Plan for 3.10 iteration 2

I complete the unload of Etoys and Nebraska from Squeak3.10alpha.7091.image.

I wish take this chance of major change for incorporate in this
Squeak3.10alpha.7091W.image. (W for without Etoys and Nebraska) , some
things required for members (Mulltillingual and others) what at this point I
could't do using the "Monticello way" what we are using now.

Things what you wish know.

Some class of "old Etoys" are keep into image, my experiments with
SqueakLight shows what soon or later you need one.
Etoys and Nebraska was reshaped taking this and MinimalMorphic lessons.
One of 3.10 targets is for easier the transition to smaller images what all
could use in safe form.
Etoys and Nebraska Nebraska unload an load again.
Without changes his classes intialize and with our current Ralph version of
Monticello !! (This wonder me too )
The only what I need in the first ever try was:
To first complain is fileIn  Vocabulary >> changeMadeToViewerAdditions
and at the second for TileMorph initialize
"TileMorph readInArrowGraphics    -- call manually if necessary to bring
graphics forward"
this steps requires three .morphs in user directory.

Off course , all need use and test and sure many changes for all works as if
never Etoys and Nebraska go out or image.

SystemNavigation default allUnimplementedCalls size 183 grows to 352.
Cleaning take some more time.

Image without Etoys and Nebraska is .8 Mb smaller (first unpolished try).

So I required your feedback.

I should put this Squeak3.10alpha.7091W.image into Squeak ftp for all could
try and give opinions ?

Or ask to Klaus for he put in his ftp ?

Edgar

PS> If wait what I do all the methods rewrite , send note to my Apple dealer
for they lower his prices. :=)

Thank you for your work. Can you please upload your image to the ftp
so that I can base the next squeak-dev image on it?

--
Damien Cassou

El 4/22/07 6:00 AM, "Damien Cassou" <[hidden email]> escribió:

> Thank you for your work. Can you please upload your image to the ftp
> so that I can base the next squeak-dev image on it?
Yes, but I wish more feedback before.
I zipped and put on my ftp this afternoon if no feedback for you could have
and use.
For .dev image could be more adequate, as I don't think what people using
.dev care about Etoys and Nebraska.

And the only way of know what problems could have is using and polishing.

Remember what image works but still needs lots of cleaning.

What I wish is agreement for put into this image all Multillingual and enh
like http://bugs.squeak.org/view.php?id=6348 of Martin what at this point I
could't do in our "Monticello way" process, but yes could go into image via
old style .cs without problem.

This way we could have a very interesting release to work on.

Edgar

How do you reload eToys?  As you know, my goal is that as much as
possible should be taken out of the image and loaded.  So, if you can
reload it then I am in favor ot taking it out.  We will need some
tests for it, so we know it has been reloaded successfully.

Damien, it sounded to me that Edgar said that the image still needed a
lot of work.  So, yu should not base swueak-dev on it.  Instead, you
should make sure squeak-dev works on both the current 3.10 and
3.10alpha7091W

-Ralph

2007/4/22, Ralph Johnson <[hidden email]>:
> How do you reload eToys?  As you know, my goal is that as much as
> possible should be taken out of the image and loaded.  So, if you can
> reload it then I am in favor ot taking it out.  We will need some
> tests for it, so we know it has been reloaded successfully.
>
> Damien, it sounded to me that Edgar said that the image still needed a
> lot of work.  So, yu should not base swueak-dev on it.  Instead, you
> should make sure squeak-dev works on both the current 3.10 and
> 3.10alpha7091W

Squeak-dev image version 115 is based on 3.10alpha7091 (with etoys and
nebraska). I'm currently working with it and it works very well. No
problem detected.

--
Damien Cassou

On 4/22/07, Damien Cassou <[hidden email]> wrote:
> Squeak-dev image version 115 is based on 3.10alpha7091 (with etoys and
> nebraska). I'm currently working with it and it works very well. No
> problem detected.

I, too, have been using squeak-dev based on 3.10alpha7091, and it
works for me.  I was concerned about the "W" part.  I wouldn't want
you to make a version of squeak-dev that worked only in the version of
3.10 without eToys.  That isn't likely, but if it happened, people who
wanted to use eToys couldn't work with squeak-dev.

All I am saying is to make sure squeak-dev keeps working with
3.10alpha7091, at least until we have a version that can successfully
load eToys.  This shouldn't be hard because squeak-dev is now a
package in Package Universe rather than an image.  But it does require
extra testing.

-Ralph

El 4/22/07 7:47 AM, "Ralph Johnson" <[hidden email]> escribió:


I do two new Etoys and Nebraska .mcz.
I need do some trick for unloading, but unlike Pavel , I don't change
initialize on Etoys and Nebraska classes.
No Undeclared , no obsoletes.
Using yours Monticello version I could load this .mcz from my repository (I
don't put into 310 still).
Don't need of special things.
Only the two minors complaints what I said in previous mail.
The cleaning what I saying is the growing of undeclared of 183 what we have
now to 352.
This don't should raise problems for packages not using some of Etoys and
Nebraska.
I keep the classes what I learn we could need sooner or later .
This still in image classes are
#(ActorState MethodMorph Player Presenter StickySketchMorph UnscriptedPlayer
SlotInformation)
Still remains remove the Test classes of Etoys what now give 7 errors, but
is a easy step.

Maybe Damien (and others) like to have the chance of see the  kind of
problems what this image have ?
For not having obsoletes, I must recode several of Flaps (and discover one
or two things ).
Follow the same style of commenting out
example:

defaultsQuadsDefiningSuppliesFlap
    "Answer a list of quads which define the objects to appear in the
default Supplies flap.
    previously in quadsDefiningSuppliesFlap"

    ^  #(
    (RectangleMorph         authoringPrototype        'Rectangle'         'A
rectangle')
    (RectangleMorph        roundRectPrototype        'RoundRect'        'A
rectangle with rounded corners')
    (EllipseMorph            authoringPrototype        'Ellipse'
'An ellipse or circle')
    (StarMorph                authoringPrototype        'Star'            'A
star')
    (CurveMorph            authoringPrototype        'Curve'            'A
curve')
    (PolygonMorph            authoringPrototype        'Polygon'        'A
straight-sided figure with any number of sides')
    (TextMorph                boldAuthoringPrototype        'Text'
'Text that you can edit into anything you desire.')
    "(ScriptingSystem        prototypicalHolder         'Holder'
'A place for storing alternative pictures in an animation, etc.')"
    (ImageMorph            authoringPrototype        'Picture'        'A
non-editable picture of something')
    "(ScriptableButton        authoringPrototype        'Button'
'A Scriptable button')"
    (SimpleSliderMorph        authoringPrototype        'Slider'
'A slider for showing and setting numeric values.')
    (PasteUpMorph            authoringPrototype        'Playfield'        'A
place for assembling parts or for staging animations')
    (BookMorph                authoringPrototype        'Book'            'A
multi-paged structure')
    (TabbedPalette            authoringPrototype        'TabbedPalette'
'A structure with tabs')
    (JoystickMorph            authoringPrototype        'Joystick'        'A
joystick-like control')
    (ClockMorph                authoringPrototype        'Clock'
'A simple digital clock')
    (BookMorph                previousPageButton         'PreviousPage'
'A button that takes you to the previous page')
    (BookMorph                nextPageButton            'NextPage'        'A
button that takes you to the next page')
        ) asOrderedCollection

Edgar, in case it is not clear, I think you definitely should publish
what you have done.  I was just warning Damien not to force people to
depend on it until you are done.  But people need to use it to give
feedback to you.

-Ralph

El 4/22/07 8:20 AM, "Ralph Johnson" <[hidden email]> escribió:

People could load Etoys.
What don't means all works .
Etoys on 3.9 have problems, and OLPC is 3.8 based.
I polish problems with time.
And I asking OLPC people for doing a OLPC to 3.10 , 3.10 to OLPC
"intelligent bridge" or iBridge.

Edgar

El 4/22/07 8:34 AM, "Ralph Johnson" <[hidden email]> escribió:

> Edgar, in case it is not clear, I think you definitely should publish
> what you have done.  I was just warning Damien not to force people to
> depend on it until you are done.  But people need to use it to give
> feedback to you.
>
> -Ralph

Ok.

I could put all what is waiting in old .cs style and I don't was able to do
using our procedure ?

We have the Multillingual thing and the
http://bugs.squeak.org/view.php?id=6348

Using Monticello, blows or hang target image.
Using old .cs, no problems and we have more interesting things into.

Or we could have two branch.
One regular with Etoys.
One "Wild" what don't have Etoys and Nebraska , but people could load.
As is alpha, they should know what at this date not all works.

Damien Cassou schrieb: Hello Damien,

I am using your dev-Images, too. Alas I have problems with it.
For example:
1. Start the Refactoring Browser.
    Select a unary message selector. Press "R"-button and then "Method"
-> "10. Rename Method".
    Enter a new name and accept it.
    => MNU: BrowserRequestor>>getArgumentPermutation

2. Open the Preferences Browser.
    Select "general" category. Select "bigDisplay" and revert the selection.
    Now a TTCFont schema is active.
    Create a new morphic project. Enter this new project and then return
to the previous.
    => MNU: TTCFont>>fontArray
    This is bug #5451 and it also happens in 3.10 alpha.

Regards
Andreas

Hello Andreas,

2007/4/22, Andreas Wacknitz <[hidden email]>:
> Hello Damien,
>
> I am using your dev-Images, too. Alas I have problems with it.

Thank you for the report.


> For example:
> 1. Start the Refactoring Browser.
>     Select a unary message selector. Press "R"-button and then "Method"
> -> "10. Rename Method".
>     Enter a new name and accept it.
>     => MNU: BrowserRequestor>>getArgumentPermutation


This is a known bug: http://bugs.squeak.org/view.php?id=6354

You may want to try using OmniBrowser instead of the refactoring
browser. In recent squeak-dev images, it contains refactorings too.
Those ones seems to work.

> 2. Open the Preferences Browser.
>     Select "general" category. Select "bigDisplay" and revert the selection.
>     Now a TTCFont schema is active.
>     Create a new morphic project. Enter this new project and then return
> to the previous.
>     => MNU: TTCFont>>fontArray
>     This is bug #5451 and it also happens in 3.10 alpha.

This one is not due to the squeak-dev image. Let's hope it will be
corrected soon.

Thank you for your reports.

--
Damien Cassou

Damien Cassou schrieb: Doh, I haven't seen this bug report, thank you.

>
> You may want to try using OmniBrowser instead of the refactoring
> browser. In recent squeak-dev images, it contains refactorings too.
> Those ones seems to work.
I have perceived  OmniBrowser  as the least interesting browser so far.
Normally I open a browser by means of the "open" menu.
OmniBrowser is only accessible via the "class browser" entry,
if you have chosen it as the default. Otherwise it vanishes and isn't
accessible via a menu entry anymore.
By the way, the open menu should have an overhaul.
There are four source browsers available in the open menu of your
dev-image:
    1. class browser                 : initially let you choose one of
                                                    #PackagePaneBrowser
                                                    #Browser
                                                    #RefactoringBrowser
                                                    #OBSystemBrowserAdaptor
    2. package pane browser   : opens a browser with title "Package Browser"
    3. Image Browser              : opens a browser with title "System
Browser"
    4. Refactoring Browser     : opens a browser with title "RB"

(#Browser and #OBSystemBrowserAdaptor seem to be the same as number 3 -
Image Browser.)
IMO the menu entries and the titles shall match. Now it is a little bit
confusing.
(And the Refactoring Browser shall have a sound title.)
Furthermore it would be nice to have a convention for capitalisation
(e.g. "class browser" versus "Image Browser").
Thus, the open menu lacks consistency.


Regards
Andreas

El 4/22/07 9:07 AM, "Andreas Wacknitz" <[hidden email]> escribió:

>   This is bug #5451 and it also happens in 3.10 alpha.

This is not more a problem, thanks Jerome work.
I sending the .mcz to 310 repository and some time this day you could hit
the update button and receive 7092TTCFontDNUfix.cs.


Edgar

2007/4/22, Andreas Wacknitz <[hidden email]>:
> I have perceived  OmniBrowser  as the least interesting browser so far.
> Normally I open a browser by means of the "open" menu.

I would like to have only the OmniBrowser in the squeak-dev image. I'm
waiting for the massive clean of 3.10.

However, OmniBrowser is easy to start in a squeak-dev image:

- it's always accessible in Open...->Image Browser (what a clear name,
isn't it?)
- it can be drag&dropped from the Tools flap on the right side of the world.

The squeak-dev image can be enhance by everyone. The package to be
modified is on squeaksource :
http://www.squeaksource.com/ImageForDevelopers. Feel free to rename
things :-)

--
Damien Cassou

El 4/22/07 10:54 AM, "Damien Cassou" <[hidden email]> escribió:

> I'm
> waiting for the massive clean of 3.10.
Have nice couch ?  :=)

Hello,

when I update a 3.10alpha7081 image to 7092, I'll get a merge window
with "Merging Multilingual-edc.22" title. Selecting the "All Newer" button
leads to a MNU: UndefinedObject>>fullTimeStamp.
This happens in MCConflict>>isLocalNewer.

Regards
Andreas

On Apr 22, 2007, at 6:14 AM, Andreas Wacknitz wrote:

*sigh*

Yes, this bit is kind of ugly.

1. is especially ugly, because the mechanism for registering browsers  
for use as the default browser insists on displaying the class name  
in the menu, rather than letting the registered class decide what it  
should be called. It would be really nice to get that fixed. In fact,  
Browser and OBSystemBrowserAdaptor are quite different, but the  
*look* quite similar when they are first opened.

3. Image browser isn't the most descriptive name. It's a holdover  
from the early days of OmniBrowser, where it was useful for  
distinguishing between File Browser and Package Browser, which were  
also included in OB. Probably the most accurate name would be  
Environment Browser, but that's not terribly informative either. I'm  
happy to change it if we can come up with something better. I don't  
want to just call it "OmniBrowser" since there are lots of browsers  
built using the library, and OBSystemBrowser is just one of them.

Colin

El 4/22/07 2:37 PM, "Andreas Wacknitz" <[hidden email]> escribió:

> Hello,
>
> when I update a 3.10alpha7081 image to 7092, I'll get a merge window
> with "Merging Multilingual-edc.22" title. Selecting the "All Newer" button
> leads to a MNU: UndefinedObject>>fullTimeStamp.
> This happens in MCConflict>>isLocalNewer.
>
> Regards
> Andreas

Try to update again, please.
Sorry what Mulltilingual is  very sensitive and what I test on Mac could
fail when go to Linux server and to any OS machine.
Now works on complete Argentina to Deutschland and back travel.
And thanks for reporting .

Edgar

Edgar J. De Cleene schrieb:

El 4/22/07 2:37 PM, "Andreas Wacknitz" [hidden email] escribió:

  

Hello,

when I update a 3.10alpha7081 image to 7092, I'll get a merge window
with "Merging Multilingual-edc.22" title. Selecting the "All Newer" button
leads to a MNU: UndefinedObject>>fullTimeStamp.
This happens in MCConflict>>isLocalNewer.

Regards
Andreas
    

Try to update again, please.
Sorry what Mulltilingual is  very sensitive and what I test on Mac could
fail when go to Linux server and to any OS machine.
Now works on complete Argentina to Deutschland and back travel.
And thanks for reporting .

Edgar



  

OK, this time all went fine. Thank you for your hard work.

Best regards
Andreas