Smalltalk › Pharo › Pharo Smalltalk Developers

[update 2.0] #20011

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

8 messages Options

EstebanLM

[update 2.0] #20011

20011
-----

introducing:

-RPackage
-Spec
-Keymapping
-Refactoring Engine
-AST
-Nautilus

- Issue 5665: bug on new ecompletion/ocompletion
http://code.google.com/p/pharo/issues/detail?id=5665

- Issue 5671: Polymorph fixes
http://code.google.com/p/pharo/issues/detail?id=5671

Stéphane Ducasse

Re: [update 2.0] #20011

Esteban

are you sure that loading everything is the best move right now.
I would have started one by one. Now this is ok too just more stress.

Stef

On Apr 25, 2012, at 5:31 PM, Esteban Lorenzano wrote:

> 20011
> -----
>
> introducing:
>
> -RPackage
> -Spec
> -Keymapping
> -Refactoring Engine
> -AST
> -Nautilus
>
> - Issue 5665: bug on new ecompletion/ocompletion
> http://code.google.com/p/pharo/issues/detail?id=5665
>
> - Issue 5671: Polymorph fixes
> http://code.google.com/p/pharo/issues/detail?id=5671
>

Tudor Girba-2

Re: [update 2.0] #20011

Hi,

I am a bit puzzled by the current trend of 2.0.

Are these additions part of the image, or will they be loaded continuously on top of a seed image?

If it's the former, I thought that we discussed a while ago to keep the image to a minimum, and, in particular, to not load RB in it. Or did anything change in the meantime?

Cheers,
Doru

On 25 Apr 2012, at 19:25, Stéphane Ducasse wrote:

> Esteban
>
> are you sure that loading everything is the best move right now.
> I would have started one by one. Now this is ok too just more stress.
>
> Stef
>
> On Apr 25, 2012, at 5:31 PM, Esteban Lorenzano wrote:
>
>> 20011
>> -----
>>
>> introducing:
>>
>> -RPackage
>> -Spec
>> -Keymapping
>> -Refactoring Engine
>> -AST
>> -Nautilus
>>
>> - Issue 5665: bug on new ecompletion/ocompletion
>> http://code.google.com/p/pharo/issues/detail?id=5665
>>
>> - Issue 5671: Polymorph fixes
>> http://code.google.com/p/pharo/issues/detail?id=5671
>>
>
>

--
www.tudorgirba.com

"Problem solving efficiency grows with the abstractness level of problem understanding."

EstebanLM

Re: [update 2.0] #20011

In reply to this post by Stéphane Ducasse

Well... secure steps are overrated this days ;)

Ok, being serious: we were introducing Nautilus, and Nautilus has all those requirements... it is maybe a bit too much, but since many of them does not overlap with existing base (AFAIK, just keymapping can cause some interaction problems, if it fails), I don't think is a big issue right now. And you know, this is really early 2.0, things are going to get worse before it starts to get better, but once we stabilize, it will be not better, just great :)

Esteban

On Apr 25, 2012, at 7:25 PM, Stéphane Ducasse wrote:

EstebanLM

Re: [update 2.0] #20011

In reply to this post by Tudor Girba-2

Hi Doru,

I will quote something I hear here, and that I strongly believe:

"Pharo is not an unusable set of classes, is the tool we use everyday , so it need to have inside everything we need".

But that's not complete, to illustrate I need to quote another phrase:

"Also, Pharo should be modular and everybody should be able to start from a minimal image and build on top".

So, short answer: Yes, Pharo will have all of that inside, because we need it to evolve Pharo itself, and build a better environment.

But long answer: You also will be able to take a really small part of Pharo (let's say Pavel's kernel for 2.0 and maybe hazelnut bootstrap for 3.0) and use metacello to build your own image, the size you want and the packages you need :)

To support that, I'm already working on a metacello configuration for Pharo (Idea is to replace current release process with one based on metacello). Of course this will take time , and our objectives for 2.0 are ambitious, but previous releases of Pharo had been created enough bases to allow such a great step.

For more details on what we are trying to accomplish, I posted another mail trying to enumerate them: http://forum.world.st/About-the-new-release-process-td4574410.html

Hope this works as an explanation.

Esteban

On Apr 25, 2012, at 7:49 PM, Tudor Girba wrote:

Hi,

I am a bit puzzled by the current trend of 2.0.

Are these additions part of the image, or will they be loaded continuously on top of a seed image?

If it's the former, I thought that we discussed a while ago to keep the image to a minimum, and, in particular, to not load RB in it. Or did anything change in the meantime?

Cheers,
Doru

On 25 Apr 2012, at 19:25, Stéphane Ducasse wrote:

Esteban

are you sure that loading everything is the best move right now.
I would have started one by one. Now this is ok too just more stress.

Stef

On Apr 25, 2012, at 5:31 PM, Esteban Lorenzano wrote:

20011
-----

introducing:

-RPackage
-Spec
-Keymapping
-Refactoring Engine
-AST
-Nautilus

- Issue 5665: bug on new ecompletion/ocompletion
http://code.google.com/p/pharo/issues/detail?id=5665

- Issue 5671: Polymorph fixes
http://code.google.com/p/pharo/issues/detail?id=5671

--
www.tudorgirba.com

"Problem solving efficiency grows with the abstractness level of problem understanding."

Stéphane Ducasse

Re: [update 2.0] #20011

In reply to this post by Tudor Girba-2

> Hi,
>
> I am a bit puzzled by the current trend of 2.0.
>
> Are these additions part of the image, or will they be loaded continuously on top of a seed image?
>
> If it's the former, I thought that we discussed a while ago to keep the image to a minimum, and, in particular, to not load RB in it. Or did anything change in the meantime?
>
> Cheers,
> Doru

Doru I guess that esteban is putting pressure on himself like that. :)
You know that I want a small image and loading code. So I would have been a bit less wild
for a while but may be this is the right decision. Now let us see.
I think that it will stress a bit the system, so do not move moose to 2.0 for now.

Stef

Marcus Denker-4

Re: [update 2.0] #20011

In reply to this post by Stéphane Ducasse

On Apr 25, 2012, at 7:49 PM, Tudor Girba wrote:

> Hi,
>
> I am a bit puzzled by the current trend of 2.0.
>
> Are these additions part of the image, or will they be loaded continuously on top of a seed image?
>

Yes, loaded on top of a seed image.

For one, the Pharo image that we call "Pharo" needs to be what we use daily. Already in 1.4, most
people at RMoD use the Nautilus build, not 1.4 as is. And this is a problem, because getting the details right
is very hard if the artefact that we call Pharo is not used by anyone.

So in some way, the "Pharo" of 2.0 is more like "Pharo Full" then the old 1.3 Core. And this is good, because
the Core will shrink until it is not usable in itself. Only this way it can be really core.

Marcus

--
Marcus Denker -- http://marcusdenker.de

Sven Van Caekenberghe

Re: [update 2.0] #20011

On 26 Apr 2012, at 09:36, Marcus Denker wrote:

> Yes, loaded on top of a seed image.
>
> For one, the Pharo image that we call "Pharo" needs to be what we use daily. Already in 1.4, most
> people at RMoD use the Nautilus build, not 1.4 as is. And this is a problem, because getting the details right
> is very hard if the artefact that we call Pharo is not used by anyone.
>
> So in some way, the "Pharo" of 2.0 is more like "Pharo Full" then the old 1.3 Core. And this is good, because
> the Core will shrink until it is not usable in itself. Only this way it can be really core.

+1

(And well said)