I have a persistence annotated POJO for a domain class and all works great, but I can'f figure out how to make one of the fields transient. The examples I've seen and the field to transients right in the class. Can this be done in the controller instead? Thanks.
Turns out I had to annotate the field in POJO itself:
#Transient
String someString;
// getters, setters, etc. omitted for brevity
Related
Is it possible to trim a string value before it is set against a bean property of type string in the destination bean?
Dozer offers such a facility through its mapping configuration for example,
<configuration>
<trim-strings>true</trim-strings>
</configuration>
Also see Dozer Global Configuration
With MapStruct 1.0.0.Final I can achieve this through Expressions or Before/After Mapping customization.
But wanted to know if there is a better way to handle such use cases.
Thanks in advance.
It appears MapStruct in its current form does not support this.
However one can achieve this effect with custom mapper methods, for example implement a class with a method that trims a String argument passed to it and then reference this class in the use attribute of the #Mapper annotation.
More at Invoking other mappers
If you require fine gained access control you could use
Selection based on Qualifiers
I was made aware of these approaches in response to a question I posted in mapstruct Google group
Example from #venkat-srinivasan answer's:
public class StringTrimmer {
public String trimString(String value) {
return value.trim();
}
}
and then in your mapper interface or class:
#Mapper(uses = StringTrimmer.class)
public interface MyMapper {
I have a class with the following structure.
Is it "legal" in Dagger to #Inject beans in field variables and constructors at the same time, as I have done below? If no - I have a MyActivityModule and MyApplicationModule, how can I get the dependencies from MyApplicationModule and add them to the constructor I use in the provideWhatEvery in the MyActivityModule ?
#Inject SmsFormatter mSmsFormatter;
#Inject SmsGuardiansUtils smsGuardiansUtils;
#Inject BipperMediaPlayer bipperMediaPlayer;
#Inject MixPanelUtils mMixpanelUtils;
#Inject
public ImHereController(View view, Context context, AlarmModel alarmModel, ActionBarListener actionBarListener,
FragmentController fragmentController){
super(view, context, alarmModel, actionBarListener, fragmentController);
}
You can inject fields and constructors as you have. The constructor parameters will be resolved first and injected upon construction, and after that the fields will be injected.
The other parts of your question are unclear - it doesn't matter whether you add dependencies by field injection or constructor injection - if you wanted to add them all with constructor injection you could.
The only time you must use field injection is where you have an object whose instantiation you cannot control, and therefore dagger cannot itself instantiate (like Activity and Application subtypes.)
All that said, I would not use both without some compelling reason - constructor injection is more semantically clear and you can make the instance variables final. Alternately, field injection is more terse, and possibly more readable in cases. I would pick one, and not both.
I have a class defined as:
class Person {
public int age;
public String firstName;
}
Note that I use camel case for the field names. Also, I know that I could have generated getters and setters but I tend to not do that for simple domain objects.
When I deserialize a JSON or XML response in my REST API, it should spit out:
<Person><Age>11</Age><FirstName>Johnson</FirstName></Person>
You will notice that the first letter is upper-cased.
I could use, for example, #JsonPoperty("FirstName") on my POJO to get the output the way I need it, but this doesn't scale when there are too many fields. I'd like to use a custom property naming strategy (as described in How To Use Property Naming Strategy In Jackson). But instead of configuring an ObjectMapper, I was wondering if its possible to specify a naming strategy using annotations?
Thanks
I am very new to GWT.
I am using ext-gwt widgets.
I found many places in my office code containing like,
class A extends BaseModel{
private UserAccountDetailsDto userAccountDetailsDto = null;
//SETTER & GETTER IN BASEMODEL WAY
}
Also, the DTO reference is unused.
public class UserAccountDetailsDto implements Serializable{
private Long userId=null;
private String userName=null;
private String userAccount=null;
private String userPermissions=null;
//NORMAL SETTER & GETTER
}
Now, I am able to get the result from GWT Server side Code and things Work fine, but when I comment the DTO reference inside the class A, I am not getting any Result.
Please explain me the need of that.
Thanks
Well the problem is in implementation of GXT BaseModel and GWT-RPC serialization.
BaseModel is based around special GXT map, RpcMap. This map has defined special serialization rules, which let's avoid RPC type explosion, but as side effect, only some simple types stored in map will be serialized. E.g. you can put any type inside the map, but if you serialize/deserialize it, only values of type Integer, String ,Double,Byte, Float and Short (and arrays of this types) will be present. So the meaning behind putting reference to the DTO inside BaseModel, is to tell GWT-RPC that this type is also have to be serialized.
Detailed explanation
Basically GWT-RPC works like this:
When you define an interface for service, GWT-RPC analyzes all the classes used in parameters/ return type, to create serializers/deserializers. If you return something like Map<Object,Object> from your service, GWT-RPC will have to create a serializer for each class which implements Map and Serializable interfaces, but also it will generate serializers for each class which implements Serializable. In the end it is quite a bad situation, because the size of your compiled js file will be much biggger. This situation is called GWT-RPC type explosion.
So, in the BaseModel, all values are stored in RpcMap. And RpcMap has custom written serializer (RpcMap_CustomFieldSerializer you can see it's code if you interested how to create such things), so it doesn't cause the problem described above. But since it has custom serializer GWT dosn't know which custom class have been put inside RpcMap, and it doesn't generate serializers for them. So when you put some field into your BaseModel class, gwt knows that it might need to be able to serialize this class, so it will generate all the required stuff for this class.
Porting GXT2 Application code using BaseModel to GXT3 Model is uphill task. It would be more or less completely rewrite on model side with ModelProviders from GXT3 providing some flexibility. Any code that relies on Model's events, store, record etc are in for a rewrite.
I have a method on a stateless session bean which creates a new instance of an entity and persists it. You might normally use new MyEntity() to create the object but I would like injection to populate some of the properties of the entity for me.
I got partial success using
#Inject
#New
private MyEntity myNewEntity;
in the session bean and then using that instance in my method.
The problem I have now is that the second time the method is called, myNewEntity isn't a new object, its the same object as the one created the first time. As a result I'm getting
com.mysql.jdbc.exceptions.jdbc4.MySQLIntegrityConstraintViolationException: Duplicate entry '9' for key 'PRIMARY'
Or at least that's why I think I'm getting this exception. Certainly if I use new MyEntity() I don't get the exception but my injection doesn't happen.
Am I on the wrong track? How can I create a new local entity object while enabling injection?
Any help would be great!
First of all - I have serious doubts that it's a good idea to use CDI to control the lifecycle of a Entity. See this quote from the documentation (here):
According to this definition, JPA
entities are technically managed
beans. However, entities have their
own special lifecycle, state and
identity model and are usually
instantiated by JPA or using new.
Therefore we don't recommend directly
injecting an entity class. We
especially recommend against assigning
a scope other than #Dependent to an
entity class, since JPA is not able to
persist injected CDI proxies.
What you should do to create new instances of entities is adding a layer of indirection, either with #Produces or #Unwraps (Seam Solder, if you need it to be truly stateless), and thereby making sure that you code explicitly calls new.
I think I have a working solution now which seems okay, though I'm not quite sure why it works so I welcome your feedback on a better solution. I am now injecting a DAO-style bean into my stateless session bean:
#Stateless
public class PhoneService {
#Inject
protected ProblemReports problemReports;
An injecting my entity into the ProblemReports bean:
public class ProblemReports {
#Inject
#New
private ProblemReport newProblemReport;
I assume that ProblemReports defaults to #Dependant scope which as I understand it should be the same as the stateless session bean which containts it. I could understand this if the scope of ProblemReports was shorter, causing a new instance of ProblemReport to be created when the new ProblemReports is created; but it isn't.
Is this just an example of EJB and CDI not playing well together?