Closure fixes

If you are interested in testing the closure fixes, I have scripted
it.  Run each section in turn.  I have a few things for discussion,
but first some feedback would be useful.
Cheers,
Mike


"A. Load Gofer"
ScriptLoader new installGofer.


"B. Install packages"
| gofer |
gofer := Gofer new.
gofer
        url: 'http://www.squeaksource.com/PharoTaskForces';
        addAll: #(
                "Part 2"
                'Kernel-MikeRoberts.405'

                "Part 3"
                'Compiler-MikeRoberts.130'
                'CompilerTests-MikeRoberts.21'
                'KernelTests-MikeRoberts.150'

                "Part 4"
                'Compiler-MikeRoberts.131'

                "Part 5"
                'Kernel-MikeRoberts.406'
               
                "Part 6"
                'Kernel-MikeRoberts.407'
               
                "Part 7"
                'System-Changes-MikeRoberts.13'
                'System-Download-MikeRoberts.11'
                'System-Support-MikeRoberts.89'
               
                "Part 8"
                'Tools-MikeRoberts.202'
               
                "Part 9 - setup recompile"
                'System-Support-MikeRoberts.90'
                ).
gofer load.


"C. Recompile"
Utilities initializeClosures.
Utilities compileUsingClosures.


"D. Run the tests please"
TestRunner open

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Does this mean that gofer depends in yet-to-be-integrated closure fixes?  That's fine, but it would probably stomp on my plan/hope to have gofer help me organize getting out of my current image and into an up-to-date one.

Maybe the refactoring code checker that Lukas mentioned is "hidden" by my current choice of toolset and will do the job.

Bill



-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of Michael Roberts
Sent: Sunday, September 20, 2009 9:27 AM
To: Pharo-project
Subject: [Pharo-project] Closure fixes

If you are interested in testing the closure fixes, I have scripted it.  Run each section in turn.  I have a few things for discussion, but first some feedback would be useful.
Cheers,
Mike


"A. Load Gofer"
ScriptLoader new installGofer.


"B. Install packages"
| gofer |
gofer := Gofer new.
gofer
        url: 'http://www.squeaksource.com/PharoTaskForces';
        addAll: #(
                "Part 2"
                'Kernel-MikeRoberts.405'

                "Part 3"
                'Compiler-MikeRoberts.130'
                'CompilerTests-MikeRoberts.21'
                'KernelTests-MikeRoberts.150'

                "Part 4"
                'Compiler-MikeRoberts.131'

                "Part 5"
                'Kernel-MikeRoberts.406'
               
                "Part 6"
                'Kernel-MikeRoberts.407'
               
                "Part 7"
                'System-Changes-MikeRoberts.13'
                'System-Download-MikeRoberts.11'
                'System-Support-MikeRoberts.89'
               
                "Part 8"
                'Tools-MikeRoberts.202'
               
                "Part 9 - setup recompile"
                'System-Support-MikeRoberts.90'
                ).
gofer load.


"C. Recompile"
Utilities initializeClosures.
Utilities compileUsingClosures.


"D. Run the tests please"
TestRunner open

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

cool to see that you can script that.

Stef

On Sep 20, 2009, at 4:27 PM, Michael Roberts wrote:


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On 20.09.2009, at 10:27, Michael Roberts wrote:

> If you are interested in testing the closure fixes, I have scripted
> it.  Run each section in turn.  I have a few things for discussion,
> but first some feedback would be useful.

I will have a look!

        Marcus




--
Marcus Denker - http://marcusdenker.de
PLEIAD Lab - Computer Science Department (DCC) - University of Chile


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Apparently when I run the tests the image freezes.

I also thought that the image was recompiled in squeak.

Stef


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

>> If you are interested in testing the closure fixes, I have scripted
>> it.  Run each section in turn.  I have a few things for discussion,
>> but first some feedback would be useful.
>
> I will have a look!

Excellent!
Once this is working I would declare it release candidates (more or  
less) (and tag red tests as broken ones).

Stef

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

No, the fixes are not dependent on Gofer. I am just using it to script
the load.  There's been a lot of chat about Gofer - i had no idea what
it was. so i figured i'd use it for something!

If your image freezes that's bad. Did it ever come back?  I did this
from #10451.  The only thing i was worried about was if I had not
copied all the packages to task forces, since they are all in my local
cache.

There are also a few test classes to change, but i'll do that after
this is in the update stream. i.e there are 2 unexpected passes in my
image because some closure/compiler tests now pass. (excellent!)

I've done the recompilation slightly differently. This was one of the
things i wanted to talk about.  Andreas made various changes to MC
packages to get it in, overriding initialize and things on various
classes. I'm not sure we need that since we can use ScriptLoader.
However this needs to be tested.  I just trigger the recompile that
squeak used.

There is not a lot of UI feedback. the whole thing could do with being
inside a Cursor busy during: [] etc.

cheers,
Mike

On Sun, Sep 20, 2009 at 6:01 PM, Stéphane Ducasse
<[hidden email]> wrote:
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

> No, the fixes are not dependent on Gofer. I am just using it to script
> the load.  There's been a lot of chat about Gofer - i had no idea what
> it was. so i figured i'd use it for something!

Looks cool.

However, from your comments I doubt that Gofer does what you expect.
Gofer does not load the packages in the order specified, but instead
merges the contents of all specified packages into one single load
operation. From the comments in the script I think that every Part 1
to 9 should all use their own Gofer instance to be truly sequential.

Cheers,
Lukas

--
Lukas Renggli
http://www.lukas-renggli.ch

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

hah! ok, no that wasn't what i expected. that will be why stef's image
dies. but i guess i was lucky when i tried it.
let me test out a new script and i will repost.

cheers,
Mike

On Sun, Sep 20, 2009 at 7:30 PM, Lukas Renggli <[hidden email]> wrote:
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Great!

Stef

On Sep 20, 2009, at 8:36 PM, Michael Roberts wrote:


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

This is tricky stuff. I have done this 4/5 times now, and managed to
get to the end and run all the tests twice.. but now it doesn't work.
oh well, it falls over in Part 7.  DNU method:. need to go through the
changesets and try and work out what's going on.  obviously the more
eyes on this the better.

cheers,
Mike

"A. Load Gofer"
ScriptLoader new installGofer.

"B. Install packages"
| gofer |
gofer := Gofer new.
gofer url: 'http://www.squeaksource.com/PharoTaskForces'.
Transcript cr; show: 'Part 2'; cr.
gofer copy
        addAll: #(
                "Part 2"
                'Kernel-MikeRoberts.405');
        load.
Transcript cr; show: 'Part 3'; cr.
gofer copy
        addAll: #(
                "Part 3"
                'Compiler-MikeRoberts.130'
                'CompilerTests-MikeRoberts.21'
                'KernelTests-MikeRoberts.150');
        load.
Transcript cr; show: 'Part 4'; cr.
gofer copy
        addAll: #(
                "Part 4"
                'Compiler-MikeRoberts.131');
        load.
Transcript cr; show: 'Part 5'; cr.
gofer copy
        addAll: #(
                "Part 5"
                'Kernel-MikeRoberts.406');
        load.
Transcript cr; show: 'Part 6'; cr.
gofer copy
        addAll: #(
                "Part 6"
                'Kernel-MikeRoberts.407');
        load.
Transcript cr; show: 'Part 7'; cr.
gofer copy
        addAll: #(
                "Part 7"
                'System-Changes-MikeRoberts.13');
        load.
gofer copy
        addAll: #(
                'System-Download-MikeRoberts.11');
        load.
gofer copy
        addAll: #(
                'System-Support-MikeRoberts.89');
        load.
Transcript show: 'Part 8'; cr.
gofer copy
        addAll: #(
                "Part 8"
                'Tools-MikeRoberts.202');
        load.
Transcript cr; show: 'Part 9'; cr.
gofer copy
        addAll: #(
                "Part 9 - setup recompile"
                'System-Support-MikeRoberts.90');
        load.
       
       
"C. Recompile"
Utilities initializeClosures.
Utilities compileUsingClosures.

"D. Run the tests please"
TestRunner open






On Sun, Sep 20, 2009 at 7:47 PM, Stéphane Ducasse
<[hidden email]> wrote:
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

The case for installer... QED

Keith



_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

For anyone actually reviewing this, it breaks after Part 4. Inability
to recompile the system.  There are MethodProperties around but i'm
not sure there should be.  need to look closer at the trunk update
mechanism.

cheers,
Mike

"A. Load Gofer"
ScriptLoader new installGofer.

"B. Install packages"
| gofer packages |
gofer := Gofer new.
gofer url: 'http://www.squeaksource.com/PharoTaskForces'.

packages := #(
                "Part 2"
                'Kernel-MikeRoberts.405'

                "Part 3"
                'Compiler-MikeRoberts.130'
                'CompilerTests-MikeRoberts.21'
                'KernelTests-MikeRoberts.150'
       
                "Part 4"
                'Compiler-MikeRoberts.131').
packages do: [:each |
        gofer copy
                add: each;
                load].

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

I have a few new methods from Eliot to include after he looked at my
debug log. I will push out some packages tonight hopefully.

Cheers mike

On Monday, September 21, 2009, Michael Roberts <[hidden email]> wrote:
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

ok

Stef

On Sep 23, 2009, at 9:36 AM, Michael Roberts wrote:


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

"A. Load Gofer"
ScriptLoader new installGofer.

"B. Install packages"
| gofer packages |
gofer := Gofer new.
gofer url: 'http://www.squeaksource.com/PharoTaskForces'.

packages := #(
                "Part 2.1"
                'Kernel-MikeRoberts.408'
               
                "Part 3"
                'CompilerTests-MikeRoberts.21'
                'KernelTests-MikeRoberts.150'
                'Compiler-MikeRoberts.130'
               
                "Part 4"
                'Compiler-MikeRoberts.131'

                "Part 5"
                'Kernel-MikeRoberts.409'

                "Part 6"
                'Kernel-MikeRoberts.411'
               
                "Part 7"
                'System-Changes-MikeRoberts.13'
                'System-Download-MikeRoberts.11'
                'System-Support-MikeRoberts.89'

                "Part 8"
                'Tools-MikeRoberts.202'

                "Part 9 - setup recompile"
                'System-Support-MikeRoberts.90').
               
packages do: [:each |
        gofer copy
                add: each;
                load].
       
"C. Recompile"
Utilities initializeClosures.
Utilities compileUsingClosures.

"D. Run the tests please"
TestRunner open

_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Mike this is really cool.
Thanks a ***lot***

I got
7281 run, 7182 passes, 2 expected failures, 80 failures, 15 errors, 2  
unexpected passes


while reading the code and checking differences. I noticed that the  
CompiledMethod>>compilerClass, parserClass and decompilerClass got  
changed and hardcoded the Compiler instead of using compilerClass and  
others.

http://code.google.com/p/pharo/issues/detail?id=1238


Now for the integration


On Sep 24, 2009, at 12:23 AM, Michael Roberts wrote:


_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

hey that's cool. the 2 unexpected passes should be closure tests. they
basically prove the integration from what i can tell since they were
marked as expected failures. The decompiler tests fail a lot, but we
can work away at that for 1.1 1.2 etc. It looks quite a big job from
what Eliot tells me.  Shows how important test coverage is!!

Mike

On Thu, Sep 24, 2009 at 9:03 AM, Stéphane Ducasse
<[hidden email]> wrote:
_______________________________________________
Pharo-project mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project