I have a problem with an automapper formatter being called twice and using the value it resolved to on the first run as the source on the second run and causing bad results.
My automapper profile contains the following global string formatter
Mapper.ForSourceType<string>().AddFormatter<StandardStringFormatter>();
I have another formatter for a specific string property. DhcpEnabledFlagFormatter turns a one character code into a meaningful string.
Mapper.CreateMap<NetInterface, NetInterfaceList>()
.ForMember(x => x.DhcpEnabledFlag,
o => o.AddFormatter<DhcpEnabledFlagFormatter>());
DhcpEnabledFlagFormatter gets called twice. On the second pass, it uses the meaningful string it resolved to as the source value and then always returns "None" which is default response of the Formatter.
If I remove the global StandardStringFormatter from my automapper profile, the problem goes away and DhcpEnabledFlagFormatter is only called once.
I know I can solve problem by turning DhcpEnabledFlagFormatter into a custom resolver, but I am curious if there is another option or if this is a known bug.
Related
In my AutoMapper scenario, I have a string property on my source type but a Uri property on the destination. Occasionally, this conversion is expected to fail when the source data is bad:
UriFormatException: Invalid URI: The hostname could not be parsed.
at Uri.ctor(String uriString)
When this happens, I would like AutoMapper to set the problematic property to null on the destination & continue on with the process of mapping the other elements. Known solutions to this are:
Writing a custom converter (with try/catch) just for that element
Using BeforeMap() as described here
But I feel that if I have to take those extra steps, AutoMapper hasn't fulfilled its promise as a time-saving tool. Isn't there some global 'continue-on-error' flag that I can set? Note that I am less concerned about Uris in particular than I am about the lack of configurable error handling.
No special setup needed to repro this--just a vanilla CreateMap() call that registers two classes with differently typed properties.
The asp-items Razor "TagHelper" will add an <option> to a <select> for each value in the SelectList. I want to modify each of those children.
Specifically I want to disable some of them (i.e. add disabled="disabled").
Even more specifically I want to dynamically disable some of them; I'm using angular so I could ng-disabled="{dynamic_boolean_which_determines_disabled}". This means the option could be disabled at first, but after user makes a change, the option could be disabled (without page reload). Angular should take care of this; I think Angular and TagHelpers should work together in theory...
I expected:
I could somehow access an IEnumerable of the children <option> tags that would be created (i.e. one for each item in the SelectList), iterate the children tags, and SetAttribute("disabled") or SetAttribute("ng-disabled")...
I tried:
Creating my own TagHelper which targets the select[asp-items], and tries to GetChildContentAsync() and/or SetContent to reach an IEnumerable <option> tags and iterate them and process each, but I think this will only let me modify the entire InnerHtml as a string; feels hacky to do a String.replace, but I could do it if that's my only option? i.e. ChildrenContent.Replace("<option", "<option disabled=\"...\"")
Creating my own TagHelper which targets the option elements that are children of the select[asp-items], so I can individually process each. This works, but not on the dynamically-added <option> created by asp-items, it only works on "literal" <option> tags that I actually put into my cshtml markup.
I think this'll work but not ideal:
As I said above, I think I can get the result of TagHelper's dynamic asp-items <option></option> <option></option>, as a string, and do a string replace, but I prefer not to work with strings directly...
I suspect (I haven't tried it) that I could just do the work of asp-items myself; i.e. custom-items. But then I'm recreating the wheel by re-doing the work which asp-items could've done for me?
So I hadn't yet read the "AutoLinkHttpTagHelper" in the example which uses string replacement (specifically RegEx replace) to replace every occurrence of a URL, with an <a> pointed at that URL. The cases are slightly different*, but...
Anyway, here's my solution once I learned to stop worrying and love the string modification:
[HtmlTargetElement("select", Attributes = "asp-items")]
public class AspItemsNgDisabledTagHelper : SelectTagHelper
{
//Need it to process *after* the SelectTagHelper
public override int Order { get; } = int.MaxValue;
//https://learn.microsoft.com/en-us/aspnet/core/mvc/views/tag-helpers/authoring#ProcessAsync
public AspItemsNgDisabledTagHelper(IHtmlGenerator gen) : base(gen) {}
public override void Process(TagHelperContext context, TagHelperOutput output)
{
//Notice I'm getting the PostContent;
//SelectTagHelper leaves its literal content (i.e. in your CSHTML, if there is any) alone ; that's Content
//it only **appends** new options specified; that's PostContent
//Makes sense, but I still wasn't expecting it
var generated_options = output.PostContent.GetContent();
//Note you do NOT need to extend SelectTagHelper as I've done here
//I only did it to take advantage of the asp-for property, to get its Name, so I could pass that to the angular function
var select_for = this.For.Name;
//The heart of the processing is a Regex.Replace, just like
//their example https://learn.microsoft.com/en-us/aspnet/core/mvc/views/tag-helpers/authoring#inspecting-and-retrieving-child-content
var ng_disabled_generated_options = Regex.Replace(
generated_options,
"<option value=\"(\\w+)\">",
$"<option value=\"$1\" ng-disabled=\"is_disabled('{select_for}', '$1')\">");
//Finally, you Set your modified Content
output.PostContent.SetHtmlContent(ng_disabled_generated_options);
}
}
Few learning opportunities:
Was thinking I'd find AspForTagHelper and AspItemsTagHelper, (angular background suggested that the corresponding attributes; asp-for and asp-items, would be separate "directives" aka TagHelper).
In fact, TagHelper "matching" focuses on the element name (unlike angular which can match element name... attribute... class... CSS selector)
Therefore I found what I was looking for in SelectTagHelper, which has For and Items as properties. Makes sense.
As I said above, I extend SelectTagHelper, but that's not necessary to answer my original question. It's only necessary if you want access to the this.For.Name as I've done, but there may even be a way around that (i.e. re-bind its own For property here?)
I got on a distraction thinking I would need to override the SelectTagHelper's behavior to achieve my goals; i.e. Object-Oriented Thinking. In fact, even if I did extend SelectTagHelper, that doesn't stop a separate instance of the base SelectTagHelper from matching and processing the element. In other words, element processing happens in a pipeline.
This explains why extending and calling base.Process(), will result in Select doing its job twice; once when your instance matches, and again when the base instance matched.
(I suppose could've prevented SelectTagHelper from matching by creating a new element name like <asp-items-select>? But, not necessary... I just avoid calling base.Process(). So unless that's a bad practice...)
*Different in this way:
They want to create a tag where none exists, whereas I want to add an attribute a tag which is already there; i.e. the <option>
Though the <option> "tag" is generated by the SelectTagHelper in its PostContent (was expecting to find it in Content), and I don't think tags-generated-in-strings-by-content-mods can be matched with their corresponding TagHelper -- so maybe we really are the same in that we're just dealing with plain old strings
Their "data" aka "model" is implied in the text itself; they find a URL and that URL string becomes a unit of meaning they use. In my case, there is an explicit class for Modeling; the SelectList (<select>) which consists of some SelectListItem (<option>) -- but that class doesn't help me either.
That class only gives me attributes like public bool Disabled (remember, this isn't sufficient for me because the value of disabled could change to true or false within browser; i.e. client-side only), and public SelectListGroup Group -- certainly nothing as nonstandard as ng-disabled, nor a "catch-all" property like Attributes which could let me put arbitrary attributes (ng-disabled or anything else) in there.
I need to run some code whenever a property value is retrieved, so naturally it made sense to define the getProperty method in my class. This method will get automatically called whenever a property value is retrieved. Here's roughly what I have in my class:
class MyClass
{
def getProperty(String name)
{
// Run some code ...
return this.#"${name}"
}
}
The problem with the above method occurs when someone tries to make the following call somewhere:
MyClass.class
This call ends up in the getProperty method looking for a property named "class", however, there is not actual property named "class" so we get a MissingFieldException.
What would be the correct way to implement running code whenever a property value is retrieved and deal with these kind of situtations.
Best is not to have a getProperty method if not needed. If you need one and you want to fall back on standard Groovy logic, then you can use return getMetaClass().getProperty(this, property), as can be found in GroovyObjectSupport. This will cover more than just fields.
This seems to be a common problem with this method. Map has the same issue. The developers of groovy got around the problem with Map by saying you need to use getClass() directly.
Trying to exclude properties from a model from being included during serialization.
I am using the following syntax:
JsConfig<MyTestClass>.ExcludePropertyNames = new[] { "ShortDescription" };
Just after that I have the following:
return (from o in __someProvider.GetAll() select (new
{
o.Name,
o.ShortDescription
o.InsertDate
}).TranslateTo<MyTestClass>()).ToList()
However once result is returned from the method, it still contains "ShortDescription" field in the Json. Am I doing something wrong?
JsConfig<T>.ExcludePropertyNames appears to be checked only once for each type, in a static constructor for TypeConfig<T>. Thus, if you are configuring ExcludePropertyNames in your service class, just before returning your response, it might be too late -- the TypeConfig properties may already be set up and cached for MyTestClass. I was able to reproduce this.
A more reliable alternative is to move all of your JsConfig<T> configuration to your AppHost setup code.
If you really do need to do this in your service class, e.g. if you are only conditionally excluding property names, then an alternative approach would be to ensure that JsConfig.IncludeNullValues is false (I believe it is by default) and in your service code set ShortDescription to null when appropriate.
After spending a year working with the Microsoft.Xrm.Sdk namespace, I just discovered yesterday the Entity.FormattedValues property contains the text value for Entity specific (ie Local) Option Set texts.
The reason I didn't discover it before, is there is no early bound method of getting the value. i.e. entity.new_myOptionSet is of type OptionSetValue which only contains the int value. You have to call entity.FormattedValues["new_myoptionset"] to get the string text value of the OptionSetValue.
Therefore, I'd like to get the crmsrvcutil to auto-generate a text property for local option sets. i.e. Along with Entity.new_myOptionSet being generated as it currently does, Entity.new_myOptionSetText would be generated as well.
I've looked into the Microsoft.Crm.Services.Utility.ICodeGenerationService, but that looks like it is mostly for specifying what CodeGenerationType something should be...
Is there a way supported way using CrmServiceUtil to add these properties, or am I better off writing a custom app that I can run that can generate these properties as a partial class to the auto-generated ones?
Edit - Example of the code that I would like to be generated
Currently, whenever I need to access the text value of a OptionSetValue, I use this code:
var textValue = OptionSetCache.GetText(service, entity, e => e.New_MyOptionSet);
The option set cache will use the entity.LogicalName, and the property expression to determine the name of the option set that I'm asking for. It will then query the SDK using the RetrieveAttriubteRequest, to get a list of the option set int and text values, which it then caches so it doesn't have to hit CRM again. It then looks up the int value of the New_MyOptionSet of the entity and cross references it with the cached list, to get the text value of the OptionSet.
Instead of doing all of that, I can just do this (assuming that the entity has been retrieved from the server, and not just populated client side):
var textValue = entity.FormattedValues["new_myoptionset"];
but the "new_myoptionset" is no longer early bound. I would like the early bound entity classes that gets generated to also generate an extra "Text" property for OptionSetValue properties that calls the above line, so my entity would have this added to it:
public string New_MyOptionSetText {
return this.GetFormattedAttributeValue("new_myoptionset"); // this is a protected method on the Entity class itself...
}
Could you utilize the CrmServiceUtil extension that will generate enums for your OptionSets and then add your new_myOptionSetText property to a partial class that compares the int value to the enums and returns the enum string
Again, I think specifically for this case, getting CrmSvcUtil.exe to generate the code you want is a great idea, but more generally, you can access the property name via reflection using an approach similar to the accepted answer # workarounds for nameof() operator in C#: typesafe databinding.
var textValue = entity.FormattedValues["new_myoptionset"];
// becomes
var textValue = entity.FormattedValues
[
// renamed the class from Nameof to NameOf
NameOf(Xrm.MyEntity).Property(x => x.new_MyOptionSet).ToLower()
];
The latest version of the CRM Early Bound Generator includes a Fields struct that that contains the field names. This allows accessing the FormattedValues to be as simple as this:
var textValue = entity.FormattedValues[MyEntity.Fields.new_MyOptionSet];
You could create a new property via an interface for the CrmSvcUtil, but that's a lot of work for a fairly simple call, and I don't think it justifies creating additional properties.