Setting a Blank Default URL for an IFrame in CRM 2011

It is common that we want an IFrame in an entity form and we need the URL of the IFrame to be set dynamically when the form loads using the setSrc() method in a script. But how can we set a blank default URL?

When you insert an IFrame in a form, the form designer forces you to enter a blank URL. Before Polaris (UR12) we could enter “about:blank” and we’re all set! However, you are now forced to enter a URL with a valid prefix (http, https, ftp, fttps):



If your IFrame will embed external web pages (such as www.bing.com ) then it is not a problem to enter the default URL and have the script modify the URL if required at run time. However, if your IFrame is used to display HTML pages that you have defined as CRM Web Resources then you will not want to enter the full URL (e.g. http://mydevcrmserver.com/myorg/WebResources/test.html ) because the URL would contain a link to your DEV organization and that means that every time you promote to test/prod or install the solution on a different server, you’d have to go and modify the URL manually. So what can you do?

One option is to set the defaul URL to some hard-coded external website (e.g. http://www.bing.com) and have your script update the URL at runtime so the IFrame loads your HTML web resource. Ugly!

The best option would probably be that Microsoft allows as to enter “about:blank” but it seems this is no longer allowed after UR12 (hopefully they will allow it in the future). But you don’t want to rely on that.

The second best option that should work for you is to define your own “blank” web resource (Thanks to Alex Ries for the suggestion):

1. Create a new HTML web resource, which is blank (<HTML/>)
2. In your form, instead of inserting an IFrame, insert a web resource and use your blank web resource you just created.
3. At run time, your script should still be able to modify the URL of your web resource to point to another web resource. The setSrc() method is supported on IFrame as well as web resource controls!

With this solution, you are no longer using an IFrame so you are not forced to set a default URL! Hope this helps you!

CRM 2011 Workflow Utilities released for CRM Online

You might know the CRM 2011 Workflow Utilities Codeplex project. The most popular request is now available: CRM Online support.

Today I have created a new release (v2.0.2.0) which now supports installing the solution in a CRM Online organization. Workflow utilities is a CRM solution that allows you to extend the CRM workflow design experience by providing the following custom workflow steps available directly from the CRM process designer.:

• Delete record
• Share or “unshare” a record
• Insert hyperlink to a CRM record
• Qualify lead (convert to account/contact/opportunity)
• Bulk activate / deactivate records (no record count limit)

For more details on how to install and use the solution please visit the Codeplex site. Thanks for all the feedback received, enjoy!

JavaScript, XML processing, DOM and CRM 2011 UR12

If you have scripts in CRM which somehow retrieve or process XML, chances are you will need to do some re-write or validations when applying Update Rollup 12 (UR12). This post explains how to fix you JavaScripts when you make use of the DOM parser for processing XML.

One of the highest-risk customizations to break with UR12 are JavaScripts and these are the 2 most common reasons why:

1. It is common for developers to perform unsupported operations with the page DOM instead of using the supported xRM APIs. Since UR12 introduces cross-browser support, accessing the DOM by unsupported methods such as ‘getElementById’ on CRM forms will no longer work because it has changed. Therefore, customizations like this one will break.

2. Because CRM has historically only been supported in IE, developers might have written JavaScripts that depend on IE extensions specific to Microsoft, these will also stop working because the assumption is no longer true (will be broken even in IE!). Even though you are not doing anything unsupported in your scripts, they can break and might need to be fixed.

There are plenty of tools and blog posts that talk about how to tackle #1, so in this article I will focus on problem #2 above and I will pick the specific example of processing XML.

Assume you are parsing and processing an XmlDocument in your JavaScript, perhaps you retrieved the xDoc from and XML HTTP request, using AJAX or even if you are simply parsing text into an XmlDocument object. In my case, I was retrieving an XML web resource using the following code:

1. function ShowHelpText() {
2.     var tooltipsTextFile = "ava_TooltipsText.xml";
3.  
4.     // Use AJAX to retrieve the tooltips texts
5.     $.ajax({
6.         type: "GET",
7.         async: true,
8.         url: tooltipsTextFile,
9.         dataType: "xml",
10.         success: parseTooltipsXml,
11.         error: function (xhr, textStatus, errorThrown) {
12.             alert("Error retrieving tooltips: " + xhr.message);
13.         }
14.     }); //end ajax
15. }

Before UR12, this code used to return a documentElement of type IXMLDOMElement which is a Microsoft extension to the W3C DOM. Therefore, I was happily calling methods such as selectNode and selectSingleNode. The problem is, after UR12, the same code will return a standard Document object which does not support selectSingleNode so I get the error “Object doesn’t support property or method ‘selectSingleNode’”:

 

In part, this is great news, because now your JavaScript will work in any browser (mostly), but you still need to do some clean-up to remove any dependencies on IE DOM extensions. In my case, I had to update my script to make use of getElementsByTagName and verify that I only use methods and properties that are W3C standard as documented here and process my XML according to that documentation (which works slightly different than if you have IE-specific JS).

Although my example is specific to parsing XML, the same can apply to any other scripts that rely on IE extensions not available in other browsers or not standard. I hope this post helps others be more aware and proactive before applying UR12 and in understanding how your scripts can break even if you are not using any unsupported customizations!

Problems with leading and trailing blank spaces in CRM data

I recently ran into a number of problems that ended all up being due to blankspaces in some CRM data. I thought of sharing where these problems come from, how blank spaces can affect CRM functionality and how you can resolve them.

You might wonder: how can you end up with leading or trailing blank spaces in text fields in CRM? While it is true that CRM forms are smart enough to remove these blank spaces upon form save, it might not be the case when the data comes in from data import or during integrations with external systems. It sounds like a small problem, but if you are not careful, blank spaces can become a data integrity nightmare for the following reasons:

1. You might end up with duplicate data that is hard to resolve later on. For example you might have account “123” and account “ 123 “ which reference to your same customer. Now all your activities, invoices and all related records are potentially spread across 2 different records. Fortunately, duplicate detection ignores leading/trailing blank spaces so these 2 records would be considered duplicate if you have a published rule on account number field. However, if these accounts are created from the SDK (or through an integration layer), chances are that the duplicate detection is not enforced because when making web service calls duplicate detection is off by default unless specified in the create/update message. Fixing duplicate parent records can be a really long and boring task.

2. Assume you have multiple contacts related to account “123” that you want to import using the data import wizard. However, your account appears in CRM as “ 123 “ (with blank spaces). It will be impossible for import wizard to match the contact’s parent account and import will fail because the parent account was not found. You might say: I can just change my import spreadsheet to reference account “ 123 “ instead of “123”. The answer is: it does not matter if in your import XML file you enter “123” or “ 123 “ as the parent account, CRM will not be able to resolve the parent account!! In short: You cannot use the out-of-the-box data import feature to import related records to that account until you fix the blank spaces, argh.

3. Once you fix all your integration points to trim all text fields before committing a create / update to CRM, how do you clean the thousands of records that already exist? My first guess was: Export to Excel for re-import, use some Excel functions to remove blank spaces and import back to CRM. Again to my surprise, removing leading/trailing blankspaces in text fields is not considered a change for CRM so during import, the records were not even processed! When you export records for re-import, CRM is smart enough to know that if you did not update a row, it will not re-import that row. However, it is not smart enough to figure out that you removed blank spaces so it also ignores your update.

4. Sometimes the CRM forms trim blank spaces for text fields. This can become a problem because if your data has trailing blank spaces, the CRM form will remove then on load, and automatically the form is marked as dirty! Therefore, just by opening the record your form is dirty and you get a warning about pending changes when you close the form even if you did not update the record. It also prevents some functionality since some of our ribbon buttons require the form to not be dirty. It is also not evident what was happening and took us a while to figure out why some forms where always getting marked as dirty.

 

We let our production CRM get really messy because we did not notice this problem until too late (we never noticed the external systems were sending data padded with blank spaces) so cleansing the data became quite challenging since it was spread to multiple fields in multiple entities, many duplicate parent and child records. As I mentioned earlier, fortunately duplicate detection does consider these 2 cases as duplicates (“123” and “ 123 “) so you can execute multiple duplicate detection jobs to find and resolve duplicates. However, it can be a really long task to do. In our case, because of the magnitude of the data cleanse that was required we had to invest in a data cleansing tool that would iterate through multiple records resolving the blank spaces and the duplicates which reduced considerably the amount of manual work required to solve the problem. The “tool” in a nutshell consisted simply in a custom workflow activity that takes as input a CRM query (Rollup Query) and then retrieves and cleans all the records specified as input. Then we can use on-demand workflows and provide different queries to resolve duplicates in multiple entities. In any case our conclusion is that we should have thought about this from the beginning and I hope this post helps other avoid the same problem before it is too late.

Script# (ScriptSharp) and CRM: The good and the bad

If you are familiar with Script# you probably understand how it can accelerate JavaScript development and completely eliminate annoyances such as accidental JavaScript syntax errors. However, does it work well for Dynamics CRM for scripts and web resources?

If you are not familiar with Script#, here is a very simple definition: Write your code in C# and you can then compile it into a JavaScript (instead of programming in JS directly). I have been using Script# for many years now (and love it), but I recently took the challenge for the first time to use it for designing and deploying all script customizations for an enterprise CRM project and quickly realized that there are a few issues about Script# and CRM that will make me think twice next time. Let’s start with the “good”:



1. Forget about JS syntax. Honestly, I hate JavaScript and client programming. One of the reasons is because I find it too error prone, I don’t know how many times I have run into issues with my JavaScript and often turns out to be something like forgetting a curly braket “}” somewhere in the syntax or accidentally typing a capital letter when it shouldn’t be. I am devoted to C# and that’s why I’m so motivated with Script#; isn’t it just awesome that I can write all my JavaScripts in C#, a language that I know inside out, gives me compile-time errors and intelli-sense? I’m sold already! Also because you write your form scripts in C#, you could even share some code between plugins (server) and web resources (client-side JS)!



2. Excellent Script# Xrm.Page Library. You can download for free this library that Gayan Perera has shared with the community. It is as simple as adding a reference assembly to your Script# project and now you have access to the entire Xrm.Page object model from Script# including intellisense for different APIs and enumerations available in the CRM SDK. This is an amazing productivity gainer.



3. Access to mscorlib from your JavaScripts. JavaScripts can get ugly with complex operations. Having a subset of mscorlib available from Script# makes things much simpler. You can have access to useful classes like String, Dictionary, Queue, List<T>, CultureInfo, XmlDocument, XmlHttpRequest , Math and many more, and all this with the intelli-sense we love. Some of these functions would be painful and/or ugly to implement in JavaScript alone. You can also use OOP/C# concepts such as inheritance, static classes and members, Properties, Events and Delegates. This way you can program your JavaScripts in C#and you can use API’s and syntax you are familiar with from .Net. Although it would be possible to do al this in JavaScript, for a C# developer, the usability and efficiency to write Script# does not compare with that of writing JavaScript.



On the other hand, I ran into the following “issues”:

1. Library size. When you make use of any API from a referenced assembly (except Xrm.Page) then you need to include the JavaScript version of that referenced assembly (e.g. mscorlib.js) to be loaded with the CRM form. These libraries can be big and typically, you only need a small subset of the library. It could slow down loading the forms, however, you will have the same problem even without Script# when for example you want to use the JQuery library.

2. You cannot reference any JavaScript code from Script#. If you want to re-use a JavaScript library from a previous project or a library you found online, you cannot reference any JS function from a Script# project. You would have to write “empty wrappers” in C# for each JavaScript function you plan to use, then compile it as a Script# library and then add the reference assembly to your code. It is not too complicated but it adds some pain.

3. AJAX incompatibility. I noticed that my dates in the form were not getting formatted as per the CRM user settings and some other fields and controls had strange issues (e.g. missing the month from the calendar control). After some investigation the culprit ended up being Script#. The reason: mscorlib.js defines classes and functions such as Date.prototype.format and Date.prototype.localeFormat. These functions are also defined in the AJAX framework that the CRM application uses in global.ashx and guess what: They are not implemented the same way. Because the CRM application relies on the MicrosoftAjax.js implementation it will be buggy if it uses the mscorlib.js implementation. Unfortunately, as soon as you add mscorlib.js library to your form, it will overwrite the functions from global.ashx and there is no way to “force” the CRM application to use the functions defined in global.ashx instead. I had to implement an emergency workaround to fix this problem which consisted in adding another JS library to the form, which would define the functions back with the AJAX implementation (something like overwriting the overwrite). Ugly, but it fixed the problem.



In summary, I had a good experience using Script# for CRM, but the issue with AJAX framework became a deal-breaker for me. Other things that I will explore in the future are TypeScript and the new Xrm.Page script template available with the CRM SDK 5.0.13+ which gives you intellisense for your CRM JavaScripts. TypeScript sounds promising since it is simply a superset of JavaScript which allows you to reference other JavaScript libraries; however, it is not C#-based and the syntax is not the same, so it would be yet another language to learn.

Mobile Express: Features and Limitations

Mobile Express for CRM 2011 offers a simplified mobile client that is integrated into the product, however, it has some key limitations. This post seeks to explain with pictures which are the pros/cons, main features and limitations of Mobile Express.

FEATURES

 

Homepage

This is the default “landing page” when opening CRM via mobile client

 

Views

All views on the selected entities will be available in Mobile Express:

 

Forms: Read

You can select which fields are available in the form. You will also be able to navigate to the related (child) entities

 

Forms: Edit

The fields which are available to edit via ME will appear in the form “edit” mode. Note that Field Level Security works in ME the same way as in the web client.

 

 

 

 

LIMITATIONS

 

Nota that this is not a complete list of limitations but represent those that in my experience have been the most significant limitations of Mobile Express.

 

 

1. Lookup fields have no “browse” or “auto-complete”. This is one of the most significant limitations to end users when working with ME. Imagine for example that you need to escalate a case immediately and you need to select the team to which you need to escalate. If you are in the web client, you can easily browse the team and there is also auto-complete feature:

 

 

However, in Mobile Express you will need to type the exact team name and hope you did not make any spelling mistakes because if it does not match exactly with an existing record, the save operation will fail. This limitation applies to all lookup fields. As a work-around you might consider using OptionSet fields in ME instead of lookups (and you might need to build a plugin or some logic to map between optionset values and lookup values on save, but this topic alone can be an entire blog post).

 

 

 

2. No Ribbon = Few buttons. Because there is not ribbon in Mobile Express, the buttons available are very limited and you can certainly not implement custom buttons. Mobile Express supports the following buttons to act on a specific record:
- New (Create new record)
- Edit
- Save
- Delete

So for example, assigning a record to another user/team is challenging from Mobile Express because there is no “Assign” button and you can also not edit the “Owner” fields. A workaround would be to create a custom Lookup field to a user/team and expose this field in the mobile form (call it “Owner”) and then have a plugin that will do the actual assignment whenever this field changes. However, this workaround will also suffer from limitation (1) above.

Custom actions (such as “Escalate”) would have to be implemented as fields instead of as custom buttons (following the plugin pattern described above).

 

 

3. No sorting / filtering. Sorting and filtering which can be easily achieved in the web client as shown in the screenshot below are not possible in ME:

 

 

4. No “select multiple”. When you have a view in ME you cannot select multiple records at once so you cannot bulk-edit, bulk-assign, etc.

 

 

5. Maximum 2 fields per view. Mobile Express views will only display the first 2 fields of the view.

 

 

6. No visualizations. ME does not support charts, dashboards or running reports.

 

 

7. On-demand processes. While automatic workflows will trigger from Mobile Express, you cannot start dialogs or workflows on-demand.

 

 

8. Form limitations. You cannot have custom JavaScript or web resources embedded in the mobile form.

 

 

9. Not available offline.

 

 

 

 

PROS/CONS OF ME VS. OTHER CRM MOBILE CLIENTS

Again, this is not a complete list of pros and cons, so feel free to contribute if I left out something significant.

PROS	CONS
No additional license required.	Limited functionality (see above)
Native application within CRM/xRM.	Not optimized for specific device (same interface for all phones and tablets).
Same support channel as the rest of the CRM/xRM application (Microsoft directly, no third-party vendors).
No additional IT infrastructure required.
Simple to implement, manage and maintain.
Access via web (no need to install applications in mobile devices).
Lowest cost.

CRM 2011: Plug-in assembly does not contain the required types or assembly content cannot be updated

This post explains why you receive this error message when trying to update a plugin assembly in CRM 2011 and how to correct it.
This is another very popular question in the community, why do I get the following error message when updating a plugin assembly?
Plug-in assembly does not contain the required types or assembly content cannot be updated (Error Code -2147204725)
The answer is not very simple as there are a number of conditions that would produce that error message, and these conditions are barely covered in the official documentation from Microsoft. So here are the things you need to check:

1. You are not allowed to update the assembly metadata (strong name)
The new plugin assembly must have the same fully qualified name (same culture, publickeytoken, name and version). You are allowed (and should!) modify the version build and/or revision number but you cannot change the major or minor version. Your assembly version is always in the form <major.minor.build.revision> and you can specify the assembly version in Visual Studio before you build. Note that because the publickeytoken must be the same, then you must use the same key to sign your assembly as you used for the original assembly you are trying to replace.

2. You cannot remove or rename classes which are already registered as plugin types
A plugin type is basically a class which implements IPlugin and they might or might not be registered in CRM as plugins. If your plugin assembly has any plugin types registered under it, then you need to make sure that your new assembly contains those registered types with the same class names. It is important that the signature of each class which is registered as a plugin is not changed in your new assembly (internal and helper classes can always change with no problem, but not the public plugin classes). If you wish to change the class name of a registered plugin (for example you had a plugin called "MyPlugin" and then you changed the name of a class to "MyNewPlugin") you will have to first unregister that plugin type and then you can successfully update the plugin assembly which contains new plugin type and then you’d have to re-register the new plugin type.

3. You cannot change or remove arguments of custom workflow activities
If your plugin assembly contains custom workflow activities which are registered and which have In or Out arguments, then you are not allowed to update the assembly if you are making changes to the custom workflow activity arguments. Removing or changing the datatype or the name of the arguments is not allowed; however, adding new arguments is OK (not recommended though). If you would like to change the arguments then you’d have to unregister the custom workflow activity from the system, update the new assembly and then re-register the custom workflow activity with the updated arguments. You can find more details about how workflows behave with different version of the plugin assembly: http://gonzaloruizcrm.blogspot.ca/2011/08/assembly-versioning-in-crm-2011.html

If you absolutely need to make a change which is not allowed in the list above, then you would need to either unregister the old assembly and re-register everything back or you can increase the major/minor version of your assembly and then register it as a new assembly (but you cannot update the existing one). Also keep in mind that in general it is a good practice to always change the build/revision number before you update an assembly. That way, the new assembly will be used immediately without having to run iisreset or restart CRM services.
If you find other conditions that cause this error message you can comment on this post. Also, I