[squeak-dev] [ANN] Public discuss about 'rules' of 3.10 and his possible branches

To all:

At the moment I'm doing 3.10.1 , my decision is going back to old trusted
.cs policy for have updates as fast we could have.

Now , Matthew and Keith come with outstanding technology ( Congratulations !
BRAVO!) , I have #unloadTraits and a Squeak3.10-7164-WT-basic.image for
daily work , so I know if any wrong could come.

I joining to #IRC with this and I was working on my game SlidesStones.pr
also in this image.
The .pr until now could be saved and loaded into old dirty
Squeak3.9b-7051.image I have and also into FunSqueak.
All this could give me info about tons of troubles in advance.

The complete test was run as we do on Mac, Windows XP, Simply Linux 6.5.
Same results as regular release.

By now, I left Traits until Board give us his final decision.

We could have a 3.10.2 version 'regular' , and a SqueakWT.3.10.2 or Squeak
3.10.2-WT or appropriate name picked for a NO Traits version.

So we could have two branches and all could be happy or maybe not.

>From the five meeting all we doing on web and Matthew was hosting and
documenting.

We need DS or MC2

Both "teams" say in turn need SystemEditor.

My suggestion to Randall and Ken is we need rebuild 3.10 from where 3.9 was.
I need first hand experience for troubles.
It's my decision take MC2 as I meet with Colin in person and as he
understand enough Spanish seems me a sure bet.

Anybody don't trust Colin ?

We need improve the 3.10 for the record of changes , too mumbo jumbo.
Lots of people is buying excellent book of Stephanne and unloading 3.9

At some point this people realize they should upgrade.
The current way of updates (the button on the Squeak flap ) is rough, not
always could work, need more as newbie knowledge, using our actual updates.

That's is why I say we should have a complete rebuild.
I say I put 3.10.2 and 3.10.2WT on the Board table at time ESUG begins.
3.10.2 is 3.10 rebuilded and Traits
3.10.2-WT means WITHOUT TRAITS

So I ask for feedback .
And no more F1 , no jokes, strict to business.
Ustedes se lo pierden :=)

Edgar




Edgar

2008/5/15 Edgar J. De Cleene <[hidden email]>:
I think it would be better when
3.10.2-WT (without traits)
can be easily transformed to
3.10.2-WT (with traits)
by loading package or changeset.

Making traits loadable / unloadable gives us much more freedom in choice.


--
Best regards,
Igor Stasenko AKA sig.

El 5/15/08 10:34 AM, "Igor Stasenko" <[hidden email]> escribió:

> I think it would be better when
> 3.10.2-WT (without traits)
> can be easily transformed to
> 3.10.2-WT (with traits)
> by loading package or changeset.


Yes, but just now we could have a "normal" image as we have now or one
without Traits.

Is not my job doing a Traits package, no idea if this could be done , when ,
for who.

So again I ask to Board...

It's redoing all again using MC2 from 3.9 ok ?

Or NOT redoing...

It's Board decision have a 3.10.x2-basic and a 3-10.2-basic-WT ?.
For playing safe.
For people like Andreas could have a cleaner system for put Croquet into (
maybe )?

I don't know if Andreas could consider this is a good or bad idea or if he
could consider is best some variant of I saying.

I don't wish work and be burned at to stake in ESUG.
Please let me do something...

Edgar

>>>>> "Edgar" == Edgar J De Cleene <[hidden email]> writes:

Edgar> So again I ask to Board...

Edgar> It's redoing all again using MC2 from 3.9 ok ?

Edgar> Or NOT redoing...

Edgar> It's Board decision have a 3.10.x2-basic and a 3-10.2-basic-WT ?.
Edgar> For playing safe.
Edgar> For people like Andreas could have a cleaner system for put Croquet into (
Edgar> maybe )?

I don't understand this.

You and I have had some private conversations.  At the end, I said
what I thought we should do.  I was acting as "the board" when we had
those conversations.

You now replay all this in public again, undoing anything you and I have
agreed upon.

This confuses me.

Please tell me what part of our private conversations you now are ignoring,
and why.

--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<[hidden email]> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

To my understanding, a 3.10.xx images should mean following:
- an image based on 3.10 release, which contains hot-fixes of bugs,
discovered since release.

This means, that image should be STABLE, no experimental code (like
removing traits), only small fixes.

You free to go with removing traits from base 3.11 (and with
supporting option to load them back, of course), but not in any
3.10.xx images.

--
Best regards,
Igor Stasenko AKA sig.

On Thu, May 15, 2008 at 10:10:20AM -0300, Edgar J. De Cleene wrote:
There is another very good reason for redoing 3.10: all
contributions to 3.10 were made after the effective date of the
license agreements (November 2006), so, unless the authors of
the 3.10 contributions made an explicit statement about the
license of their contributions, all the 3.10 contributions are
license-unclean.

For 3.11, the board has asked the release team to not accept any
contributions without an explicit license of MIT. For 3.10, we
should retroactively enforce this policy as well.

For more information, see my license pages:
- http://installer.pbwiki.org/Squeak40
  - Top page
- http://installer.pbwiki.org/LicenseAuditing
  - A breakdown of what is and is not license-clean

--
Matthew Fulmer -- http://mtfulmer.wordpress.com/

 
> There is another very good reason for redoing 3.10: all
> contributions to 3.10 were made after the effective date of the
> license agreements (November 2006), so, unless the authors of
>  
They were and all contributors knew this, there is NO WAY that 3.10
should be considered unclean!

Keith

BTW.: etoys.image is licensed officially under Apache License, Version
2.0 now because VPRI has taken the responsibility for this step. Does
it mean that we can take any code from this image without fear that
there will be any licensing quarrel at law in future? Does it mean
that we can tell in this case that we used this code in good faith
that it's clean from the law POV and delegate it to VPRI? Is it ok for
Software Freedom Conservancy?

-- Pavel

On Thu, May 15, 2008 at 7:00 PM, Keith Hodges <[hidden email]> wrote:

Depends on how confident you are in VPRI's license cleanliness. I
don't mean any offense, to VPRI.

A more egregeous example is Heumel Smalltalk which forked a Squeak
(which only 1.1 seems clean).

Of course, you can never be too sure in different jurisdicitons.
That's why we have lawyers.

On 5/15/08, Pavel Krivanek <[hidden email]> wrote:

>>>>> "David" == David Mitchell <[hidden email]> writes:

David> Depends on how confident you are in VPRI's license cleanliness. I
David> don't mean any offense, to VPRI.

If VPRI would also promise to indemnify any derivative work by setting up a
legal defense fund, this would be one possibility.  I mean, the real purpose
of any license is to have something to help you in court.  And money goes a
long ways. :)

However, I can't imagine any shareholder that would approve such a deal,
unless there was a good business case.

That said, I haven't asked anyone at VPRI.  I would just be surprised.

--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<[hidden email]> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

Sigh, more license discussions. What fun.

I am not a lawyer, though I do get to shoot them sometimes in my other  
job.

a) the licensing agreement that people have been signing (thank you  
people!) only explicitly covers past code. It doesn't mention future  
code. See http://www.netjam.org/squeak/contributors/ and click the  
link  'get a copy of it here' to read it for yourself.

b) there are still a number of contributors that have not signed, or  
that have signed but we don't know about it yet - see http://www.netjam.org/squeak/contributors/missingSignatories 
  (and this list is certainly a few items out right now)

c) there are some sets of initials to which we can't map email  
addresses - see http://www.netjam.org/squeak/contributors/missing (and  
again, we know a few items are incorrect)

d) until we are able to trace the initials attached to every revision  
of every method in the current image and match those to signed  
licences the SFLC will not be content.

e) we need to modify all the tools/sites such as SM, SqS so that it is  
made completely clear what license is in use for *new* contributions.  
Anything that would logically go into any typical default image *must*  
be MIT licensed.

f) it seems pretty clear to me that any expanded image - such as  
FunSqueak - that people would like to see distributed by any means  
connected to the foundation would need to require clear licensing for  
every method added to the core. One of the reasons for trying to build  
a smaller core image was to reduce the volume of code needing checking  
and clearing.

g) there is still a bunch of work to do

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Don't document the program; program the document.

El 5/15/08 1:29 PM, "Matthew Fulmer" <[hidden email]> escribió:

> There is another very good reason for redoing 3.10: all
> contributions to 3.10 were made after the effective date of the
> license agreements (November 2006), so, unless the authors of
> the 3.10 contributions made an explicit statement about the
> license of their contributions, all the 3.10 contributions are
> license-unclean.

Ok. Glad we going to some place.
I say BRAVO again to your work (and Keith) for I could have #unloadTraits.

And I also suggest to "FIA Stewards" for redoing all with DS,

We could have four branches.

3.10  bussiness as usual, having Traits.
3.10.x-WT
3.10.x-DS (I want also test work of "Hodges&Fullmer Racing" , excuse my
humor.
3.10.x-MC2


Off course I can't do all, so I must made a decision.

And Board say "go" "stop" "wait until x" "we need know more , but you could
do XYZ  , so have first hand feedback"

All begin to understand ?

Edgar

  I do not think that there is a need to actually publish branches.

option 2 is available by doing.

Installer upgrade install: 'UnloadTraits'.

option 3 will be avialble by doing

Installer upgrade install: 'DeltaStreams'.

option 4 will be available by doing.

Installer upgrade wiresong project: 'MC2'; install: 'MC2'. or something
similar.

best regards

Keith

>>>>> "Edgar" == Edgar J De Cleene <[hidden email]> writes:

Edgar> We could have four branches.

Edgar> 3.10  bussiness as usual, having Traits.

Well, 3.10.x are hot patches for the stable version, as need.

3.10 is *done*
NO MORE DEVELOPMENT.

All the next 3 should be 3.11alpha branches.  3.10.x should be reserved for
hot patches to a *stable* 3.10, not for introducing unstable new features.

Why do you keep ignoring what you agreed on this?

Edgar> 3.10.x-WT
Edgar> 3.10.x-DS (I want also test work of "Hodges&Fullmer Racing" , excuse my
Edgar> humor.
Edgar> 3.10.x-MC2

--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<[hidden email]> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion