Hi milan

Stef, Does it help to just test the upgrade from 3.8 final to 3.9alpha, and is it more valuable to test from full or basic or both?

I think that for now we should really work more on putting in place mechanisms to support 3.9 (use of MC) so testing may be of limited use. But if you do it please let us know if you get problems Stef

There seem many places pointing to 3.7 as "latest", I assume a good link to 3.8 is

http://box1.squeakfoundation.org/files/3.8/

Thanks Milan

...

Stef

thanks for the report. I will look at that. There are some funny things because I got a lot of updates skipped one day and I do not know why.

Hi Stef,

OK, sorry I misunderstood I thought you ment help testing the update from 3.8 to 3.9 in your previous email.

I have tried the update before I realized it, because I wanted to try latest 3.9alpha, so FWIW, here's what hapened

I did: - download 3.8 Full from http://box1.squeakfoundation.org/files/ 3.8/ - ran the 3.8 Full image under squeak version = 3.8a-1 #1 Tue Mar 22 14:09:10 PST 2005 ccache Squeak3.8gamma of '24 November 2004' [latest update: #6548] - Squeak asked "Do you want to check for updates" - answered Yes Update -> Squeak Public Updates, 0 updates - executed in workspace: SystemVersion newVersion:'Squeak3.9alpha' - saved image as Squeak3.9alpha.image - restarted squeak with Squeak3.9alpha.image - Squeak asked "Do you want to check for updates" - answered Yes Update -> Squeak Public Updates. When loading: Error1: Loading 6537, Server Name Not Found and other errors, finally abandod. Restarted update, 9 more updates after tat, no more errors - saved - did a few very simple things with Projects, Morphs etc - simple stuff looks good

Then I wanted to use it as webserver instead of 3.7 I use now, so I: - OPened SqueakMap PackageLoader - When Loading it: error occured when updating map: UndefinedObject>>allInstances. Proceeded, but keeps repeating, finally abandoned it

So I was unable to open PackageLoader ... Perhaps a known problem anyway (I don't need to fix it in any way, just thought I'd report it as I've already kept the log)

milan

On June 9, 2005 04:34 pm, stéphane ducasse wrote:

...

Hi milan

...

Stef, Does it help to just test the upgrade from 3.8 final to 3.9alpha, and is it more valuable to test from full or basic or both?

I think that for now we should really work more on putting in place mechanisms to support 3.9 (use of MC) so testing may be of limited use. But if you do it please let us know if you get problems Stef

...

There seem many places pointing to 3.7 as "latest", I assume a good link to 3.8 is

http://box1.squeakfoundation.org/files/3.8/

Thanks Milan

...

Stef

On June 17, 2005 08:27 am, Alexandre Bergel wrote:

Hello Milan,

I do not know if someone changed the configuration of the stream, but I did an update from the Squeak Public Server using a 3.8-6548 and several changes got loaded without any error.

Hi Alexandre,

I think there is a bit of a number confusion :) - #6548 in my test and message below referred to version of the squeak executable:

Squeak3.8gamma of '24 November 2004' [latest update: #6548]

The image itself was the latest from:

http://box1.squeakfoundation.org/files/3.8/

which 2 weeks ago happened to be (probably still is)

Squeak3.8-6665.image

During loading the updates to 3.9, the error "Error1: Loading 6537, Server Name Not Found" is still there.

Just wanted to clarify,

Later Milan

Cheers, Alexandre

On Fri, Jun 10, 2005 at 02:11:11AM -0400, Milan Zimmermann wrote:

...

Hi Stef,

OK, sorry I misunderstood I thought you ment help testing the update from 3.8 to 3.9 in your previous email.

I have tried the update before I realized it, because I wanted to try latest 3.9alpha, so FWIW, here's what hapened

I did:

• download 3.8 Full from http://box1.squeakfoundation.org/files/3.8/
  • ran the 3.8 Full image under squeak version = 3.8a-1 #1 Tue Mar 22 14:09:10 PST 2005 ccache
  Squeak3.8gamma of '24 November 2004' [latest update: #6548]
• Squeak asked "Do you want to check for updates" - answered Yes Update

-> Squeak Public Updates, 0 updates - executed in workspace: SystemVersion newVersion:'Squeak3.9alpha' - saved image as Squeak3.9alpha.image - restarted squeak with Squeak3.9alpha.image

• Squeak asked "Do you want to check for updates" - answered Yes Update

-> Squeak Public Updates. When loading: Error1: Loading 6537, Server Name Not Found and other errors, finally abandod. Restarted update, 9 more updates after tat, no more errors - saved - did a few very simple things with Projects, Morphs etc - simple stuff looks good

Then I wanted to use it as webserver instead of 3.7 I use now, so I:

• OPened SqueakMap PackageLoader - When Loading it: error occured when updating map:

UndefinedObject>>allInstances. Proceeded, but keeps repeating, finally abandoned it

So I was unable to open PackageLoader ... Perhaps a known problem anyway (I don't need to fix it in any way, just thought I'd report it as I've already kept the log)

milan

On June 9, 2005 04:34 pm, stéphane ducasse wrote:

...

Hi milan

...

Stef, Does it help to just test the upgrade from 3.8 final to 3.9alpha, and is it more valuable to test from full or basic or both?

I think that for now we should really work more on putting in place mechanisms to support 3.9 (use of MC) so testing may be of limited use. But if you do it please let us know if you get problems Stef

...

There seem many places pointing to 3.7 as "latest", I assume a good link to 3.8 is

http://box1.squeakfoundation.org/files/3.8/

Thanks Milan

...

Stef

Le 2005/06/20, Don McLane dmclane@u.washington.edu écrivait :

How do I break out of an endless loop in Windows? I'm running 3.7. I've try control-c, alt-c, control-pause (break) and everything else I could think of.

Thanks, Don McLane

alt-dot or ctrl-dot dont remember for sur but it is with the 'dot'

This is a new-user question. I'm typing things in a workspace and sending them to the transcript (and, as it turns out, learning to use the debugger). I decided to challenge myself by reviewing some discrete mathematics. So one statement read:

(5 raisedTo: (r-1) modulo: r) = 1 ifFalse: [successFlag := false].

where r was big (in one of my tests it was 501659). The debugger reported that the left hand side of the '=' sign evaluated to "01"--what does the leading '0' mean? Why does the equality test fail? Am I crazy?

Thanks, Don

Everyone's done a great job on Squeak, I'm not criticizing. But this stuff about needing to 'normalize' a Large...Integer before comparing seems dangerous. It appears as though integers can be magically coerced into LargeIntegers (which is nice), but if you don't expect it, it can be quite surprising. Surprises should be minimized.

I've only be using Squeak for a couple weeks, so I'm not comfortable trying to fix this myself.

Don

Don McLane wrote:

Never mind! I discovered "normalize". Thanks.

Don

Don McLane wrote:

...

This is a new-user question. I'm typing things in a workspace and sending them to the transcript (and, as it turns out, learning to use the debugger). I decided to challenge myself by reviewing some discrete mathematics. So one statement read:

(5 raisedTo: (r-1) modulo: r) = 1 ifFalse: [successFlag := false].

where r was big (in one of my tests it was 501659). The debugger reported that the left hand side of the '=' sign evaluated to "01"--what does the leading '0' mean? Why does the equality test fail? Am I crazy?

Thanks, Don

Hi,

I inspected the result and noticed that the '01' is actually a LargePositiveInteger. Then I used the 'normalize' message and got true for your expression

|r| r := 501659. (5 raisedTo: (r-1) modulo: r) normalize = 1 =>true

It would be nice that comparison primitives do automatically send normalize message.

Noury Le 21 juin 05, à 07:03, Don McLane a écrit :

This is a new-user question. I'm typing things in a workspace and sending them to the transcript (and, as it turns out, learning to use the debugger). I decided to challenge myself by reviewing some discrete mathematics. So one statement read:

(5 raisedTo: (r-1) modulo: r) = 1 ifFalse: [successFlag := false].

where r was big (in one of my tests it was 501659). The debugger reported that the left hand side of the '=' sign evaluated to "01"--what does the leading '0' mean? Why does the equality test fail? Am I crazy?

Thanks, Don

-------------------------------------------------------------- Dr. Noury Bouraqadi - Enseignant/Chercheur Ecole des Mines de Douai - Dept. G.I.P http://csl.ensm-douai.fr/noury

European Smalltalk Users Group Board http://www.esug.org

Squeak: an Open Source Smalltalk http://www.squeak.org --------------------------------------------------------------

Hi andreas

I was visiting mike last week and I really want to use the same setup. I took the mcconfiguration package and I'm trying to understand it but time is flying. I think (in fact since ginsu and even before) that this is the way to go: decompose to control complexity. This will be also the way to manage branches, distribution because of the granularity.

...

I would like to see how we can manage gradually the image via MC so may be we will have several back and worth into the update streams. People

willing to help are welcome.

Actually, I think we've nailed this problem in Tweak by using the right mixture of packages and updates. If you haven't looked at it, here is how it works: In Tweak, *all* code is in packages[*]. Updates are exclusively used for in-image reshapes, e.g., when part of the system has undergone significant changes that require manual intervention. In order to synchronize these modifications with the package versions that we expect, we typically post a configuration map, e.g., a list of packages where we expect some specific version to be present.

When you update, we always consult the update stream first. This will suck in any intermediate configurations and perform the necessary in-image modifications. Once this is done, we merely upgrade all packages in a well-defined order to their latest version.

[*] The one exception from that rule being code that needs to modify existing Squeak code which doesn't come in packages. Since overrides are evil we leave these changes alone in the updates. What we *should* be doing here (if Squeak were in packages) is to maintain our own branches of the packages in question. As a matter of fact I started looking into this issue lately - if you check out http://source.impara.de/iSqueak.html you'll find a small change set for reorganizing Squeak 3.8 (6662) so that *all* code is in packages (look at the wiki page for reference).

Thanks for the pointer. I really want this to happen ;)

Personally, I think this is the way to go

Me too. I think that the email archive can show that :)

• we have used this for several months in Tweak and besides some

screwups that we're responsible for we definitely haven't found any fundamental flaw in this setup.

Cheers,

• Andreas

As a matter of fact I started looking into this issue lately - if you check out http://source.impara.de/iSqueak.html you'll find a small change set for reorganizing Squeak 3.8 (6662) so that *all* code is in packages (look at the wiki page for reference).

Can you point me exactly to it because I think that it would be good to build 3.9 from there. As a side effect we should be able to package in classes classes incrementally.

Stef