Command Pattern implementation with bound winforms controls?

[Samuel R. Neff]

We're working on implementing a Command Pattern design with our
application in order to help facilitate undo/redo. This is fine four
user actions we control which basically means menu actions.

However, it's not quite so clear for bound controls. We don't
directly interact with the change--the user makes a change in a
control, such as a grid, and the dataset stores the update. We
haven't interjected any code which we can funnel through a Command
object.

One work-around we're thinking is to react to the change events and
put a Command object in the stack wich some flag to indicate that the
action has already been accomplished (and thus the Execute method only
should be called for the Redo operation and not for the originating
operation).

How have others addressed this situation? Are there any examples
available of using Command pattern in a bound-control environment?

Thanks,

Sam

B-Line is now hiring one Washington D.C. area VB.NET
developer for WinForms + WebServices position.
Seaking mid to senior level developer. For
information or to apply e-mail resume to
sam_blinex_com.

[Peter Huang [MSFT]]

Hi

I think it is not a trivial job.
Here is a link for your reference.
http://www.thecodeproject.com/vb/net...f=100&forumid=
14573&select=817584#xx817584xx

Also I think you may try to post the databinding issue in the newsgroup
below.
Thanks!
microsoft.public.dotnet.framework.windowsforms.dat abinding

Best regards,

Peter Huang
Microsoft Online Partner Support

Get Secure! - www.microsoft.com/security
This posting is provided "AS IS" with no warranties, and confers no rights.

[Bob Powell [MVP]]

I recently created an architecture that did this.

The undo-redo architecture was interface based with the specific
implementations for the various objects conforming to the contract of the
interface but imlementing differently depending upon object type. The
interface included a do, undo and redo method. This separated the do from
the redo and enables you to create a null do with an operative redo.

For objects such as textboxes, a change notification occurred via the
standard events which created and stacked the command object as per normal,
the system called the null do command because the assumption was that
databinding would already have effected the change and subsequent undo /
redo actions were implemented using specialised code that had knowledge of
the bound control type being undone.

I can't post code as it was for a commercial client but the process was
pretty standard Command Pattern.

--
Bob Powell [MVP]
Visual C#, System.Drawing

Find great Windows Forms articles in Windows Forms Tips and Tricks
http://www.bobpowell.net/tipstricks.htm

Answer those GDI+ questions with the GDI+ FAQ
http://www.bobpowell.net/faqmain.htm

All new articles provide code in C# and VB.NET.
Subscribe to the RSS feeds provided and never miss a new article.

"Samuel R. Neff" <bl****@newsgroup.nospam> wrote in message
news:8p********************************@4ax.com...

We're working on implementing a Command Pattern design with our
application in order to help facilitate undo/redo. This is fine four
user actions we control which basically means menu actions.

However, it's not quite so clear for bound controls. We don't
directly interact with the change--the user makes a change in a
control, such as a grid, and the dataset stores the update. We
haven't interjected any code which we can funnel through a Command
object.

One work-around we're thinking is to react to the change events and
put a Command object in the stack wich some flag to indicate that the
action has already been accomplished (and thus the Execute method only
should be called for the Redo operation and not for the originating
operation).

How have others addressed this situation? Are there any examples
available of using Command pattern in a bound-control environment?

Thanks,

Sam

B-Line is now hiring one Washington D.C. area VB.NET
developer for WinForms + WebServices position.
Seaking mid to senior level developer. For
information or to apply e-mail resume to
sam_blinex_com.

[Samuel R. Neff]

Thanks Bob, that looks like a good solution.

Peter's post brought up an issue I hadn't considered which is when
responding to events and putting commands with empty Do methods on the
undo stack, how do you differentiate from user initiated events and
code initiated events? Seems like that still is an issue with your
proposal.

Thanks,

Sam
On Wed, 23 Mar 2005 13:50:14 +0100, "Bob Powell [MVP]"
<bob@_spamkiller_bobpowell.net> wrote:

I recently created an architecture that did this.

The undo-redo architecture was interface based with the specific
implementations for the various objects conforming to the contract of the
interface but imlementing differently depending upon object type. The
interface included a do, undo and redo method. This separated the do from
the redo and enables you to create a null do with an operative redo.

For objects such as textboxes, a change notification occurred via the
standard events which created and stacked the command object as per normal,
the system called the null do command because the assumption was that
databinding would already have effected the change and subsequent undo /
redo actions were implemented using specialised code that had knowledge of
the bound control type being undone.

I can't post code as it was for a commercial client but the process was
pretty standard Command Pattern.

B-Line is now hiring one Washington D.C. area VB.NET
developer for WinForms + WebServices position.
Seaking mid to senior level developer. For
information or to apply e-mail resume to
sam_blinex_com.

[Peter Huang [MSFT]]

Hi

What do you mean by initiated events and code initiated events?
Do you mean the event which is defined by the user code or by the
DataGrid's buildin event?
If so, I think we can add more argument in the User Code Event to indicate
the difference.

Best regards,

Peter Huang
Microsoft Online Partner Support

Get Secure! - www.microsoft.com/security
This posting is provided "AS IS" with no warranties, and confers no rights.

[Samuel R. Neff]

I mean if we're listening to the ColumnChanged event on a DataTable,
we don't know if the event was triggered because the user did
something in the bound grid or because some other code applied a
change to the table.

However, another colleague pointed out the grid we're using,
Infragistics UltraGrid, has change events we can use instead of the
DataTable events and this will provide for the necessary
differentiation in our case--grids are the only place we're using data
binding.

Thanks,

Sam
On Thu, 24 Mar 2005 08:33:03 GMT, v-******@online.microsoft.com
("Peter Huang" [MSFT]) wrote:

Hi

What do you mean by initiated events and code initiated events?
Do you mean the event which is defined by the user code or by the
DataGrid's buildin event?
If so, I think we can add more argument in the User Code Event to indicate
the difference.

Best regards,

Peter Huang
Microsoft Online Partner Support

Get Secure! - www.microsoft.com/security
This posting is provided "AS IS" with no warranties, and confers no rights.

B-Line is now hiring one Washington D.C. area VB.NET
developer for WinForms + WebServices position.
Seaking mid to senior level developer. For
information or to apply e-mail resume to
sam_blinex_com.

[Peter Huang [MSFT]]

Hi Samuel,

Thanks for your feedback.
I agree that if we want to check who fire the columnchanged event, we may
try to monitor the grid's related change event.
And thanks for your sharing the experience in the community.
Best regards,

Peter Huang
Microsoft Online Partner Support

Get Secure! - www.microsoft.com/security
This posting is provided "AS IS" with no warranties, and confers no rights.