Boring confirmation message...

Hi All,

Often, I do a print-it in a debugger or browser. For some mysterious  
reason, I subsequently press the delete key that erases what was just  
print. Some changes occurred therefore. When I then select another  
method (or frame if I am in a debugger), a confirmation whether I want  
to save my changes is done. I find this very boring. Am i the only one ?

A simple check whether the textual content has been modified would  
prevent this stupid question.

Cheers,
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






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

Alexandre,

There is a preference for confirm on close or something like that - I think.  Your test for change in value is an interesting idea, but I am often willing/expecting to toss whatever I type into an inspector.  I very much want the system to warn me if I am about to discard code in the debugger or browser source pane, and certainly in workspaces.

Bill


----
Wilhelm K. Schwab, Ph.D.
bschwab AT anest DOT ufl DOT edu

________________________________________
From: [hidden email] [[hidden email]] On Behalf Of Alexandre Bergel [[hidden email]]
Sent: Saturday, February 14, 2009 2:09 PM
To: Pharo Development
Subject: [Pharo-project] Boring confirmation message...

Hi All,

Often, I do a print-it in a debugger or browser. For some mysterious
reason, I subsequently press the delete key that erases what was just
print. Some changes occurred therefore. When I then select another
method (or frame if I am in a debugger), a confirmation whether I want
to save my changes is done. I find this very boring. Am i the only one ?

A simple check whether the textual content has been modified would
prevent this stupid question.

Cheers,
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






_______________________________________________
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

2009/2/14 Alexandre Bergel <[hidden email]>: pressing Alt-l in a text morph will restore it contents back to
original, and clear 'modified' flag.
I using it often to avoid answering this question. :)



--
Best regards,
Igor Stasenko AKA sig.

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

Wow... I did not know that
thanks Igor

Alexandre


On 14 Feb 2009, at 21:48, Igor Stasenko wrote:

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






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

Hi Bill

> There is a preference for confirm on close or something like that -  
> I think.  Your test for change in value is an interesting idea, but  
> I am often willing/expecting to toss whatever I type into an  
> inspector.  I very much want the system to warn me if I am about to  
> discard code in the debugger or browser source pane, and certainly  
> in workspaces.

I agree with you, and everybody wants this behavior. If I type  
something in a debugger or browser, I should get a warning to prevent  
loss upon closing it. However, if I just do a 'printIt' of an  
expression (without typing anything else), i do not think there is a  
reason to raise a warning if I want to close the window. I see myself  
confirming "yes, I am sure I want to discard the content" without  
having typed any code. Am I the only one to think this?

For workspace, I added an history few weeks ago. You can now close a  
workspace without having a warning. The content is inserted into a  
history collection, that once can retrieve at anytime. I hope this  
behavior will satisfy you.

Cheers,
Alexandre

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






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

Alexandre,

How far back does the workspace history go?  Can it be cleared?  Perhaps unique to my working style, I end up with a workspace that stays open and becomes a journal; I save its contents to disk every so often, and finally it gets so big that I start a new one under a new file name.  They are not quite as useful as they once were because I have started using comments in methods to hold expressions to make things happen.  It is nice to have the resulting "context sensitive help" but the journal idea suffers for it.

Back on topic, I typically inspect rather than print, probably for just this type of reason.  Most changes to text in the debugger are "real" as a result, so a warning is appreciated more than dreaded.  Most things I type in an inspector I am willing to lose rather than be bothered every time I close one, especially since I close so many because I am using them to avoid making changes in the debugger (now that you've gotten me to think about it).

Putting all of that together, I think I would want the following:

 * debugger and browser code panes warn
 * workspaces warn on close; history is a nice addition
 * inspectors have history and do not warn on close

Just what a history means for inspectors (class level, chronological, sorted based on the species of the inspected object??) is open for debate.

Bill


----
Wilhelm K. Schwab, Ph.D.
bschwab AT anest DOT ufl DOT edu

________________________________________
From: [hidden email] [[hidden email]] On Behalf Of Alexandre Bergel [[hidden email]]
Sent: Sunday, February 15, 2009 4:23 AM
To: [hidden email]
Subject: Re: [Pharo-project] Boring confirmation message...

Hi Bill

> There is a preference for confirm on close or something like that -
> I think.  Your test for change in value is an interesting idea, but
> I am often willing/expecting to toss whatever I type into an
> inspector.  I very much want the system to warn me if I am about to
> discard code in the debugger or browser source pane, and certainly
> in workspaces.

I agree with you, and everybody wants this behavior. If I type
something in a debugger or browser, I should get a warning to prevent
loss upon closing it. However, if I just do a 'printIt' of an
expression (without typing anything else), i do not think there is a
reason to raise a warning if I want to close the window. I see myself
confirming "yes, I am sure I want to discard the content" without
having typed any code. Am I the only one to think this?

For workspace, I added an history few weeks ago. You can now close a
workspace without having a warning. The content is inserted into a
history collection, that once can retrieve at anytime. I hope this
behavior will satisfy you.

Cheers,
Alexandre

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






_______________________________________________
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

> How far back does the workspace history go?  Can it be cleared?  
> Perhaps unique to my working style, I end up with a workspace that  
> stays open and becomes a journal; I save its contents to disk every  
> so often, and finally it gets so big that I start a new one under a  
> new file name.  They are not quite as useful as they once were  
> because I have started using comments in methods to hold expressions  
> to make things happen.  It is nice to have the resulting "context  
> sensitive help" but the journal idea suffers for it.

First time I've heard about this way of using workspaces. Keith  
produced an enhancement to save a workspace content in a file.

> Putting all of that together, I think I would want the following:
>
> * debugger and browser code panes warn

Me too. You do not have a warning if you haven't touch a text pane. If  
I just do a print-It in a text pane, I haven't produced any new code  
that deserve to be warned against loosing it.

> * workspaces warn on close; history is a nice addition

Before I produce my change that is now in Pharo, I did a small poll on  
this mailing list: who actually answer yes when trying to close a  
workspace. Two people answered (as far as I recall) saying that they  
never press 'Yes'. Since this is something I wanted to change for  
long, I did it.

> * inspectors have history and do not warn on close

There is no warning on inspector.

> Just what a history means for inspectors (class level,  
> chronological, sorted based on the species of the inspected  
> object??) is open for debate.

Agree. For now, it is just an ordered collection of texts contained in  
a pane.

Cheers,
Alexandre

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






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

Why not combine your history idea with ScriptManager package, have
workspaces broadcast their contents and have ScriptManager subscribe to
the broadcast, recording the history, it then provides a browser for the
history. The whole thing then becomes an optionally loadable package.

regards

Keith

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

Looks exiting. But I've never used ScriptManager. And my workspace  
content are highly disposable. I would need a concrete need actually...

Cheers,
Alexandre


On 15 Feb 2009, at 17:05, Keith Hodges wrote:

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.






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