The Trunk: Tools-mt.990.mcz

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

The Trunk: Tools-mt.990.mcz

commits-2
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+ super modelWakeUpIn: aWindow.
+ self updateInspectors.!


Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

Christoph Thiede

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!




Carpe Squeak!
Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

marcel.taeumel
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!




Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

Christoph Thiede

Hi Marcel,


just an idea - as we are talking about adding possible mode checkboxes to an inspector in [1], we could also add a checkbox for stepping there. We already have such a menu item in the Process Browser ... And then we could add a preference for choosing the default stepping boolean. Hypothetically. :-)


Best,

Christoph


[1] http://forum.world.st/The-Trunk-Tools-eem-994-mcz-td5122902.html


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 13:58:32
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!




Carpe Squeak!
Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

marcel.taeumel
Hi Christoph, hi all.

Every time you provide an option, you expect the user to make a decision. Easy-to-use interfaces might be found on a different path. ;-) Just my two cents.

Best,
Marcel

Am 06.10.2020 10:40:46 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just an idea - as we are talking about adding possible mode checkboxes to an inspector in [1], we could also add a checkbox for stepping there. We already have such a menu item in the Process Browser ... And then we could add a preference for choosing the default stepping boolean. Hypothetically. :-)


Best,

Christoph


[1] http://forum.world.st/The-Trunk-Tools-eem-994-mcz-td5122902.html


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 13:58:32
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!




Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

Christoph Thiede

Also right. So let's defer this question again until someone has a real use case for stepping in debuggers ... :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Mittwoch, 7. Oktober 2020 09:04:52
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph, hi all.

Every time you provide an option, you expect the user to make a decision. Easy-to-use interfaces might be found on a different path. ;-) Just my two cents.

Best,
Marcel

Am 06.10.2020 10:40:46 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just an idea - as we are talking about adding possible mode checkboxes to an inspector in [1], we could also add a checkbox for stepping there. We already have such a menu item in the Process Browser ... And then we could add a preference for choosing the default stepping boolean. Hypothetically. :-)


Best,

Christoph


[1] http://forum.world.st/The-Trunk-Tools-eem-994-mcz-td5122902.html


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 13:58:32
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!




Carpe Squeak!
Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

Eliot Miranda-2


On Oct 7, 2020, at 7:46 AM, Thiede, Christoph <[hidden email]> wrote:



Also right. So let's defer this question again until someone has a real use case for stepping in debuggers ... :-)


I don’t understand.  I step in debuggers every day.  It is too slow.  That’s a use case,an important enough use case.  And I’d rather throw out the emphasis that I’ve done without for years than accept a debugger I can no longer trust to behave controllably.


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Mittwoch, 7. Oktober 2020 09:04:52
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph, hi all.

Every time you provide an option, you expect the user to make a decision. Easy-to-use interfaces might be found on a different path. ;-) Just my two cents.

Best,
Marcel

Am 06.10.2020 10:40:46 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just an idea - as we are talking about adding possible mode checkboxes to an inspector in [1], we could also add a checkbox for stepping there. We already have such a menu item in the Process Browser ... And then we could add a preference for choosing the default stepping boolean. Hypothetically. :-)


Best,

Christoph


[1] http://forum.world.st/The-Trunk-Tools-eem-994-mcz-td5122902.html


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 13:58:32
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!





Reply | Threaded
Open this post in threaded view
|

Re: The Trunk: Tools-mt.990.mcz

Christoph Thiede

Hi Eliot,


I think this is a misconception - in this thread, I was not referring to the step buttons in the debugger, but the - never implemented - automatic stepping (self-updating) logic of the debugger's inspectors' contents. Debuggers shouldn't be slow, of course, that's what we are discussing and trying to fix in that other thread. :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von Eliot Miranda <[hidden email]>
Gesendet: Mittwoch, 7. Oktober 2020 18:35:11
An: The general-purpose Squeak developers list
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 


On Oct 7, 2020, at 7:46 AM, Thiede, Christoph <[hidden email]> wrote:



Also right. So let's defer this question again until someone has a real use case for stepping in debuggers ... :-)


I don’t understand.  I step in debuggers every day.  It is too slow.  That’s a use case,an important enough use case.  And I’d rather throw out the emphasis that I’ve done without for years than accept a debugger I can no longer trust to behave controllably.


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Mittwoch, 7. Oktober 2020 09:04:52
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph, hi all.

Every time you provide an option, you expect the user to make a decision. Easy-to-use interfaces might be found on a different path. ;-) Just my two cents.

Best,
Marcel

Am 06.10.2020 10:40:46 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just an idea - as we are talking about adding possible mode checkboxes to an inspector in [1], we could also add a checkbox for stepping there. We already have such a menu item in the Process Browser ... And then we could add a preference for choosing the default stepping boolean. Hypothetically. :-)


Best,

Christoph


[1] http://forum.world.st/The-Trunk-Tools-eem-994-mcz-td5122902.html


Von: Squeak-dev <[hidden email]> im Auftrag von Taeumel, Marcel
Gesendet: Donnerstag, 1. Oktober 2020 13:58:32
An: squeak-dev; [hidden email]
Betreff: Re: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Hi Christoph,

it has "always" been that way.. I suppose that its (a) for performance reasonse and (b) you are "stopping time" when debugging anyway. :-)

Best,
Marcel

Am 01.10.2020 13:56:55 schrieb Thiede, Christoph <[hidden email]>:

Hi Marcel,


just a basic question - why do we disable stepping in Debuggers? :-)


Best,

Christoph


Von: Squeak-dev <[hidden email]> im Auftrag von [hidden email] <[hidden email]>
Gesendet: Dienstag, 29. September 2020 10:20:36
An: [hidden email]; [hidden email]
Betreff: [squeak-dev] The Trunk: Tools-mt.990.mcz
 
Marcel Taeumel uploaded a new version of Tools to project The Trunk:
http://source.squeak.org/trunk/Tools-mt.990.mcz

==================== Summary ====================

Name: Tools-mt.990
Author: mt
Time: 29 September 2020, 10:20:33.276444 am
UUID: e7da396b-a8ec-7a44-8f15-5ba7d0d30b37
Ancestors: Tools-mt.989

In debuggers, forward model-wake-up call to update inspectors. This fixes the bug where a 'nil' receiver had no fields in the receiver inspector. For example, try debug-it on any expression in the workspace to trigger that bug.

Note that the underlying challenge is that the inspectors in debuggers are not stepping. Field updates are triggered manually via #updateInspectors. An inspector's initial object is 'nil'. So, "changing" that to 'nil' will not update the field list in inspectors bc. it is considered redundant and will be ignored there. While stepping is not required to produce the initial field list, the model-wake-up call will usually trigger that initial update in a stand-alone inspector.

That's why I think that it is not necessary to change anything in the inspector code but only here in the debugger.

=============== Diff against Tools-mt.989 ===============

Item was added:
+ ----- Method: Debugger>>modelWakeUpIn: (in category 'user interface') -----
+ modelWakeUpIn: aWindow
+
+        super modelWakeUpIn: aWindow.
+        self updateInspectors.!





Carpe Squeak!