak community obviously wants beginners to do. If Squeak was a so
hard to use back in 2000, I would have abandoned it immediately.
I think the AIO causes more problems than it’s worth. Does it work in
Windows better than it does in Mac OS and Linux? I hope so, because it only
works for rank beginners on the latter platforms.
This is the question of the target user group.
As Bert mentions Etoys-to-go on a pen drive (use your environment in
school, e.g. Mac and at home Windows-PC) _is_ an issue. But as I just
write in the previous mail probably it has to be put on the
back-burner at the moment ....
The user group of "beginners" is also the one which is the largest
group of users!
/————————————————————/
Web: http://www.objectnets.net and http://www.objectnets.org
HI
Chris Muller
Tue, Feb 28, 2017 at 3:14 AM
AIO provides a compact example file of everything needed to deploy an
application on each of the top platforms. For no more than its
instructional value, it is something worth keeping, IMO.
The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.
But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before
starting it the first time, else things just do not work. The only viable
Way I see for that is building read-only disk images, so that people must
move the app.
- ZIP
- DMG
And now that snaps (https://snapcraft.io/) enter the stage for linux, we
have at least to _consider_ that, too.
And when signing comes into play, this is getting too complex for me.
I completely the understand the desirability of a portable app, but given
we;re not a near-stateless browser and given our "workforce", I don't see
this happen reliably in the forthcoming time.
Best regards
-Tobias
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org
This scared beginners
All in one should have the most stable and recent
Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.
This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dave

Fabio Niephaus

2018-05-31 11:46:01 UTC

Hi all,

Post by H. Hirzel
John, I agree with you.
Before thinking about a All-In-One which might cover the Mac as well
there needs to be an stand-alone installation package for 5.2 (VM,
sources, changes, image)
for the Mac which actually runs ......

We recently discussed the necessity of a read-only DMG bundle for macOS
which solves this issue. And I've actually implemented that already (e.g.
[1]).
Thing is, we cannot provide a DMG for the All-In-One nor is there an easy
way around the sandbox disaster on macOS.
However, we should probably extend the read-only warning in Squeak when
running on macOS and instruct users to move the .app bundle before opening
it.

Fabio

[1]
http://files.squeak.org/5.2alpha/Squeak5.2alpha-18028-32bit/Squeak5.2alpha-18028-32bit-201804030952-macOS.dmg

Post by H. Hirzel
H.

Post by John Pfersich
Well, at least on Mac OS, the fact that Squeak doesnât work right off the
bat is probably a turnoff for most users, especially since there donât

seem

Post by John Pfersich
to be any instructions on what to after you get this (I just tested
I mean, it's a 32 bit app (running on Mac OS 10.13.4) and you can't save

the

Post by John Pfersich
image. What good is it? It's worse than non-functional. It seems to me

that

Post by John Pfersich
a lot of people would just give up on it. I'm still running 4.6 because

Post by John Pfersich
least it works on Mac OS without having to screw around, trying to figure
out what does work. And doing things not related to programming is what

the

Post by John Pfersich
small Squeak community obviously wants beginners to do. If Squeak was a

Post by John Pfersich
hard to use back in 2000, I would have abandoned it immediately.
I think the AIO causes more problems than itâs worth. Does it work in
Windows better than it does in Mac OS and Linux? I hope so, because it

only

Post by John Pfersich
works for rank beginners on the latter platforms.
This is the question of the target user group.
As Bert mentions Etoys-to-go on a pen drive (use your environment in
school, e.g. Mac and at home Windows-PC) _is_ an issue. But as I just
write in the previous mail probably it has to be put on the
back-burner at the moment ....
The user group of "beginners" is also the one which is the largest
group of users!
/ââââââââââââââââââââ/
Web: http://www.objectnets.net and http://www.objectnets.org
HI
Chris Muller
Tue, Feb 28, 2017 at 3:14 AM
AIO provides a compact example file of everything needed to deploy an
application on each of the top platforms. For no more than its
instructional value, it is something worth keeping, IMO.
The AIO _is_ the application. Good to run off a pen drive as well in a
platform independent way.
It does not need to be built regularily.
Just for the release is fine.
But the AIO is impractical.
Eg, for OSX we _must_ force users now to _move_ the .app bundle before
starting it the first time, else things just do not work. The only viable
Way I see for that is building read-only disk images, so that people must
move the app.
- ZIP
- DMG
And now that snaps (https://snapcraft.io/) enter the stage for linux, we
have at least to _consider_ that, too.
And when signing comes into play, this is getting too complex for me.
I completely the understand the desirability of a portable app, but given
we;re not a near-stateless browser and given our "workforce", I don't see
this happen reliably in the forthcoming time.
Best regards
-Tobias
Making it easier to find and download new VM builds should be a priority.
It's quite hard to find a recent VM following links from squeak.org
This scared beginners
All in one should have the most stable and recent
Let's face it: the All-in-ones are dead.
Apple makes it harder than ever and for linux we should have start
building platform packages long ago.
This topic deserves a new subject line, and it would be great to get some
more input regarding who prefers using the All-In-One distribution for
regular use, versus other approaches for organizing their image and VM.
My personal view is that the All-In-One is a valuable enhancement to
the basic image and VM downloads. I do not think that it is practical
to maintain it as the primary release artifact, but I do think that we
should provide it, as best we can, in addition to the primary image and
VM release downloads.
Dave

Paul DeBruicker

2018-06-06 20:32:06 UTC

Post by Fabio Niephaus
We recently discussed the necessity of a read-only DMG bundle for macOS
which solves this issue. And I've actually implemented that already (e.g.
[1]).
Thing is, we cannot provide a DMG for the All-In-One nor is there an easy
way around the sandbox disaster on macOS.
However, we should probably extend the read-only warning in Squeak when
running on macOS and instruct users to move the .app bundle before opening
it.
Fabio

Is homebrew (https://brew.sh/) a valid way to appease the people that, on
mac, think the all-in-one isn't sufficient for real work?

--
Sent from: http://foru

Edgar J. De Cleene

2018-05-31 09:26:08 UTC

This do ot happen if you move the .app to applications folder.
Is what you do with ANY app, don¹t¹ you ?
And the not optimized is for any not 64 bits

Well, at least on Mac OS, the fact that Squeak doesn¹t work right off the bat
is probably a turnoff for most users, especially since there don¹t seem to be

John Pfersich

2018-06-01 07:50:08 UTC

No, with Pharo, I copy the files to subdirectories of my home folder and run without the problems I have with Squeak. I have dozens of images. But then again, Smalltalk is a hobby; I havenât had an offer for Smalltalk work since 2005.

/ââââââââââââââââââââ/
Encrypted email at ***@protonmail.com
Web: www.objectnets.net and www.objectnets.org

Post by Edgar J. De Cleene
This do ot happen if you move the .app to applications folder.
Is what you do with ANY app, donâtâ you ?
And the not optimized is for any not 64 bits

David T. Lewis

2018-06-01 12:33:06 UTC

John,

What you describe is exactly how I work with Squeak. VMs are installed
system wide (on my small Linux laptop "system"). Images are in directories
in my home directory, with many separate directories for various Squeak
things (including some very old images), as well as directories for Cuis
and Pharo images. I use a /usr/local/bin/run shell script to find the
right VM based on image type, so for example if I want to run an image
called "cuis.image" in my Cuis directory, the command is:

$ run cuis

I rarely use the All-In-One, but I can see that it is very important
for some people in their daily use, and it is also very important for
classrooms and for making Squeak easily downloadable for people who
would like to just download and easily try it. So I think that the
"download and run" packages are important, even though I personally
prefer to work in the way that you describe.

As with you, Smalltalk is a hobby for me.

Dave

/????????????????????????????????????????????????????????????/
Web: www.objectnets.net and www.objectnets.org

Post by Edgar J. De Cleene
This do ot happen if you move the .app to applications folder.
Is what you do with ANY app, don???t??? you ?
And the not optimized is for any not 64 bits

Germán Arduino

2018-06-01 16:01:31 UTC

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted 20
years with Smalltalk, who barely bought me a cup or two of tea :(

David T. Lewis

2018-06-02 02:39:35 UTC

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted 20
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was going
to make you rich, then you are probably a sad and disappointed person. If
you thought that Smalltalk might be a tool to help you think clearly and
explore new ideas, then you may be a happy and satisfied person.

I don't know if anyone will ever make any money by being happy and satisfied,
but I can say that in my own experience that the learning that I have done
with Squeak has been a big part of enabling me to make a decent income in
the last 10 or 15 years of my professional life.

To put it in more concrete terms: I do systems integration work for a
manufacturing company, and this involves (among other things) some custom
software development (not Squeak, not Smalltalk). I am good at what I
do, and I get paid for it. A significant part of the "I am good at what
I do" aspect is attributable to Squeak, or more specifically the learning
that I have derived from Squeak.

So I am a happy "hobby" Smalltalker. I make zero income from Squeak, but I
can honestly tell you that a good deal of my actual professional

tim Rowledge

2018-06-02 03:32:34 UTC

[snip]
So I am a happy "hobby" Smalltalker. I make zero income from Squeak, but I
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.

Now I'm almost the exact opposite in some ways. Smalltalk isn't exactly a hobby in my life; you could almost say it *is* my life. Over the last 35+ years I'd say 90+% of my income has come from Smalltalk and probably 75% specifically from Smalltalk on ARM. Which has to be one of the weirdest specialisations ever.

And I'm inclined to say that the current AIO setup is not a very effective one. An all-in-one package is a fine idea - providing the Smalltalk system and suitable VMs for all the major OS/hardware we can run on is very sensible. I think we've got to the stage where trying to make a unified one-click-run system probably causes more confusion than anything. Perhaps if we went for a simpler directory with the image/changes/sources and appropriate separate VM (sub)directories it might be easier to handle. Plus an actually helpful README to explain things.

tim
--
tim Rowledge; ***@rowledge.org; http://www.rowledge.org/tim
Quality assurance: A way to ensure you neve

John Pfersich

2018-06-06 08:08:14 UTC

Post by tim Rowledge

[snip]
So I am a happy "hobby" Smalltalker. I make zero income from Squeak, but I
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.

Now I'm almost the exact opposite in some ways. Smalltalk isn't exactly a hobby in my life; you could almost say it *is* my life. Over the last 35+ years I'd say 90+% of my income has come from Smalltalk and probably 75% specifically from Smalltalk on ARM. Which has to be one of the weirdest specialisations ever.
And I'm inclined to say that the current AIO setup is not a very effective one. An all-in-one package is a fine idea - providing the Smalltalk system and suitable VMs for all the major OS/hardware we can run on is very sensible. I think we've got to the stage where trying to make a unified one-click-run system probably causes more confusion than anything. Perhaps if we went for a simpler directory with the image/changes/sources and appropriate separate VM (sub)directories it might be easier to handle. Plus an actually helpful README to explain things.

Here, here, there there

Post by tim Rowledge
tim
--
Quality assurance: A way to ensur

Germán Arduino

2018-06-06 11:20:13 UTC

Hi Dave:

I enjoyed a lot smalltalk and its communities but I failed thinking that I
could make a living with it. Might be my lack of talent but other people
with an infinite more capability than me failed at same way and my
stubbornness for use Smalltalk in my business only made me lost lot of time
and money.

Of course I miss the technical purity of Smalltalk and objects everywhere,
but it is not viable, in my pov, to develop massive business applications,
neither web, nor desktop and much less mobile.

Just my opinion.

Saludos / Regards,
GermÃ¡n Arduino
@garduino

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work

since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted 20
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was going
to make you rich, then you are probably a sad and disappointed person. If
you thought that Smalltalk might be a tool to help you think clearly and
explore new ideas, then you may be a happy and satisfied person.
I don't know if anyone will ever make any money by being happy and satisfied,
but I can say that in my own experience that the learning that I have done
with Squeak has been a big part of enabling me to make a decent income in
the last 10 or 15 years of my professional life.
To put it in more concrete terms: I do systems integration work for a
manufacturing company, and this involves (among other things) some custom
software development (not Squeak, not Smalltalk). I am good at what I
do, and I get paid for it. A significant part of the "I am good at what
I do" aspect is attributable to Squeak, or more specifically the learning
that I have derived from Squeak.
So I am a happy "hobby" Smalltalker. I make zero income from Squeak, but I
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.
YMMV,
Dave

Louis LaBrunda

2018-06-06 13:41:59 UTC

Hi Germán,

Just an FYI, and not to knock Squeak or Pharo you might take a look at
VA Smalltalk http://www.instantiations.com or GemStoneS https://gemtalksystems.com. May people
use them to develop business and web applications with them.

Lou

Post by GermÃ¡n Arduino
I enjoyed a lot smalltalk and its communities but I failed thinking that I
could make a living with it. Might be my lack of talent but other people
with an infinite more capability than me failed at same way and my
stubbornness for use Smalltalk in my business only made me lost lot of time
and money.
Of course I miss the technical purity of Smalltalk and objects everywhere,
but it is not viable, in my pov, to develop massive business applications,
neither web, nor desktop and much less mobile.
Just my opinion.
Saludos / Regards,
Germán Arduino
@garduino

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work

since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted 20
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was going
to make you rich, then you are probably a sad and disappointed person. If
you thought that Smalltalk might be a tool to help you think clearly and
explore new ideas, then you may be a happy and satisfied person.
I don't know if anyone will ever make any money by being happy and satisfied,
but I can say that in my own experience that the learning that I have done
with Squeak has been a big part of enabling me to make a decent income in
the last 10 or 15 years of my professional life.
To put it in more concrete terms: I do systems integration work for a
manufacturing company, and this involves (among other things) some custom
software development (not Squeak, not Smalltalk). I am good at what I
do, and I get paid for it. A significant part of the "I am good at what
I do" aspect is attributable to Squeak, or more specifically the learning
that I have derived from Squeak.
So I am a happy "hobby" Smalltalker. I make zero income from Squeak, but I
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.
YMMV,
Dave

--
Louis LaBrunda
Keystone Software Corp.
SkypeMe callto://Ph

Eliot Miranda

2018-06-06 16:05:20 UTC

Hi Edwin,

Please ask your question again using a separate thread instead of hijacking
a very important thread on a specific subject.

Thank you.

I can imagine Squeak suffering a bit on the UI-part, but would there be
much difference in supported functionalities, you think?
All applications are for personnal use, and as good looking an interface
can be, I attach more importance to the usability of my applications
anyway. Of course if one can get both....
Kind regards,
Edwin Ancaer

Hi GermÃ¡n,
Just an FYI, and not to knock Squeak or Pharo you might take a look at
VA Smalltalk http://www.instantiations.com or GemStoneS
https://gemtalksystems.com. May people
use them to develop business and web applications with them.
Lou

Post by GermÃ¡n Arduino
I enjoyed a lot smalltalk and its communities but I failed thinking that

Post by GermÃ¡n Arduino
could make a living with it. Might be my lack of talent but other people
with an infinite more capability than me failed at same way and my
stubbornness for use Smalltalk in my business only made me lost lot of

time

Post by GermÃ¡n Arduino
and money.
Of course I miss the technical purity of Smalltalk and objects

everywhere,

Post by GermÃ¡n Arduino
but it is not viable, in my pov, to develop massive business

applications,

Post by GermÃ¡n Arduino
neither web, nor desktop and much less mobile.
Just my opinion.
Saludos / Regards,
GermÃ¡n Arduino
@garduino

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work

since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted

Post by GermÃ¡n Arduino
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was

going

Post by David T. Lewis
to make you rich, then you are probably a sad and disappointed person.

Post by David T. Lewis
you thought that Smalltalk might be a tool to help you think clearly

and

Post by David T. Lewis
explore new ideas, then you may be a happy and satisfied person.
I don't know if anyone will ever make any money by being happy and satisfied,
but I can say that in my own experience that the learning that I have

done

Post by David T. Lewis
with Squeak has been a big part of enabling me to make a decent income

Post by David T. Lewis
the last 10 or 15 years of my professional life.
To put it in more concrete terms: I do systems integration work for a
manufacturing company, and this involves (among other things) some

custom

Post by David T. Lewis
software development (not Squeak, not Smalltalk). I am good at what I
do, and I get paid for it. A significant part of the "I am good at what
I do" aspect is attributable to Squeak, or more specifically the

learning

Post by David T. Lewis
that I have derived from Squeak.
So I am a happy "hobby" Smalltalker. I make zero income from Squeak,

but I

Post by David T. Lewis
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.
YMMV,
Dave

--
Louis LaBrunda
Keystone Software Corp.
SkypeMe callto://PhotonDemon

--
_,,,^..^,,,_
best, Eliot

Louis LaBrunda

2018-06-06 16:35:27 UTC

Hi Edwin,

I develop with VA, have made a few personal programs with Squeak and have no experience with
Gemstone.

If you use Seaside to develop web apps (I have done some with both VA and Squeak), I think they
will look much the same.

I think you can do some nice looking GUIs with Squeak but they will look like Squeak programs.
VA uses the underling OS to draw its GUI stuff, so on Windows, they look like most Windows
programs and on Linux, they look like Linux programs.

Depending upon what you want your application to do, there may be things in one of these
Smalltalks that makes it easier that the others.

As you know Squeak is free. VA and Gemstone are not although they may offer something free for
personal use, I'm not sure.

Lou

Hello guys,
Now I am not counting on making massive amounts of money with Squeak, but I
would like to know if someone with a good knowledge of Squeak environment,
could arrive at applications that can relatively stand the comparison
(taking into account that Squeak is open source) with applications
developed with VA Smalltalk or Gemstone.
I can imagine Squeak suffering a bit on the UI-part, but would there be
much difference in supported functionalities, you think?
All applications are for personnal use, and as good looking an interface
can be, I attach more importance to the usability of my applications
anyway. Of course if one can get both....
Kind regards,
Edwin Ancaer

Post by Louis LaBrunda
Hi Germán,
Just an FYI, and not to knock Squeak or Pharo you might take a look at
VA Smalltalk http://www.instantiations.com or GemStoneS
https://gemtalksystems.com. May people
use them to develop business and web applications with them.
Lou

time

Post by GermÃ¡n Arduino
and money.
Of course I miss the technical purity of Smalltalk and objects everywhere,
but it is not viable, in my pov, to develop massive business applications,
neither web, nor desktop and much less mobile.
Just my opinion.
Saludos / Regards,
Germán Arduino
@garduino

Smalltalk is a hobby; I haven???t had an offer for Smalltalk work

since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have wasted 20
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was

going

Post by David T. Lewis
to make you rich, then you are probably a sad and disappointed person.

Post by David T. Lewis
you thought that Smalltalk might be a tool to help you think clearly and
explore new ideas, then you may be a happy and satisfied person.
I don't know if anyone will ever make any money by being happy and satisfied,
but I can say that in my own experience that the learning that I have

done

Post by David T. Lewis
with Squeak has been a big part of enabling me to make a decent income

custom

learning

Post by David T. Lewis
that I have derived from Squeak.
So I am a happy "hobby" Smalltalker. I make zero income from Squeak,

but I

Post by David T. Lewis
can honestly tell you that a good deal of my actual professional income
is supported by my hobby.
YMMV,
Dave

--
Louis LaBrunda
Keystone Software Corp.
SkypeMe callto://PhotonDemon

--
Louis LaBrunda
Keystone Software Co

Eliot Miranda

2018-06-06 18:19:37 UTC

Hi Louis, Hi All,

can we please move these discussions to another thread? The All-in-One
thread is very important in its own right.

Post by Eliot Miranda
Hi Edwin,
I develop with VA, have made a few personal programs with Squeak and have
no experience with
Gemstone.
If you use Seaside to develop web apps (I have done some with both VA and
Squeak), I think they
will look much the same.
I think you can do some nice looking GUIs with Squeak but they will look
like Squeak programs.
VA uses the underling OS to draw its GUI stuff, so on Windows, they look like most Windows
programs and on Linux, they look like Linux programs.
Depending upon what you want your application to do, there may be things in one of these
Smalltalks that makes it easier that the others.
As you know Squeak is free. VA and Gemstone are not although they may
offer something free for
personal use, I'm not sure.
Lou

Hello guys,
Now I am not counting on making massive amounts of money with Squeak, but

would like to know if someone with a good knowledge of Squeak environment,
could arrive at applications that can relatively stand the comparison
(taking into account that Squeak is open source) with applications
developed with VA Smalltalk or Gemstone.
I can imagine Squeak suffering a bit on the UI-part, but would there be
much difference in supported functionalities, you think?
All applications are for personnal use, and as good looking an interface
can be, I attach more importance to the usability of my applications
anyway. Of course if one can get both....
Kind regards,
Edwin Ancaer

Post by GermÃ¡n Arduino
I enjoyed a lot smalltalk and its communities but I failed thinking

that I

Post by GermÃ¡n Arduino
could make a living with it. Might be my lack of talent but other

people

Post by GermÃ¡n Arduino
with an infinite more capability than me failed at same way and my
stubbornness for use Smalltalk in my business only made me lost lot of

time

Post by GermÃ¡n Arduino
and money.
Of course I miss the technical purity of Smalltalk and objects

everywhere,

Post by GermÃ¡n Arduino
but it is not viable, in my pov, to develop massive business

applications,

Post by GermÃ¡n Arduino
neither web, nor desktop and much less mobile.
Just my opinion.
Saludos / Regards,
GermÃ¡n Arduino
@garduino

Smalltalk is a hobby; I haven???t had an offer for Smalltalk

work

Post by David T. Lewis
since

2005.

Post by David T. Lewis
As with you, Smalltalk is a hobby for me.

I should have realized this before and then I would not have

wasted 20

Post by GermÃ¡n Arduino
years with Smalltalk, who barely bought me a cup or two of tea :(

You should think twice about this. If you thought that Smalltalk was

going

Post by David T. Lewis
to make you rich, then you are probably a sad and disappointed

person.

Post by David T. Lewis
you thought that Smalltalk might be a tool to help you think clearly

and

done

Post by David T. Lewis
with Squeak has been a big part of enabling me to make a decent

income

custom

Post by David T. Lewis
software development (not Squeak, not Smalltalk). I am good at what I
do, and I get paid for it. A significant part of the "I am good at

what

Post by David T. Lewis
I do" aspect is attributable to Squeak, or more specifically the

learning

Post by David T. Lewis
that I have derived from Squeak.
So I am a happy "hobby" Smalltalker. I make zero income from Squeak,

but I

Post by David T. Lewis
can honestly tell you that a good deal of my actual professional

income