I found a code from Spring forums that seems to be an interesting way to implement wizardForms in Spring 3:
#RequestMapping(method = RequestMethod.POST)
public ModelAndView processSubmit(
#ModelAttribute("pet") Pet pet,
SessionStatus status) {
if (pet.getFieldOne() == null) {
//return the form that will set field one's value
return new ModelAndView( ... );
} else if (pet.getFieldTwo() == null) {
//return the form that will set field two's value
return new ModelAndView( ... );
} //and so on for all the other field that need to be set...
...
else {
//once the object has all necessary fields
//set and validated, then do what needs
//to be done to finish. Store object, end
//session, and return your success view.
this.clinic.storePet(pet);
status.setComplete();
return new ModelAndView( ... );
}
}
Can anyone tell me what the storing here means, and is this a good way?
If by "storing" you mean this.clinic.storePet(pet);, it's an action of saving the complete object in your database when wizard is finished, so that it's completely unrelated to wizard implementation.
The approach itself is a standard way to implement wizard forms in Spring 3 that replaces the deprectated AbstractWizardFormController.
Note that it also requires #SessionAttribute("pet") as a class-level annotation. This annotation makes Spring to store the corresponding model attribute in the session between requests, so that each form submission sets fields of the same object. When all fields are set and wizard is finished, object is saved to the database, and removed from the session by status.setComplete();.
Related
Using ASP.NET Core 2.2 with EF Core, I have followed various guides in trying to implement the automatic creation of date/time values when creating either a new record or editing/updating an existing one.
The current result is when i initially create a new record, the CreatedDate & UpdatedDate column will be populated with the current date/time.
However first time I edit this same record, the UpdatedDate column is then given a new date/time value (as expected) BUT for some reason, EF Core is wiping out the value of the original CreatedDate which results in SQL assigning a default value.
Required result I need as follows:
Step 1: New row created, both CreatedDate & UpdatedDate column is given a date/time value (this already works)
Step 2: When editing and saving an existing row, I want EF Core to update the UpdatedDate column with the updated date/time only, BUT leave the other CreatedDate column unmodified with the original creation date.
I'm using EF Core code first, and do no want to go down the fluent API route.
One of the guides i was partially following is https://www.entityframeworktutorial.net/faq/set-created-and-modified-date-in-efcore.aspx but neither this or other solutions I've tried is giving the result I am after.
Baseclass:
public class BaseEntity
{
public DateTime? CreatedDate { get; set; }
public DateTime? UpdatedDate { get; set; }
}
DbContext Class:
public override Task<int> SaveChangesAsync(bool acceptAllChangesOnSuccess, CancellationToken cancellationToken = default(CancellationToken))
{
var entries = ChangeTracker.Entries().Where(E => E.State == EntityState.Added || E.State == EntityState.Modified).ToList();
foreach (var entityEntry in entries)
{
if (entityEntry.State == EntityState.Modified)
{
entityEntry.Property("UpdatedDate").CurrentValue = DateTime.Now;
}
else if (entityEntry.State == EntityState.Added)
{
entityEntry.Property("CreatedDate").CurrentValue = DateTime.Now;
entityEntry.Property("UpdatedDate").CurrentValue = DateTime.Now;
}
}
return base.SaveChangesAsync(acceptAllChangesOnSuccess, cancellationToken);
}
UPDATE FOLLOWING ADVICE FROM STEVE IN COMMENTS BELOW
I spent a bit more time debugging today, turns out the methods I posted above are appear to be functioning as expected i.e. when editing an existing row and saving it, only the entityEntry.State == EntityState.Modified IF statement is being called. So what I'm finding is that after saving the entity, the CreatedDate column is being overwitten with a Null value, I can see this by watching the SQL explorer after a refresh. I believe the issue is along the lines of what Steve mentions below "If it is #null then this might also explain the behavior in that it is not being loaded with the entity for whatever reason."
But i'm a little lost in tracing where this CreatedDate value is being dropped somewhere through edit/save process.
Image below shows the result at the point just before the entity is saved following an update. In the debugger I'm not quite sure where to find the entry of the CreatedDate to see what value is held at this step, but it appears to be missing from the debugger list so wandering whether somehow it doesn't know about the existence of this field when saving.
Below is the method I have in my form 'Edit' Razor page model class:
public class EditModel : PageModel
{
private readonly MyProject.Data.ApplicationDbContext _context;
public EditModel(MyProject.Data.ApplicationDbContext context)
{
_context = context;
}
[BindProperty]
public RuleParameters RuleParameters { get; set; }
public async Task<IActionResult> OnGetAsync(int? id)
{
if (id == null)
{
return NotFound();
}
RuleParameters = await _context.RuleParameters
.Include(r => r.SystemMapping).FirstOrDefaultAsync(m => m.ID == id);
if (RuleParameters == null)
{
return NotFound();
}
ViewData["SystemMappingID"] = new SelectList(_context.SystemMapping, "ID", "MappingName");
return Page();
}
public async Task<IActionResult> OnPostAsync()
{
if (!ModelState.IsValid)
{
return Page();
}
_context.Attach(RuleParameters).State = EntityState.Modified;
try
{
await _context.SaveChangesAsync();
}
catch (DbUpdateConcurrencyException)
{
if (!RuleParametersExists(RuleParameters.ID))
{
return NotFound();
}
else
{
throw;
}
}
return RedirectToPage("./Index");
}
private bool RuleParametersExists(int id)
{
return _context.RuleParameters.Any(e => e.ID == id);
}
}
Possibly one of the reasons for this issue is the fact that I have not included the CreatedDate field in my Edit Razor Page form, so when I update the entity which in turn will run the PostAsync method server side, there is no value stored for the CreatedDate field and therefore nothing in the bag by the tine the savechangesasync method is called in my DbContext Class. But I also didn't think this was necessary? otherwise I'd struggle to see what value there is in the this process of using an inherited BaseEntity class i.e. not having to manually add the CreatedDate & UpdatedDate attribute to every model class where I want to use it...
It may be easier to just give your BaseEntity a constructor:
public BaseEntity()
{
UpdatedDate = DateTime.Now;
CreatedDate = CreatedDate ?? UpdatedDate;
}
Then you can have your DbContext override SaveChangesAsync like:
public override Task<int> SaveChangesAsync(
bool acceptAllChangesOnSuccess,
CancellationToken token = default)
{
foreach (var entity in ChangeTracker
.Entries()
.Where(x => x.Entity is BaseEntity && x.State == EntityState.Modified)
.Select(x => x.Entity)
.Cast<BaseEntity>())
{
entity.UpdatedDate = DateTime.Now;
}
return base.SaveChangesAsync(acceptAllChangesOnSuccess, token);
}
Possibly one of the reasons for this issue is the fact that I have not included the CreatedDate field in my Edit Razor Page form, so when I update the entity which in turn will run the PostAsync method server side, there is no value stored for the CreatedDate field and therefore nothing in the bag by the tine the savechangesasync method is called in my DbContext Class.
That's true.Your post data does not contains the original CreatedDate,so when save to database, it is null and could not know what the exact value unless you assign it before saving.It is necessary.
You could just add below code in your razor form.
<input type="hidden" asp-for="CreatedDate" />
Update:
To overcome it in server-side,you could assign data manually:
public async Task<IActionResult> OnPostAsync()
{
RuleParameters originalData = await _context.RuleParameters.FirstOrDefaultAsync(m => m.ID == RuleParameters.ID);
RuleParameters.CreatedDate = originalData.CreatedDate;
_context.Attach(RuleParameters).State = EntityState.Modified;
await _context.SaveChangesAsync();
}
I don't suspect EF is doing this, but rather your database, or you're inadvertently inserting records instead of updating them.
A simple test: Put break-points in your SaveChangesAsnc method within both the Modified and Added handlers and then run a unit test that loads an entity, edits it, and saves. Which breakpoint is hit? If the behavior seems to be normal with a simple unit test, repeat again with your code.
If the Modified breakpoint is hit, and only the Modified handler is hit then check the state of the CreatedDate value in the entity modified. Does it still reflect the original CreatedDate? If yes, then it would appear that something in your schema will be overwriting it on save. If no then you have a bug in your code that has caused it to update. If it is #null then this might also explain the behaviour in that it is not being loaded with the entity for whatever reason. Check that the property has not been configured as something like a Computed property.
If the Added breakpoint is hit at all, then this would point at a scenario where you're dealing with a detached entity, such as an entity that was read from a different DB Context and being associated to another entity in the current DB Context and saved as a byproduct. When a DbContext encounters an entity that was loaded and disassociated with a different DbContext, it will treat that entity as a completely new entity and insert a new record. The biggest single culprit for this is invariably MVC code where people pass entities to/from views. Entity references are loaded in one request, serialized to the view, and then passed back on another request. Devs assume they are receiving an entity that they can just associate to a new entity and save, but the Context of this request doesn't know about that entity, and that "entity" isn't actually an entity, it is now a POCO shell of data that the serializer created. It's no different to you newing up a new class and populating fields. EF won't know the difference. The result of this is you will trip the Added condition for your entity, and after completion you will have a duplicate record. (with different PK if EF is configured to treat PKs as Identity)
So an example is an Order screen: When presenting a screen to create a new order I may have loaded the Customer and passed that to the view to display customer information and will want to associate to the new order:
var customer = context.Customers.Single(x => x.CustomerId == 15);
var newOrder = new Order { Customer = customer };
return View(newOrder);
This looks innocent enough. When we go to save the new order after setting their details:
public ActionResult Save(Order newOrder)
{
context.Orders.Add(newOrder);
newOrder.Customer.Orders.Add(newOrder);
context.SaveChanges();
// ...
}
newOrder had a reference to Customer #14, so all looks good. We're even associating the new order to the customer's order collection. We might even want to have updated fields on the customer record to reflect a change to the Modified date. However, newOrder in this case, and all associated data including .Customer are plain 'ol C# objects at this point. We've added the new order to the Context, but as far as the context is concerned, the Customer referenced is also a new record. It will ignore the Customer ID if that is set as an Identity column and it will save a brand new Customer record (ID #15 for example) with all of the same details as Customer ID 14 and associate that to the new order. It can be subtle and easy to miss until you start querying Customers and spotting duplicate looking rows.
If you are passing entities to/from views, I'd be very wary of this gotcha. Attaching and setting modified state is one option, but that involves trusting that the data has not been tampered with. As a general rule, calls to update entities should never pass entities & attach them, but rather re-load those entities, validate row version, validate the data coming in, and only copy across fields you expect should ever be modified before saving the entity associated to the DbContext.
Hopefully that gives you a few ideas on things to check to get to the bottom of the issue.
Hy,
I'm trying to call my action with allways a fixed Uid (configured by TS) so I could put a plugin on my page to register for a specific Event. And don't have to go over a Event List click the Event click register.
I tried the following which did not work out:
public function newAction(
\XYZ\xyz\Domain\Model\Registration $newRegistration = NULL,
\XYZ\xyz\Domain\Model\Event $event = 'DD8B2164290B40DA240D843095A29904'
)
The next didn't one work either!
public function newAction(
\XYZ\xyz\Domain\Model\Registration $newRegistration = NULL,
\XYZ\xyz\Domain\Model\Event $event = Null
) {
$myinstance = \TYPO3\CMS\Core\Utility\GeneralUtility::makeInstance(
'XYZ\\xyz\\Domain\\Model\\Event'
);
$event = $myinstance->findByUid('DD8B2164290B40DA240D843095A29904');
.......
}
So I was woundering is there a way to give my fixed Uid to the action?
In TYPO3 calling Extbase actions is done in the routing and dispatching components - to pass anything from the outside that is different from a numeric uid value a custom property TypeConverter would have to be implemented that transforms a particular string pattern into a value domain object of type Event.
However, there's a simpler approach by using configuration:
1) Provide configuration in TypoScript
Extbase uses a strong naming convention based on the extension name and optionally the plugin name. Thus, either tx_myextension or tx_myextension_someplugin can be used - latter is more specific for for according somePlugin. Besides that settings are automatically forwarded and provided in an Extbase controller context - accessible by $this->settings.
plugin.tx_xyz {
settings {
newActionEventIdentifier = DD8B2164290B40DA240D843095A29904
}
}
2) Retrieve data via repository
\XYZ\xyz\Domain\Repository\EventRepository
Use a dedicated EventRepository::findByIdentifier(string) method to retrieve the data. The property names are just assumptions since there are no explicit mentions how exactly the event data is persisted and whether it is persisted in a relational DBMS at all.
<?php
namespace XYZ\xyz\Domain\Repository;
class EventRepository
{
public function findByIdentifier($identifier)
{
$query = $this->createQuery();
$query->matching(
$query->equals('event_id', $identifier)
);
return $query->execute();
}
}
3) Putting all together in the according controller
The $event property was removed from the action since that entity is pre-defined and cannot be submitted from the outside (and to support the string to Event entity transformation a custom TypeConverter would be required as mentioned earlier).
public function newAction(
\XYZ\xyz\Domain\Model\Registration $newRegistration = null
) {
$event = $this->eventRepository->findByIdentifier(
$this->settings['newActionEventIdentifier']
);
if ($event === null) {
throw new \RuntimeException('No event found', 1522070079);
}
// the regular controller tasks
$this->view->assign(...);
}
I want to create input forms which validate user input and prevent the model from being saved with invalid data. I have been using databinding which works up to a point but my implementation is not as intuitive as I would like.
Imagine an input which contains '123' and the value must not be empty. The user deletes the characters one by one until empty. The databinding validator shows an error decoration.
However, if the user saves the form and reloads it, then a '1' is displayed in the field - i.e. the last valid input. The databinding does not transmit the invalid value into the model.
I have a ChangeListener but this is called before the databinding so at that point the invalid state has not been detected.
I would like the error to be displayed in the UI but the model remains valid (this is already so). Also, for as long as the UI contains errors, it should not be possible to save the model.
/**
* Bind a text control to a property in the view model
**/
protected Binding bindText(DataBindingContext ctx, Control control,
Object viewModel, String property, IValidator validator)
{
IObservableValue value = WidgetProperties.text(SWT.Modify).observe(
control);
IObservableValue modelValue = BeanProperties.value(
viewModel.getClass(), property).observe(viewModel);
Binding binding = ctx.bindValue(value, modelValue, getStrategy(validator), null);
binding.getTarget().addChangeListener(listener);
ControlDecorationSupport.create(binding, SWT.TOP | SWT.LEFT);
return binding;
}
private UpdateValueStrategy getStrategy(IValidator validator)
{
if (validator == null)
return null;
UpdateValueStrategy strategy = new UpdateValueStrategy();
strategy.setBeforeSetValidator(validator);
return strategy;
}
private IChangeListener listener = new IChangeListener()
{
#Override
public void handleChange(ChangeEvent event)
{
// notify all form listeners that something has changed
}
};
/**
* Called by form owner to check if the form contains valid data e.g. before saving
**/
public boolean isValid()
{
System.out.println("isValid");
for (Object o : getDataContext().getValidationStatusProviders())
{
ValidationStatusProvider vsp = (ValidationStatusProvider) o;
IStatus status = (IStatus)vsp.getValidationStatus()
.getValue();
if (status.matches(IStatus.ERROR))
return false;
}
return true;
}
Your best bet is to steer clear of ChangeListeners - as you've discovered, their order of execution is either undefined or just not helpful in this case.
Instead, you want to stick with the 'observable' as opposed to 'listener' model for as long as possible. As already mentioned, create an AggregateValidationStatus to listen to the overall state of the DataBindingContext, which has a similar effect to your existing code.
Then you can either listen directly to that (as below) to affect the save ability, or you could even bind it to another bean.
IObservableValue statusValue = new AggregateValidationStatus(dbc, AggregateValidationStatus. MAX_SEVERITY);
statusValue.addListener(new IValueChangeListener() {
handleValueChange(ValueChangeEvent event) {
// change ability to save here...
}
});
You can use AggregateValidationStatus to observe the aggregate validation status:
IObservableValue value = new AggregateValidationStatus(bindContext.getBindings(),
AggregateValidationStatus.MAX_SEVERITY);
You can bind this to something which accepts an IStatus parameter and it will be called each time the validation status changes.
We need users to be able to save a draft of their form in MVC4.
We already implemented a "Save Draft" button which disables the
client side validation entirely and submits the form.
Now on the server side the plan is to ignore any form fields which
have validation errors. For example, if a zip code field has "abc"
then we'll ignore that field and save the rest. If any field are
required but not provided those will also be ignored.
We're trying to figure out an elegant way to do this in MVC4 via
IModelBinder.
public class CustomModelBinder : DefaultModelBinder
{
public override object BindModel(ControllerContext controllerContext, ModelBindingContext bindingContext)
{
var result = base.BindModel(controllerContext, bindingContext);
if (!controllerContext.Controller.ViewData.ModelState.IsValid)
{
foreach( var state in controllerContext.Controller.ViewData.ModelState)
{
if( state.Value.Errors.Count > 0) {
var fieldName = state.Key;
// Clear original data field that has the error.
}
}
// Now redo the binding without the fields that presented a problem.
result = base.BindModel(controllerContext, bindingContext);
}
return result;
}
}
So the problem above is how to clear the specific field data so that it can re-run
the data binding and generate the model without the fields that cause validation errors?
I have two pick-list in Car details entity. I'm setting the Model (cir_model) Picklist value with from the input parameter (that is CrmNumber) of Custom Workflow activity and it's working as expected, and the second pick-list Marque (cir_marque) will be set logically using the Model pick-list.
Logic should be if Model is set to 'Ac Ace' then Marque should be set to 'Ac'. Take value 'Ac' using Split() from the string 'Ac Ace'.
Normally in C# this can be done easily but in CRM 4.0 how this can be achieve (How I'll set 'Ac' to Marque)
public static DependencyProperty modelProperty = DependencyProperty.Register("model",
typeof(int), typeof(CreateCardetails));
[CrmInput("Model")]
public int model
{
get
{
return (int)base.GetValue(modelProperty);
}
set
{
base.SetValue(modelProperty, value);
}
}
public static DependencyProperty ContactProperty =
DependencyProperty.Register("Contact", typeof(Lookup), typeof(CreateCardetails));
[CrmInput("Contact ID")]
[CrmReferenceTarget("contact")]
public Lookup Contact
{
get
{
return (Lookup)base.GetValue(ContactProperty);
}
set
{
base.SetValue(ContactProperty, value);
}
}
protected override ActivityExecutionStatus Execute(ActivityExecutionContext
executionContext)
{
//Create an car details record which will be linked to the contact record
DynamicEntity cardetails = new DynamicEntity("cir_cardetails");
cardetails["cir_carsdetailsid"] = Contact;
//Setting the picklist value of Model
Picklist modelPickList = new Picklist();
modelPickList.Value = model.Value;
cardetails.Properties.Add(new PicklistProperty("cir_model",modelPickList));
/*
Here the logic should be done for setting Marque (cir_model) value
Picklist marquePickList = new Picklist();
marquePickList.Value = ???
cardetails.Properties.Add(new PicklistProperty("cir_marque",marquePickList));
*/
//Creating the car details record
Guid carkey = crmService.Create(cardetails);
}
How we can set the Marque value logically, I have left the code blank for this like below
/*
Here the logic should be done for setting Marque (cir_marque) value
Picklist marquePickList = new Picklist();
marquePickList.Value = ???
cardetails.Properties.Add(new PicklistProperty("cir_marque",marquePickList));
*/
Please arrange to help me out on this, all suggestions are welcome.
There is no language CRM 4.0, in CRM 4.0 you code in c#. The only thing that change is the way you work with new types.
In Workflow you don't work with controls, you work with entities and the related attributes. So you "just" need to get the attribute cir_model, do a subtring and find the available options in Marque and set the corrected value. Check this sample from SDK.
You can use JavaScript or C# (Plug-In, Workflow) to accomplish this. There are some considerations to think of when choosing which approach to use.
If you want the user to be able to see the result in real time (when they select) then you can use JavaScript.
If you don't care for the user to see the result, or there is data coming in from an outside source (not the user form), then think about using a plugin.
I don't think you should have to use a WF to do this, plugins are just as easy to write and will happen instantaneously instead of waiting for the async process to complete.