Smalltalk › Pharo › Pharo Smalltalk Developers

[update 3.0] #30022

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

15 messages Options

EstebanLM

[update 3.0] #30022

30022
-----

10032 Integrate new debugger
https://pharo.fogbugz.com/f/cases/10032

Diff information:
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33

tinchodias

Re: [update 3.0] #30022

Sounds great. No idea how to notice the difference, but I tried the "Fuel out stack" option and it works.

On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:

30022
-----

10032 Integrate new debugger
https://pharo.fogbugz.com/f/cases/10032

Diff information:
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33

Sven Van Caekenberghe-2

Re: [update 3.0] #30022

On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:

> Sounds great. No idea how to notice the difference, but I tried the "Fuel out stack" option and it works.

Indeed, sounds very promising but what does it mean ?
There does not seem to be any description in the issue tracker either, or maybe I missed it.

> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
> 30022
> -----
>
> 10032 Integrate new debugger
> https://pharo.fogbugz.com/f/cases/10032
>
> Diff information:
> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33

Camillo Bruni-3

Re: [update 3.0] #30022

On 2013-04-05, at 17:29, Sven Van Caekenberghe <[hidden email]> wrote:

>
> On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:
>
>> Sounds great. No idea how to notice the difference, but I tried the "Fuel out stack" option and it works.
>
> Indeed, sounds very promising but what does it mean ?
> There does not seem to be any description in the issue tracker either, or maybe I missed it.

AFAIK that's a "very" old feature no?
=> open debugger => right click => last entry/second last entry => Fuel out stack

It will just serialize the whole stack and put it in a file next to the image.

philippeback

Re: [update 3.0] #30022

One question: why is the debugged menu (right click thing) taking so long to show up where the world menu pops almost immediately?

I wonder.

2013/4/5 Camillo Bruni <[hidden email]>

On 2013-04-05, at 17:29, Sven Van Caekenberghe <[hidden email]> wrote:

>
> On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:
>
>> Sounds great. No idea how to notice the difference, but I tried the "Fuel out stack" option and it works.
>
> Indeed, sounds very promising but what does it mean ?
> There does not seem to be any description in the issue tracker either, or maybe I missed it.

AFAIK that's a "very" old feature no?
=> open debugger => right click => last entry/second last entry => Fuel out stack

It will just serialize the whole stack and put it in a file next to the image.

Camillo Bruni-3

Re: [update 3.0] #30022

indeed strange...
To me it looks like the pragma collector spends a lot of time fetching stuff.
And I would even say it happens twice per menu popup?

On 2013-04-05, at 17:53, "[hidden email]" <[hidden email]> wrote:

> One question: why is the debugged menu (right click thing) taking so long
> to show up where the world menu pops almost immediately?
>
> I wonder.
>
>
> 2013/4/5 Camillo Bruni <[hidden email]>
>
>>
>> On 2013-04-05, at 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>>
>>>
>>> On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:
>>>
>>>> Sounds great. No idea how to notice the difference, but I tried the
>> "Fuel out stack" option and it works.
>>>
>>> Indeed, sounds very promising but what does it mean ?
>>> There does not seem to be any description in the issue tracker either,
>> or maybe I missed it.
>>
>>
>> AFAIK that's a "very" old feature no?
>> => open debugger => right click => last entry/second last entry => Fuel
>> out stack
>>
>> It will just serialize the whole stack and put it in a file next to the
>> image.
>>

philippeback

Re: [update 3.0] #30022

Where should I look to see that? Curious me :-)

2013/4/5 Camillo Bruni <[hidden email]>

indeed strange...
To me it looks like the pragma collector spends a lot of time fetching stuff.
And I would even say it happens twice per menu popup?

On 2013-04-05, at 17:53, "[hidden email]" <[hidden email]> wrote:

> One question: why is the debugged menu (right click thing) taking so long
> to show up where the world menu pops almost immediately?
>
> I wonder.
>
>
> 2013/4/5 Camillo Bruni <[hidden email]>
>
>>
>> On 2013-04-05, at 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>>
>>>
>>> On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:
>>>
>>>> Sounds great. No idea how to notice the difference, but I tried the
>> "Fuel out stack" option and it works.
>>>
>>> Indeed, sounds very promising but what does it mean ?
>>> There does not seem to be any description in the issue tracker either,
>> or maybe I missed it.
>>
>>
>> AFAIK that's a "very" old feature no?
>> => open debugger => right click => last entry/second last entry => Fuel
>> out stack
>>
>> It will just serialize the whole stack and put it in a file next to the
>> image.
>>

Camillo Bruni-3

Re: [update 3.0] #30022

I opened a debugger,
then I just sampled the UI process (WorldMenu >> System >> Start profiling UI)
then I right clicked a couple of times and moved the cursor out of the screen on top to stop the profiling.

After that you have a pretty nice list :)

On 2013-04-05, at 18:51, "[hidden email]" <[hidden email]> wrote:

> Where should I look to see that? Curious me :-)
>
>
> 2013/4/5 Camillo Bruni <[hidden email]>
>
>> indeed strange...
>> To me it looks like the pragma collector spends a lot of time fetching
>> stuff.
>> And I would even say it happens twice per menu popup?
>>
>> On 2013-04-05, at 17:53, "[hidden email]" <[hidden email]> wrote:
>>
>>> One question: why is the debugged menu (right click thing) taking so long
>>> to show up where the world menu pops almost immediately?
>>>
>>> I wonder.
>>>
>>>
>>> 2013/4/5 Camillo Bruni <[hidden email]>
>>>
>>>>
>>>> On 2013-04-05, at 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>>>>
>>>>>
>>>>> On 05 Apr 2013, at 17:18, Martin Dias <[hidden email]> wrote:
>>>>>
>>>>>> Sounds great. No idea how to notice the difference, but I tried the
>>>> "Fuel out stack" option and it works.
>>>>>
>>>>> Indeed, sounds very promising but what does it mean ?
>>>>> There does not seem to be any description in the issue tracker either,
>>>> or maybe I missed it.
>>>>
>>>>
>>>> AFAIK that's a "very" old feature no?
>>>> => open debugger => right click => last entry/second last entry => Fuel
>>>> out stack
>>>>
>>>> It will just serialize the whole stack and put it in a file next to the
>>>> image.
>>>>
>>
>>
>>

Andrei Chis

Re: [update 3.0] #30022

In reply to this post by EstebanLM

Just a few comments about this update.

Right now the current debugger is unchanged. This update only added some classes to the image

The new debugger is far from completed and until it's finished both will be in the image.

You can only open the new debugger from the Pre-Debug window if you choose the last option 'SpecDebug'.

Cheers,

Andrei

On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:

30022
-----

10032 Integrate new debugger
https://pharo.fogbugz.com/f/cases/10032

Diff information:
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33

Sean P. DeNigris

Re: [update 3.0] #30022

Administrator

In reply to this post by Camillo Bruni-3

Camillo Bruni-3 wrote

I just sampled the UI process

Cool trick!!

Cheers,
Sean

Igor Stasenko

Re: [update 3.0] #30022

In reply to this post by Sven Van Caekenberghe-2

On 5 April 2013 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>
> There does not seem to be any description in the issue tracker either, or maybe I missed it.
>
indeed, the record on issue tracker looks very informative. (in the attached).

>> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
>> 30022
>> -----
>>
>> 10032 Integrate new debugger
>> https://pharo.fogbugz.com/f/cases/10032
>>
>> Diff information:
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33
>

--
Best regards,
Igor Stasenko.

Screen Shot 2013-04-05 at 9.23.06 PM.png (175K) Download Attachment

Nicolas Cellier

Re: [update 3.0] #30022

Yep, very good example of faked quality.

Making changes at the rate of Pharo is demanding and I understand the will to by-pass some quality rules. But then, any one wanting to give lessons should observe the real usage of the bug tracker and think again, because this is far from being an isolated case.

Nicolas

2013/4/5 Igor Stasenko <[hidden email]>

On 5 April 2013 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>
> There does not seem to be any description in the issue tracker either, or maybe I missed it.
>
indeed, the record on issue tracker looks very informative. (in the attached).

>> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
>> 30022
>> -----
>>
>> 10032 Integrate new debugger
>> https://pharo.fogbugz.com/f/cases/10032
>>
>> Diff information:
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33
>

--
Best regards,
Igor Stasenko.

stephane ducasse

Re: [update 3.0] #30022

In reply to this post by Sven Van Caekenberghe-2

>>
>
> Indeed, sounds very promising but what does it mean ?
> There does not seem to be any description in the issue tracker either, or maybe I missed it.

What we integrated is a model (two classes) that andrei wrote based on the old debugger. The old debugger
does not make the distinction cleanly between a model and the view. So the logic is mixed with the display.
When andrei came at lille he wrote with ben a Spec interface. We should rethink it by the way because I prefer
the glamourst one much more :).

So this model is good because we can build tools on top.
What we plan in the future is to have
- to get a new UI based on the model and that this ui via this model still use the old logic (bytecode interpretation and the bugs we get with it)
- an AST interpreter and plug it with this model so get a really new debugger execution logic (not a bytecode mapping range to text one)

So Pharo30 will have at least two debuggers for a while
- the old one but we are rewamped ui and using the model of andrei
- the one based on AST interpretation using the rewamped ui and using the model of andrei

Stef

>
>> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
>> 30022
>> -----
>>
>> 10032 Integrate new debugger
>> https://pharo.fogbugz.com/f/cases/10032
>>
>> Diff information:
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33
>

stephane ducasse

Re: [update 3.0] #30022

In reply to this post by Igor Stasenko

Hi igor

since you were at the meeting where we built the list of actions, you could also have edited the entry
and add more information. Camille was nice to take notes for everbody and to open issues.
I added contents.

Now we can also use a bad and private wiki to store information that is clear for the core team
but it will not be nice.

Stef

> On 5 April 2013 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>>
>> There does not seem to be any description in the issue tracker either, or maybe I missed it.
>>
> indeed, the record on issue tracker looks very informative. (in the attached).
>
>>> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
>>> 30022
>>> -----
>>>
>>> 10032 Integrate new debugger
>>> https://pharo.fogbugz.com/f/cases/10032
>>>
>>> Diff information:
>>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
>>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33
>>
>
>
>
> --
> Best regards,
> Igor Stasenko.
> <Screen Shot 2013-04-05 at 9.23.06 PM.png>

stephane ducasse

Re: [update 3.0] #30022

In reply to this post by Nicolas Cellier

Yep, very good example of faked quality.
Making changes at the rate of Pharo is demanding and I understand the will to by-pass some quality rules. But then, any one wanting to give lessons should observe the real usage of the bug tracker and think again, because this is far from being an isolated case.

Nicolas

this is always easier to talk than to do :)

And you got the prove of it so can we look at the 95% of good practices instead of the 5% wrong?

Stef

Nicolas

2013/4/5 Igor Stasenko <[hidden email]>
On 5 April 2013 17:29, Sven Van Caekenberghe <[hidden email]> wrote:
>
> There does not seem to be any description in the issue tracker either, or maybe I missed it.
>
indeed, the record on issue tracker looks very informative. (in the attached).

>> On Fri, Apr 5, 2013 at 1:51 PM, Esteban Lorenzano <[hidden email]> wrote:
>> 30022
>> -----
>>
>> 10032 Integrate new debugger
>> https://pharo.fogbugz.com/f/cases/10032
>>
>> Diff information:
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/DebuggerModel-EstebanLorenzano.23
>> http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Debugger-EstebanLorenzano.33
>

--
Best regards,
Igor Stasenko.