Code highlighting

Saturday, May 07, 2011

Tutorial: Table Relation properties in AX 2012

As many of you already know, Microsoft has put in a lot of effort into normalizing the tables in AX 2012 and consolidating all the data modeling tools in one place - the table itself. As part of this effort, a number of new properties have been introduced on Tables, and in this post I would like to cover some of them, namely the properties on Table Relations.

The new properties you will find on a Table Relation in AX 2012 are:
  • Cardinality
  • RelatedTableCardinality
  • RelationshipType
  • Role
  • RelatedTableRole
  • UseDefaultRoleNames
  • CreateNavigationPropertyMethods
  • NavigationPropertyMethodNameOverride

Hua Chu from the AX team has written a Guideline document, explaining how these properties should be set for Relations you add to Tables in AX. Note that in AX 2012 most of the above information is not actually used at runtime. This is something that will happen in future releases.
I have modified the document so that it contains the information relevant for partners and customers extending the standard application and have uploaded it to my OneDrive.

The document requires certain knowledge of Entity Relationship Modeling (ERM) and UML notation.

Disclaimer:
This document is intended as a guideline only, and should not be used as a Step-by-Step instruction.
Changes to any of the described functionality might still happen before AX 2012 RTM.


Your feedback and questions are, as always, welcome.

Thursday, May 05, 2011

Tutorial: lockWindowUpdate() vs. lock()/unlock()

There are two method pairs in X++, that are used throughout the application by everyone writing some processing on application forms. These are:

element.lock();
element.unLock();


and

element.lockWindowUpdate(true);
element.lockWindowUpdate(false);


Now, not that many people know the difference between the two methods, and only very few think about why and when should each of them be used.
I will try to describe the behavior of these methods and at the end give some recommendations on how to use them. I have done some kernel code reading (with help of kernel dev. Andy Stach, who I would like to mention here), so what I write below is more or less backed up by code.
If you disagree with some of the recommendations though, please share your experience in using these methods through comments for this post.

FormRun.lockWindowUpdate()

is basically a wrapper around the LockWindowUpdate Win32 function. What it does is pretty simple:
When a window is locked, all attempt to draw into it or its children fail. Instead of drawing, the window manager remembers which parts of the window the application tried to draw into, and when the window is unlocked, those areas are invalidated so that the application gets another WM_PAINT message, thereby bringing the screen contents back in sync with what the application believed to be on the screen.
See the link on MSDN for a detailed description.
Note, that according to MSDN, it should not be used for general purpose suppression of redraw operations, but only when dealing with drag&drop operations. This does not hold true for AX, where this method is used all over the place to prevent redraw of controls on the form.
Another interesting point is that only one window can be locked at the same time. So, any nested calls to lockWindowUpdate will be ignored, but when unlocking, only the outer-most unlock will actually invoke the Win32 counterpart. Now, I have not seen this used in X++, which is for the better.

FormRun.lock()

is internally invoking lockWindowUpdate to prevent the redraw of the window, and then also prevents the IntelliMorph control layout engine from running. This is commonly used in scenarios where control properties affecting control arrangement are being set in a loop, which provides a performance optimization as it avoids redundant arrange calls being processed. On the other hand, when calling FormRun.unlock, more work will need to be done, compared to using lockWindowUpdate(false), where the control layout changes were actually processed by the layout engine, but simply not displayed.

So, based on my investigation, I would suggest to use the following recommendations when doing form development:

  • When formRun.resetSize() is used, specifically, when some controls become visible, increasing the form size, always use formRun.lock()/unlock(), otherwise the change in the size of the form might not get reflected on the screen correctly.
  • When changing multiple layout properties (Left, Width, etc.) on one or more controls, use lock/unlock
  • When you modify the properties that do not impact the layout of controls on the form, use formRun.lockWindowUpdate(), or, if there are only very few control properties being modified, do not lock the form window at all.