I am learning factory method in dart. However I encounter a problem, which says
A value of type 'Person?' can't be returned from a function with return type 'Person' because 'Person?' is nullable and 'Person' isn't.
3-1.dart:13
Here is my code:
void main(List<String> args) {
Person a = new Person("zarkli");
a.say("hello");
}
class Person{
final name;
static final Map<String, Person> _cache = new Map<String,Person>();
factory Person(String name){
if (_cache.containsKey(name)){
return _cache[name]; // error
} else {
final person = new Person.newPerson(name);
_cache[name] = person;
return person;
}
}
Person.newPerson(this.name);
void say(String content){
print("$name: $content");
}
}
Some maps allow null as a value. For those maps, a lookup using this operator cannot distinguish between a key not being in the map, and the key being there with a null value.
This is from the get operator documentation for map. When you use containsKey it could be the case that the key exists but its value is null.
You can use a local variable to store _cache[name] and check if it's null or not. Dart then can promote that local variable and not give you any errors. So your factory will look like this:
factory Person(String name) {
final cached = _cache[name];
if (cached != null) {
return cached;
} else {
final person = Person.newPerson(name);
_cache[name] = person;
return person;
}
}
Related
In Dart, is it possible for a constructor to cancel object creation and return a different object instead?
Use case:
Users contains a static map that maps ids to User objects.
When a User is initialized, I want the User constructor to check if User with id is already created, if so: return existing User object, else create a new User object
Example (of-course not working):
class Users {
static const Map<String, User> users = {};
}
class User {
final String id;
final String firstName;
User({required id, required firstName}) {
// If user with id already exists, return that object
if (Users.users.containsKey(id) {
return Users.users[id];
}
// Else, initialize object and save it in Users.users
this.id = id;
this.firstName = firstName;
Users.users[id] = this;
}
}
Question: IS there any way to get the above pseudo code to work?
As mentioned by jamesdlin you should use a factory constructor. Here's what is mentioned in the documentation:
Use the factory keyword when implementing a constructor that doesn’t
always create a new instance of its class.
And in your case this is exactly what you want. Now here's a code sample that does what you want:
Code sample
class Users {
// Also your Map cannot be const if you want to edit it.
static Map<String, User> users = {};
}
class User {
final String id;
final String firstName;
/// Base private constructor required to create the User object.
User._({required this.id, required this.firstName});
/// Factory used to create a new User if the id is available otherwise return the User
/// associated with the id.
factory User({required String id, required String firstName}) {
// If user with id already exists, return that object
if (Users.users.containsKey(id)) {
// Force casting as non nullable as we already checked that the key exists
return Users.users[id]!;
}
// Else, initialize object and save it in Users.users
final newUser = User._(id: id, firstName: firstName);
Users.users[id] = newUser;
return newUser;
}
}
Try the full code on DartPad
You can create a function in class to handle things you want. Here's what you can implement.
class Player {
final String name;
final String color;
Player(this.name, this.color);
Player.fromPlayer(Player another) :
color = another.color,
name = another.name;
}
If this is for caching purposes or you are not creating multiple instances of the Users class, I would suggest using a pattern where static is responsible for a list of class instances. Sometimes this helps to significantly reduce the amount of code:
class User {
static final Map<String, User> users = {};
final String id, firstName;
User._({required this.id, required this.firstName});
factory User({required String id, required String firstName}) => users[id] ??= User._(id: id, firstName: firstName);
}
I'm getting an error like
The expression doesn't evaluate to a function, so it can't be invoked.
Is there any workaround to achieve this I have different classes being returned from a map according to the key.
void main() {
Map<String, Type> map = {'user': User};
Type T = map['user']!;
User a = T(5);
print(a.id);
}
class User {
User(this.id);
final int id;
}
struct Person {
string FirstName;
string LastName;
};
class Builder {
public:
Person Builder::Build() {
Person person;
person.FirstName = "FirstName";
person.LastName = "LastName";
return person;
};
};
When I compile this it gives me the below error:
'return' - structure have objects and cannot be copied.
I just need to create a struct or class object and return it, I don't want to do any copying.
I tried using & and * combinations but didn't work. I tried with a class instead of struct and it didn't work either.
I also tried with class as shown below:
class Person {
public:
string FirstName;
string LastName;
};
class Builder {
public:
Person* Build() {
Person person;
person.FirstName = "FirstName";
person.LastName = "LastName";
return &person;
};
};
int OnInit()
{
Builder builder;
Person* person = builder.Build();
string firstName = person.FirstName;
return(INIT_SUCCEEDED);
}
And it gives me invalid pointer access when accessing person.FirstName in the OnInit() method at runtime.
Found the answer but how to avoid memory leak? how to destruct the object and its pointer after use?
class cPerson {
public:
string FirstName;
string LastName;
};
class cBuilder {
public:
cPerson* Build() {
cPerson* person = new cPerson();
person.FirstName = "firstname";
return person;
};
};
cBuilder builder;
cPerson* person = builder.Build();
string age = person.FirstName;
You can delete the object by delete(person); and if you are unsure the object is not a null, it is better to check if(CheckPointer(object)==POINTER_DYNAMIC)delete(object);
Overall you should have all such objects as variables with corresponding variable names, or keep them all in a collection and destroy the whole collection at end. You may also create global variable of object (before OnInit, not inside any function) and it is to be deleted at end of program.
Regarding the initial question - you cannot have string inside a struct, only primitives
you should NOT delete static objects - you will get "delete invalid pointer"-message from compiler in print area... therefore the check should be done as Daniel Kniaz answered... but as so as you are creating the object inside the wrapper - you'd better have a check for deletion & delete it in the Destructor of its wrapper (though I doubt, that you really should use here another class for CPerson creation - yuo can create its object in its - CPerson's - Constructor)
First you need a default constructor and copy constructor in your code. second you need to initialize the struct to default value, the struct only allocate space in the memory and not initialize it, so you can end up with cases where a long variable has some weird value like -1823834393939, so always set your struct to default values.
then when you return a class or struct from a function. the copy constructor will be called that copy the values. so if you don't want to return the exact object you've created in your class you don't need to return a reference
struct Person
{
string FirstName;
string LastName;
Person()
{
FirstName = "";
LastName = "";
}
Person(Person &that)
{
FirstName = that.FirstName;
LastName = that.LastName;
}
};
class Builder
{
public:
Person Builder::Build(string argFirstname, string argLastName)
{
Person person;
person.FirstName = argFirstname;
person.LastName = argLastName;
return person;
};
};
void OnStart()
{
Builder b;
Person p = b.Build("Mohammad Hossein", "amri");
Print(p.FirstName + " " + p.LastName);
}
the output will be
I'm working with EF 4.5 and DbContext. At business rules layer level, I should implement checks to avoid change entity value properties in some entity scenarios. Sample: StartProjecteDate should be readonly if ProjectIsStarted but not in other status.
I follow DRY principle, for this reason, I should be able to inspect readonly properties list from context and also from UI.
My question:
Is there a DataAnnotation validator that can dynamically set properties as readonly?
(and if not, is there a different / better solution to this problem?)
Notice than I'm working with Web Forms (and Telerik) architecture, a clean and elegant pattern will be welcome.
I'm trying to set and get at run time EditableAttribute as Jesse Webb explains, but I'm not able to get dataannotation attributes from property, my code:
<EditableAttribute(False)>
<MaxLength(400, ErrorMessage:="Màxim 400 caracters")>
Public Property NomInvertebrat As String
Edited Nov 8 2013 after digging docs, it seems that dataanottions if for class but for instance object itself. Perhaps an iReadonlyableProperties interface may be a way.
I have a class containing extension methods that lets me read data annotations like this:
int maxRefLen = ReflectionAPI.GetProperty<Organisation, String>(x => x.Name)
.GetAttribute<StringLengthAttribute>()
.GetValueOrDefault(x => x.MaximumLength, 256);
So if you use it you should be able to do get the value of the EditableAttribute like this:
bool isEditable = ReflectionAPI.GetProperty<Foo, String>(x => x.NomInvertebrat)
.GetAttribute<EditableAttribute>()
.GetValueOrDefault(x => x.AllowEdit, true);
As for setting the data annotations at run-time, I haven't done it myself but I have read that there is a solution here: Setting data-annotations at runtime
Getting a list of all data annotations of a particular type I think would entail reading the entity framework metadata. Again I haven't tried this.
If you add that together I personally think it feels clunky rather than elegant, but you have asked for a solution using DataAnnotations and something more elegant would probably mean getting into your architecture.
I would be inclined to do this:
public bool StartDateIsReadOnly
{
//use this property client-side to disable the input
get{ return Project.IsStarted;}
}
//Implement IValidatable object to do server side validation
public IEnumerable<ValidationResult> Validate(ValidationContext validationContext
{
bool startdateIsChanged = // I'll leave you to work out this bit
var results = new List<ValidationResult>();
if(StartDateIsReadOnly && startdateIsChanged)
results.Add(new ValidationResult("Start Date cannot be changed after project is started");
}
Here is the ReflectionAPI class:
Please note that the class includes part of a hack that #JonSkeet posted and described as "evil". I personally think this bit ain't so bad, but you should read the following references:
Override a generic method for value types and reference types.
Evil code - overload resolution workaround
public static class ReflectionAPI
{
public static int GetValueOrDefault<TInput>(this TInput a, Func<TInput, int> func, int defaultValue)
where TInput : Attribute
//Have to restrict to struct or you get the error:
//The type 'R' must be a non-nullable value type in order to use it as parameter 'T' in the generic type or method 'System.Nullable<T>'
{
if (a == null)
return defaultValue;
return func(a);
}
public static Nullable<TResult> GetValueOrDefault<TInput, TResult>(this TInput a, Func<TInput, TResult> func, Nullable<TResult> defaultValue)
where TInput : Attribute
where TResult : struct
//Have to restrict to struct or you get the error:
//The type 'R' must be a non-nullable value type in order to use it as parameter 'T' in the generic type or method 'System.Nullable<T>'
{
if (a == null)
return defaultValue;
return func(a);
}
//In order to constrain to a class without interfering with the overload that has a generic struct constraint
//we need to add a parameter to the signature that is a reference type restricted to a class
public class ClassConstraintHack<T> where T : class { }
//The hack means we have an unused parameter in the signature
//http://msmvps.com/blogs/jon_skeet/archive/2010/11/02/evil-code-overload-resolution-workaround.aspx
public static TResult GetValueOrDefault<TInput, TResult>(this TInput a, Func<TInput, TResult> func, TResult defaultValue, ClassConstraintHack<TResult> ignored = default(ClassConstraintHack<TResult>))
where TInput : Attribute
where TResult : class
{
if (a == null)
return defaultValue;
return func(a);
}
//I don't go so far as to use the inheritance trick decribed in the evil code overload resolution blog,
//just create some overloads that take nullable types - and I will just keep adding overloads for other nullable type
public static bool? GetValueOrDefault<TInput>(this TInput a, Func<TInput, bool?> func, bool? defaultValue)
where TInput : Attribute
{
if (a == null)
return defaultValue;
return func(a);
}
public static int? GetValueOrDefault<TInput>(this TInput a, Func<TInput, int?> func, int? defaultValue)
where TInput : Attribute
{
if (a == null)
return defaultValue;
return func(a);
}
public static T GetAttribute<T>(this PropertyInfo p) where T : Attribute
{
if (p == null)
return null;
return p.GetCustomAttributes(false).OfType<T>().LastOrDefault();
}
public static PropertyInfo GetProperty<T, R>(Expression<Func<T, R>> expression)
{
if (expression == null)
return null;
MemberExpression memberExpression = expression.Body as MemberExpression;
if (memberExpression == null)
return null;
return memberExpression.Member as PropertyInfo;
}
}
.NET allows you to dynamically change structure of Class by implementing System.ComponentModel.ICustomTypeDescriptor. Most serializers support this interface.
// Sample Serialization
foreach(PropertyDescriptor p in TypeDescriptor.GetProperties(obj)){
string name = p.PropertyName;
object value = p.GetValue(obj);
}
Internally TypeDescriptor uses Reflection, but the implementation allows us to override reflection attributes easily.
Here are three steps of implementation,
// Implement System.ComponentModel.ICustomTypeDescriptor Interface on
// your Entity
public class MyEntity: System.ComponentModel.ICustomTypeDescriptor
{
....
// most methods needs only call to default implementation as shown below
System.ComponentModel.AttributeCollection
System.ComponentModel.ICustomTypeDescriptor.GetAttributes()
{
return TypeDescriptor.GetAttributes(this, true);
}
string System.ComponentModel.ICustomTypeDescriptor.GetClassName()
{
return TypeDescriptor.GetClassName(this, true);
}
string System.ComponentModel.ICustomTypeDescriptor.GetComponentName()
{
return TypeDescriptor.GetComponentName(this, true);
}
System.ComponentModel.TypeConverter System.ComponentModel.ICustomTypeDescriptor.GetConverter()
{
return TypeDescriptor.GetConverter(this, true);
}
System.ComponentModel.EventDescriptor System.ComponentModel.ICustomTypeDescriptor.GetDefaultEvent()
{
return TypeDescriptor.GetDefaultEvent(this, true);
}
System.ComponentModel.PropertyDescriptor System.ComponentModel.ICustomTypeDescriptor.GetDefaultProperty()
{
return TypeDescriptor.GetDefaultProperty(this, true);
}
object System.ComponentModel.ICustomTypeDescriptor.GetEditor(Type editorBaseType)
{
return TypeDescriptor.GetEditor(this, editorBaseType, true);
}
System.ComponentModel.EventDescriptorCollection System.ComponentModel.ICustomTypeDescriptor.GetEvents(Attribute[] attributes)
{
return TypeDescriptor.GetEvents(this, attributes, true);
}
System.ComponentModel.EventDescriptorCollection System.ComponentModel.ICustomTypeDescriptor.GetEvents()
{
return TypeDescriptor.GetEvents(this, true);
}
System.ComponentModel.PropertyDescriptorCollection System.ComponentModel.ICustomTypeDescriptor.GetProperties(Attribute[] attributes)
{
return TypeDescriptor.GetProperties(this, attributes, true);
}
object System.ComponentModel.ICustomTypeDescriptor.GetPropertyOwner(System.ComponentModel.PropertyDescriptor pd)
{
return this;
}
// The Only method that needs different implementation is below
System.ComponentModel.PropertyDescriptorCollection
System.ComponentModel.ICustomTypeDescriptor.GetProperties()
{
// ... you are supposed to create new instance of
// PropertyDescriptorCollection with PropertyDescriptor
PropertyDescriptorCollection pdc = new PropertyDescriptorCollection();
foreach(PropertyDescriptor p in TypeDescriptor.GetProperties(this,true)){
// if readonly..
AtomPropertyDescriptor ap = new AtomPropertyDescriptor(p, p.Name);
// or
AtomPropertyDescriptor ap = new AtomPropertyDescriptor(p, p.Name,
true,
new XmlIgnoreAttribute(),
new ScriptIgnoreAttribute(),
new ReadOnlyAttribute());
pdc.Add(ap);
}
return pdc;
}
}
// And here is the AtomPropertyDescriptorClass
public class AtomPropertyDescriptor : PropertyDescriptor
{
PropertyDescriptor desc;
bool? readOnly = null;
public AtomPropertyDescriptor(PropertyDescriptor pd, string name,
bool? readOnly, params Attribute[] attrs) :
base(name, attrs)
{
desc = pd;
this.readOnly = readOnly;
}
public override bool CanResetValue(object component)
{
return desc.CanResetValue(component);
}
public override Type ComponentType
{
get
{
return desc.ComponentType;
}
}
public override object GetValue(object component)
{
return desc.GetValue(component);
}
public override bool IsReadOnly
{
get
{
if (readOnly.HasValue)
return readOnly.Value;
return desc.IsReadOnly;
}
}
public override Type PropertyType
{
get { return desc.PropertyType; }
}
public override void ResetValue(object component)
{
desc.ResetValue(component);
}
public override void SetValue(object component, object value)
{
desc.SetValue(component, value);
}
public override bool ShouldSerializeValue(object component)
{
return desc.ShouldSerializeValue(component);
}
}
I think what you are looking for is a custom Annotation Attribute like this:
<DisableEditAttribute(this.IsProjectStarted)>
Public Property NomInvertebrat As String
public override bool IsValid(bool value)
{
bool result = true;
// Add validation logic here.
if(value)
{
//Compare Current Value Against DB Value.
}
return result;
}
See MSDN: http://msdn.microsoft.com/en-us/library/cc668224(v=vs.98).aspx
i have some mandatory fields and when calling the constructor of my class i got null reference exception, how prevent this please.
new Part(Partitems["Name"].ToString(), Partitems["Logo"].ToString(), Partitems["Description"].ToString(), Partitems["URL"].ToString()));
MY Class :
public class Part
{
string _Name= string.Empty;
public string Name
{
get { return _Name; }
set { _Name = value; }
}
...
..
EDIT
Here is constructor code
public Part(string Name, string Logo, string Description, string URL)
{
this.Name = Name;
this.Logo = Logo;
this.Description = Description;
this.URL = URL;
}
may be one of this properties
Partitems["Name"], Partitems["Logo"],Partitems["Description"],Partitems["URL"]
is null? Your constructor code will help to understand problem more...
There's nothing in the constructor that could cause a null reference exception, as assigning null to a string property is perfectly valid. So, either your Partitems variable is null, or one of it's properties returns null.
Example for handling both situations:
if (Partitems != null) {
object name = Partitems["Name"] ?? String.Empty;
object logo = Partitems["Logo"] ?? String.Empty;
object description = Partitems["Description"] ?? String.Empty;
object url = Partitems["URL"] ?? String.Empty;
part = new Part(name.ToString(), logo.ToString(), description.ToString(), url.ToString()));
} else {
// Oops, Partitems is null. Better tell the user or log the error...
}
This will replace any null values in the properties with empty strings. However, you might want to handle it as an error situation instead, depending on the situation.