Rectangle 27 32

It's important to remember that an Endpoint, created by a Component (i.e. Endpoint Factory), can sit at either end of a Camel Route. If you put the Component at the start of a route then there must be an implementation of the Consumer part of the Component. This does the work of converting the specific input/request (like an HTTP request) into something generic - a Camel Exchange - that can travel down a Route. Whereas if you put the Component at end of a route then you must have an implementation of a Producer. The Producer does the work of taking the Exchange from the end of a route and converting it into something specific (like a JMS message).

I find the Fuse ESB documentation to be better (in general) than the Apache Camel website. From the Fuse ESB Component page:

Consumer endpoints consume requests. They always appear at the start of a route and they encapsulate the code responsible for receiving incoming requests and dispatching outgoing replies.

Producer endpoints produce requests. They always appears at the end of a route and they encapsulate the code responsible for dispatching outgoing requests and receiving incoming replies.

Nice clear explanation thanks. The terms are often counter-intuitive, as a typical user of camel is thinking in terms of routes and messageExchanges, rather than authoring a new camel component. From the perspective of a messageExchange, the terms are the wrong way around. A Consumer produces a messageExchange, while a Producer consumes a messageExchange.

As a newbie to Camel, I'm really confused by this answer. If I look at the org.apache.camel.Producer interface, the only contracts it specifies are createExchange methods. And looking at something like the EventDrivenPollingConsumer, the only thing that does with an Exchange is process it. Doesn't an Exchange travel through a route - that's what Processors work with, so that was my impression.

@Snekse The descriptive part of the answer says exactly this. But in any case I don't think this technicality has much relevance for a conceptual understanding of producers and consumers.

messaging - Apache Camel producers and consumers - Stack Overflow

messaging apache-camel
Rectangle 27 1

  • To make sure the data is loaded before the routes are reached, you can use a resolver. The resolver will be linked to your route paths, and will be a function that will execute the Http request before the route is resolved.
  • You need to store your data in a store once the call has been made. That can be done with external stores such as @Ngrx, but the most basic example is to use a variable inside your service (even better with getter / setter) :
_equipmentDetails: any  // Should be private in use-case

...

getEquipmentDetail(clientId, modemId): Observable<any> {
  const url = 'myOwnUrl';

  return this.Http
    .get(url)
    .map(result => {
      if (result) {
        // Notice how I set the value from the Http call.
        return this._equipmentDetails = result.json();
      }
    })
    .catch(this.handleError);
}

And then you can get the property value from your components through dependency injection.

constructor(
    ...
    private equipmentService: ServiceClass,
  ) {
    this.equipmentService._equipmentDetails;
  }

Angular/rxjs/Subject/BehaviorSubject share http service data across mu...

angular rxjs
Rectangle 27 30

For one-way data binding from parent to child, use the @Input decorator (as recommended by the style guide) to specify an input property on the child component

@Input() model: any;   // instead of any, specify your type

and use template property binding in the parent template

<child [model]="parentModel"></child>

Since you are passing an object (a JavaScript reference type) any changes you make to object properties in the parent or the child component will be reflected in the other component, since both components have a reference to the same object. I show this in the Plunker.

If you reassign the object in the parent component

this.model = someNewModel;

Angular will propagate the new object reference to the child component (automatically, as part of change detection).

The only thing you shouldn't do is reassign the object in the child component. If you do this, the parent will still reference the original object. (If you do need two-way data binding, see https://stackoverflow.com/a/34616530/215945).

@Component({
  selector: 'child',
  template: `<h3>child</h3> 
    <div>{{model.prop1}}</div>
    <button (click)="updateModel()">update model</button>`
})
class Child {
  @Input() model: any;   // instead of any, specify your type
  updateModel() {
    this.model.prop1 += ' child';
  }
}

@Component({
  selector: 'my-app',
  directives: [Child],
  template: `
    <h3>Parent</h3>
    <div>{{parentModel.prop1}}</div>
    <button (click)="updateModel()">update model</button>
    <child [model]="parentModel"></child>`
})
export class AppComponent {
  parentModel = { prop1: '1st prop', prop2: '2nd prop' };
  constructor() {}
  updateModel() { this.parentModel.prop1 += ' parent'; }
}

you're doing god's work! any advice for sibling components? in my case I have 2 at root level bootstrapped. HeaderComponent has a search input which i want to share with components in the body..

@SonicSoul, put the data in a JavaScript reference type in the parent, or, if they don't share a parent, put the data in a shared service. For the service approach, you can again used a JavaScript reference type or use observables.

thanks! I am trying the service route.. didn't think of actually putting the Observable on it

Thanks it works like a charm

@Mark we have learned Angular1 with your posts and know again you are teaching us Angular2, you rock :)

angular2 directives - How to pass object from one component to another...

angular angular2-directives
Rectangle 27 30

For one-way data binding from parent to child, use the @Input decorator (as recommended by the style guide) to specify an input property on the child component

@Input() model: any;   // instead of any, specify your type

and use template property binding in the parent template

<child [model]="parentModel"></child>

Since you are passing an object (a JavaScript reference type) any changes you make to object properties in the parent or the child component will be reflected in the other component, since both components have a reference to the same object. I show this in the Plunker.

If you reassign the object in the parent component

this.model = someNewModel;

Angular will propagate the new object reference to the child component (automatically, as part of change detection).

The only thing you shouldn't do is reassign the object in the child component. If you do this, the parent will still reference the original object. (If you do need two-way data binding, see https://stackoverflow.com/a/34616530/215945).

@Component({
  selector: 'child',
  template: `<h3>child</h3> 
    <div>{{model.prop1}}</div>
    <button (click)="updateModel()">update model</button>`
})
class Child {
  @Input() model: any;   // instead of any, specify your type
  updateModel() {
    this.model.prop1 += ' child';
  }
}

@Component({
  selector: 'my-app',
  directives: [Child],
  template: `
    <h3>Parent</h3>
    <div>{{parentModel.prop1}}</div>
    <button (click)="updateModel()">update model</button>
    <child [model]="parentModel"></child>`
})
export class AppComponent {
  parentModel = { prop1: '1st prop', prop2: '2nd prop' };
  constructor() {}
  updateModel() { this.parentModel.prop1 += ' parent'; }
}

you're doing god's work! any advice for sibling components? in my case I have 2 at root level bootstrapped. HeaderComponent has a search input which i want to share with components in the body..

@SonicSoul, put the data in a JavaScript reference type in the parent, or, if they don't share a parent, put the data in a shared service. For the service approach, you can again used a JavaScript reference type or use observables.

thanks! I am trying the service route.. didn't think of actually putting the Observable on it

Thanks it works like a charm

@Mark we have learned Angular1 with your posts and know again you are teaching us Angular2, you rock :)

angular2 directives - How to pass object from one component to another...

angular angular2-directives
Rectangle 27 43

You can use PrimeFaces' <p:remoteCommand> for this.

<p:remoteCommand name="updateGrowl" update="showmessage" />

which is to be invoked as

<p:commandButton ... oncomplete="addMemberDlg.hide(); updateGrowl();" />
autoUpdate
<p:growl>
true
<p:growl autoUpdate="true" life="1500" id="showmessage"/>

It'll auto-update itself on every ajax request. If your component actually didn't support it, then you could always wrap it in a <p:outputPanel> which also supports that attribute.

<p:outputPanel autoUpdate="true">
    ...
</p:outputPanel>

Just a FYI if the reader is using a Jurassic version of PF, growl's autoupdate is only available from 3.0 and above. blog.primefaces.org/?p=1164

jsf - How to trigger component refresh from javascript in primefaces? ...

jsf jsf-2 primefaces
Rectangle 27 24

You need to provide the so called client id, which you'll find on UIComponent.

The following is a quick example of how to use this.

Consider the following bean:

@ManagedBean
@RequestScoped
public class ComponentMsgBean {

    private UIComponent component;

    public UIComponent getComponent() {
        return component;
    }

    public void setComponent(UIComponent component) {
        this.component = component;
    }

    public String doAction() {

        FacesContext context = FacesContext.getCurrentInstance();

        context.addMessage(component.getClientId(), new FacesMessage("Test msg"));

        return "";
    }

}

being used on the following Facelet:

<html xmlns="http://www.w3.org/1999/xhtml"
    xmlns:h="http://java.sun.com/jsf/html"
    xmlns:f="http://java.sun.com/jsf/core"
    xmlns:ui="http://java.sun.com/jsf/facelets" 
    >

    <h:body>

        <h:form>
            <h:outputText id="test" value="test component" binding="#{componentMsgBean.component}"/>
            <h:message for="test"/>

            <h:commandButton value="click me" action="#{componentMsgBean.doAction}" />
        </h:form>

    </h:body>
</html>

This will add a Faces message with content "Test msg" for the outputText component used in the example.

So, without binding i just need to know the generated clientId for the components' id and that's it?

Yes, you could grab this ID from the markup being send to your browser. But depending on this is rather brittle, as this ID might change when you move your component or add other components to your view.

OK, but i need to send messages without binding. Any easy way to find out the clientId from Id? (Without TLD functions)

Given the (relative) id of a component you can not universally calculate the clientId. The JSF spec only requires those Ids to be unique within the scope of a naming container. If you think those relative Ids are unique enough in your application you could locate the component starting from the view root using findComponent, e.g. context.getViewRoot().findComponent("myText")

That's what i was looking for. Thanks!

jsf 2 - How to add a message to a specific component from JSF backing ...

jsf jsf-2 richfaces
Rectangle 27 34

I am not sure why you say that using cloneWithProps is a bad solution, but here is a working example using it.

var Hello = React.createClass({
    render: function() {
        return <div>Hello {this.props.name}</div>;
    }
});

var App = React.createClass({
    render: function() {
        return (
            <Group ref="buttonGroup">
                <Button key={1} name="Component A"/>
                <Button key={2} name="Component B"/>
                <Button key={3} name="Component C"/>
            </Group>
        );
    }
});

var Group = React.createClass({
    getInitialState: function() {
        return {
            selectedItem: null
        };
    },

    selectItem: function(item) {
        this.setState({
            selectedItem: item
        });
    },

    render: function() {
        var selectedKey = (this.state.selectedItem && this.state.selectedItem.props.key) || null;
        var children = this.props.children.map(function(item, i) {
            var isSelected = item.props.key === selectedKey;
            return React.addons.cloneWithProps(item, {
                isSelected: isSelected,
                selectItem: this.selectItem,
                key: item.props.key
            });
        }, this);

        return (
            <div>
                <strong>Selected:</strong> {this.state.selectedItem ? this.state.selectedItem.props.name : 'None'}
                <hr/>
                {children}
            </div>
        );
    }

});

var Button = React.createClass({
    handleClick: function() {
        this.props.selectItem(this);
    },

    render: function() {
        var selected = this.props.isSelected;
        return (
            <div
                onClick={this.handleClick}
                className={selected ? "selected" : ""}
            >
                {this.props.name} ({this.props.key}) {selected ? "<---" : ""}
            </div>
        );
    }

});


React.renderComponent(<App />, document.body);

Here's a jsFiddle showing it in action.

EDIT: here's a more complete example with dynamic tab content : jsFiddle

This is so unbelievably complicated to do something D3/Jquery could do with just a :not selector on this. Is there any real advantage, other than having used React?

FYI.. this is not "updating children state from the parent state", this is child updating the parent state in order to update the children state. There's a difference in the verbage here. In case this confuses people.

@Incodeveritas you're right, sibling relationship, parent-child, and child-parent is a bit complicated. This is a modular way of doing things. But keep in mind JQuery is a quick hack to get things done, ReactJS keeps things according to an orchestration.

Reactjs: how to modify dynamic child component state or props from par...

reactjs
Rectangle 27 34

I've found an example on the internet.

<div ng-cloak ng-app="ManagerApp">
    <div id="MainWrap" class="container" ng-controller="ManagerCtrl">
       <span class="label label-info label-ext">Exposing Controller Function outside the module via onClick function call</span>
       <button onClick='ajaxResultPost("Update:Name:With:JOHN","accept",true);'>click me</button>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.data"></span>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.type"></span>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.res"></span>
       <br/>
       <input type="text" ng-model="sampletext" size="60">
       <br/>
    </div>
</div>
var angularApp = angular.module('ManagerApp', []);
angularApp.controller('ManagerCtrl', ['$scope', function ($scope) {

$scope.customParams = {};

$scope.updateCustomRequest = function (data, type, res) {
    $scope.customParams.data = data;
    $scope.customParams.type = type;
    $scope.customParams.res = res;
    $scope.sampletext = "input text: " + data;
};



}]);

function ajaxResultPost(data, type, res) {
    var scope = angular.element(document.getElementById("MainWrap")).scope();
    scope.$apply(function () {
    scope.updateCustomRequest(data, type, res);
    });
}

Working with a legacy jQuery validation library which MUST be used. So, it's either 1: rewrite the library, 2: create directive to wrap the library, 3: 2 lines of code to call the angular submit when valid...

if i do not use apply then function which will update view data will not be reflected ?

AngularJS. How to call controller function from outside of controller ...

angularjs
Rectangle 27 34

I've found an example on the internet.

<div ng-cloak ng-app="ManagerApp">
    <div id="MainWrap" class="container" ng-controller="ManagerCtrl">
       <span class="label label-info label-ext">Exposing Controller Function outside the module via onClick function call</span>
       <button onClick='ajaxResultPost("Update:Name:With:JOHN","accept",true);'>click me</button>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.data"></span>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.type"></span>
       <br/> <span class="label label-warning label-ext" ng-bind="customParams.res"></span>
       <br/>
       <input type="text" ng-model="sampletext" size="60">
       <br/>
    </div>
</div>
var angularApp = angular.module('ManagerApp', []);
angularApp.controller('ManagerCtrl', ['$scope', function ($scope) {

$scope.customParams = {};

$scope.updateCustomRequest = function (data, type, res) {
    $scope.customParams.data = data;
    $scope.customParams.type = type;
    $scope.customParams.res = res;
    $scope.sampletext = "input text: " + data;
};



}]);

function ajaxResultPost(data, type, res) {
    var scope = angular.element(document.getElementById("MainWrap")).scope();
    scope.$apply(function () {
    scope.updateCustomRequest(data, type, res);
    });
}

Working with a legacy jQuery validation library which MUST be used. So, it's either 1: rewrite the library, 2: create directive to wrap the library, 3: 2 lines of code to call the angular submit when valid...

if i do not use apply then function which will update view data will not be reflected ?

AngularJS. How to call controller function from outside of controller ...

angularjs
Rectangle 27 15

Oh!! may be I' m too late to answer the question! But never mind.This might help you or other to share data between components using Router,Shared-Service and Shared-object used withing shared-service. I hope this will surely help.

import {Component,bind} from 'angular2/core';

import {bootstrap} from 'angular2/platform/browser';

import {Router,ROUTER_PROVIDERS,RouteConfig, ROUTER_DIRECTIVES,APP_BASE_HREF,LocationStrategy,RouteParams,ROUTER_BINDINGS} from 'angular2/router';

import {SharedService} from 'src/sharedService';

    import {ComponentFirst} from 'src/cone';
import {ComponentTwo} from 'src/ctwo';


@Component({
  selector: 'my-app',
  directives: [ROUTER_DIRECTIVES],
  template: `
    <h1>
      Home
    </h1> 

    <router-outlet></router-outlet>
      `,

})

@RouteConfig([
  {path:'/component-first', name: 'ComponentFirst', component: ComponentFirst}
  {path:'/component-two', name: 'ComponentTwo', component: ComponentTwo}

])

export class AppComponent implements OnInit {

  constructor(router:Router)
  {
    this.router=router;
  }

    ngOnInit() {
    console.log('ngOnInit'); 
    this.router.navigate(['/ComponentFirst']);
  }



}

    bootstrap(AppComponent, [SharedService,
    ROUTER_PROVIDERS,bind(APP_BASE_HREF).toValue(location.pathname)
    ]);
import {Component,View,bind} from 'angular2/core';
import {SharedService} from 'src/sharedService';
import {Router,ROUTER_PROVIDERS,RouteConfig, ROUTER_DIRECTIVES,APP_BASE_HREF,LocationStrategy,RouteParams,ROUTER_BINDINGS} from 'angular2/router';
@Component({
  //selector: 'f',
  template: `
    <div><input #myVal type="text" >
    <button (click)="send(myVal.value)">Send</button>
      `,

})

export class ComponentFirst   {

  constructor(service:SharedService,router:Router){
    this.service=service;
    this.router=router;
  }

  send(str){
    console.log(str);
    this.service.saveData(str); 
    console.log('str');
    this.router.navigate(['/ComponentTwo']);
  }

}
import {Component,View,bind} from 'angular2/core';
import {SharedService} from 'src/sharedService';
import {Router,ROUTER_PROVIDERS,RouteConfig, ROUTER_DIRECTIVES,APP_BASE_HREF,LocationStrategy,RouteParams,ROUTER_BINDINGS} from 'angular2/router';
@Component({
  //selector: 'f',
  template: `
    <h1>{{myName}}</h1>
    <button (click)="back()">Back<button>
      `,

})

export class ComponentTwo   {

  constructor(router:Router,service:SharedService)
  {
    this.router=router;
    this.service=service;
    console.log('cone called');
    this.myName=service.getData();
  }
  back()
  {
     console.log('Back called');
    this.router.navigate(['/ComponentFirst']);
  }

}
import {Component, Injectable,Input,Output,EventEmitter} from 'angular2/core'

// Name Service
export interface myData {
   name:string;
}



@Injectable()
export class SharedService {
  sharingData: myData={name:"nyks"};
  saveData(str){
    console.log('save data function called' + str + this.sharingData.name);
    this.sharingData.name=str; 
  }
  getData:string()
  {
    console.log('get data function called');
    return this.sharingData.name;
  }
}

This is what exactly i was looking for.Thanks God i found this solution. Thanks @micronyks for this answer. :) :)

angular - How to send a value from one component to another? - Stack O...

angular angular2-routing angular2-template angular2-directives
Rectangle 27 25

If you're using at least PrimeFaces 3.3, then you can use PrimeFaces Selectors for this. This allows you using jQuery CSS selector syntax in process and update attributes of PrimeFaces ajax components.

<h:form>
    <h:inputText ... />
    <h:inputText ... />
    <h:inputText ... styleClass="noupdate" />
    <h:inputText ... />
    <h:inputText ... />
    <p:commandButton ... update="@(form :not(.noupdate))"/>
</h:form>

This example will update the entire form except for inputs having class="noupdate" in the client side.

If you want to update a all children of a certain component except one, replace 'form' by the id of the surrounding component (or a class or...)

<h:form id="form">
    <h:panel id="myPanel">
        <h:inputText ... />
        <h:inputText ... />
        <h:inputText ... styleClass="noupdate" />
    </h:panel>
    <h:inputText ... />
    <h:inputText ... />
    <p:commandButton ... update="@(form :not(.noupdate))"/>
</h:form>

<p:commandButton ... update="@(#form\:myPanel :not(.noupdate))"/>

Just make sure you use the full client-side id.

thank you for your answer ! I was not aware that jQuery selectors can be used in Primefaces update attribute.

update="@(form :not(.noupdate)),otherPanel"
update="otherPanel,@(form :not(.noupdate))"

How to do the same thing with only JSF <p:ajax>?

jsf - How to exclude child component in ajax update of a parent compon...

ajax jsf primefaces
Rectangle 27 25

If you're using at least PrimeFaces 3.3, then you can use PrimeFaces Selectors for this. This allows you using jQuery CSS selector syntax in process and update attributes of PrimeFaces ajax components.

<h:form>
    <h:inputText ... />
    <h:inputText ... />
    <h:inputText ... styleClass="noupdate" />
    <h:inputText ... />
    <h:inputText ... />
    <p:commandButton ... update="@(form :not(.noupdate))"/>
</h:form>

This example will update the entire form except for inputs having class="noupdate" in the client side.

If you want to update a all children of a certain component except one, replace 'form' by the id of the surrounding component (or a class or...)

<h:form id="form">
    <h:panel id="myPanel">
        <h:inputText ... />
        <h:inputText ... />
        <h:inputText ... styleClass="noupdate" />
    </h:panel>
    <h:inputText ... />
    <h:inputText ... />
    <p:commandButton ... update="@(form :not(.noupdate))"/>
</h:form>

<p:commandButton ... update="@(#form\:myPanel :not(.noupdate))"/>

Just make sure you use the full client-side id.

thank you for your answer ! I was not aware that jQuery selectors can be used in Primefaces update attribute.

update="@(form :not(.noupdate)),otherPanel"
update="otherPanel,@(form :not(.noupdate))"

How to do the same thing with only JSF <p:ajax>?

jsf - How to exclude child component in ajax update of a parent compon...

ajax jsf primefaces
Rectangle 27 248

Look in HTML output for actual client ID

You need to look in the generated HTML output to find out the right client ID. Open the page in browser, do a rightclick and View Source. Locate the HTML representation of the JSF component of interest and take its id as client ID. You can use it in an absolute or relative way depending on the current naming container. See following chapter.

Note: if it happens to contain iteration index like :0:, :1:, etc (because it's inside an iterating component), then you need to realize that updating a specific iteration round is not always supported. See bottom of answer for more detail on that.

If a component which you'd like to reference by ajax process/execute/update/render is inside the same NamingContainer parent, then just reference its own ID.

If it's not inside the same NamingContainer, then you need to reference it using an absolute client ID. An absolute client ID starts with the NamingContainer separator character, which is by default :.

<h:form id="form">
    <p:commandLink update=":otherform:result"> <!-- OK! -->
</h:form>
<h:form id="otherform">
    <h:panelGroup id="result" />
</h:form>

NamingContainer components are for example <h:form>, <h:dataTable>, <p:tabView>, <cc:implementation> (thus, all composite components), etc. You recognize them easily by looking at the generated HTML output, their ID will be prepended to the generated client ID of all child components. Note that when they don't have a fixed ID, then JSF will use an autogenerated ID in j_idXXX format. You should absolutely avoid that by giving them a fixed ID. The OmniFaces NoAutoGeneratedIdViewHandler may be helpful in this during development.

If you know to find the javadoc of the UIComponent in question, then you can also just check in there whether it implements the NamingContainer interface or not. For example, the HtmlForm (the UIComponent behind <h:form> tag) shows it implements NamingContainer, but the HtmlPanelGroup (the UIComponent behind <h:panelGroup> tag) does not show it, so it does not implement NamingContainer. Here is the javadoc of all standard components and here is the javadoc of PrimeFaces.

So in your case of:

<p:tabView id="tabs"><!-- This is a NamingContainer -->
    <p:tab id="search"><!-- This is NOT a NamingContainer -->
        <h:form id="insTable"><!-- This is a NamingContainer -->
            <p:dialog id="dlg"><!-- This is NOT a NamingContainer -->
                <h:panelGrid id="display">

The generated HTML output of <h:panelGrid id="display"> looks like this:

<table id="tabs:insTable:display">

You need to take exactly that id as client ID and then prefix with : for usage in update:

<p:commandLink update=":tabs:insTable:display">

If this command link is inside an include/tagfile, and the target is outside it, and thus you don't necessarily know the ID of the naming container parent of the current naming container, then you can dynamically reference it via UIComponent#getNamingContainer() like so:

<p:commandLink update=":#{component.namingContainer.parent.namingContainer.clientId}:display">

Or, if this command link is inside a composite component and the target is outside it:

<p:commandLink update=":#{cc.parent.namingContainer.clientId}:display">

Or, if both the command link and target are inside same composite component:

<p:commandLink update=":#{cc.clientId}:display">
UIComponent#findComponent()

A search expression consists of either an identifier (which is matched exactly against the id property of a UIComponent, or a series of such identifiers linked by the UINamingContainer#getSeparatorChar character value. The search algorithm should operates as follows, though alternate alogrithms may be used as long as the end result is the same:

UIComponent
  • If the search expression begins with the the separator character (called an "absolute" search expression), the base will be the root UIComponent of the component tree. The leading separator character will be stripped off, and the remainder of the search expression will be treated as a "relative" search expression as described below.
  • Otherwise, if this UIComponent is a NamingContainer it will serve as the basis.
  • Otherwise, search up the parents of this component. If a NamingContainer is encountered, it will be the base.
  • Otherwise (if no NamingContainer is encountered) the root UIComponent will be the base.
  • If the search expression is a simple identifier, this value is compared to the id property, and then recursively through the facets and children of the base UIComponent (except that if a descendant NamingContainer is found, its own facets and children are not searched).
  • If the search expression includes more than one identifier separated by the separator character, the first identifier is used to locate a NamingContainer by the rules in the previous bullet point. Then, the findComponent() method of this NamingContainer will be called, passing the remainder of the search expression.

Note that PrimeFaces also adheres the JSF spec, but RichFaces uses "some additional exceptions".

"reRender" uses UIComponent.findComponent() algorithm (with some additional exceptions) to find the component in the component tree.

Those additional exceptions are nowhere in detail described, but it's known that relative component IDs (i.e. those not starting with :) are not only searched in the context of the closest parent NamingContainer, but also in all other NamingContainer components in the same view (which is a relatively expensive job by the way).

If this all still doesn't work, then verify if you aren't using <h:form prependId="false">. This will fail during processing the ajax submit and render. See also this related question: UIForm with prependId="false" breaks <f:ajax render>.

It was for long time not possible to reference a specific iterated item in iterating components like <ui:repeat> and <h:dataTable> like so:

<h:form id="form">
    <ui:repeat id="list" value="#{['one','two','three']}" var="item">
        <h:outputText id="item" value="#{item}" /><br/>
    </ui:repeat>

    <h:commandButton value="Update second item">
        <f:ajax render=":form:list:1:item" />
    </h:commandButton>
</h:form>

However, since Mojarra 2.2.5 the <f:ajax> started to support it (it simply stopped validating it; thus you would never face the in the question mentioned exception anymore; another enhancement fix is planned for that later).

This only doesn't work yet in current MyFaces 2.2.7 and PrimeFaces 5.2 versions. The support might come in the future versions. In the meanwhile, your best bet is to update the iterating component itself, or a parent in case it doesn't render HTML, like <ui:repeat>.

@this
@form
UIForm
@all
@none

PrimeFaces has enhanced this with new keywords and composite expression support:

@parent
@namingcontainer
UINamingContainer
@widgetVar(name)
widgetVar

You can also mix those keywords in composite expressions such as @form:@parent, @this:@parent:@parent, etc.

PrimeFaces Selectors (PFS) as in @(.someclass) allows you to reference components via jQuery CSS selector syntax. E.g. referencing components having all a common style class in the HTML output. This is particularly helpful in case you need to reference "a lot of" components. This only prerequires that the target components have all a client ID in the HTML output (fixed or autogenerated, doesn't matter). See also How do PrimeFaces Selectors as in update="@(.myClass)" work?

@jack: Just read javadoc: docs.oracle.com/javaee/6/api/javax/faces/component/ Since JSF 2.0 it has become configurable instead of a constant.

context.getViewRoot().findComponent(":inputform"         + UINamingContainer.getSeparatorChar(context) + "inputtext" );

Thank you, note about failing ajax rerender inside of form with prependId="false" saved my day.

what is the exact meaning of client ID as outlined in your explaination? Is it the same as in JSF -> "The client-side identifier for this component". regards+thanks for your work.

@antonu17: As stated in the answer, it's only supported in Mojarra's f:ajax.

jsf - How to find out client ID of component for ajax update/render? C...

ajax jsf jsf-2 primefaces clientid
Rectangle 27 4

You are 99% percent there.

React uses a one-way data-flow. So, events on nested components will not propagate to their parents.

<Button>
handleClick
this.props.handleClick
<Search>
var Button = React.createClass({
  handleClick: function () {
    this.props.onClick();
  },

  ...

});

Attached is a fiddle of your original post, with the required change. Instead of logging FROM BUTTON, it will now alert searching.

reactjs - How to override event handler function of child component fr...

event-handling reactjs
Rectangle 27 7

It's quite simple but you have to keep a few points in mind which I'll detail below, first the code.

To reference your children, in this case you want your children within your View, so you must use @ViewChildren and you must wait the view to be initialized so you do

@Component({
    selector: 'hello',
    template: `<child></child>`,
    directives : [Child]
})
export class Parent implements AfterViewInit {
  @ViewChildren(Child) children: QueryList<Child>;

  afterViewInit() {
    for(let child of this.children) {
      child.doSomething();
    }
  }

}

The loop inside afterViewInit() will work if you are transpiling to ES6 since angular2 internally uses Symbol.iterator. If you are transpiling to ES5 you'll have to workaround it since typescript does not support it (see plnkr for workaround).

this.children.first

@rob If you need only one child component then use @ViewChild instead of @ViewChildren

Is there a way to access one specific child using an ID or something? Just like getElementById in JS.

Also in the cheat sheet here angular.io/cheatsheet it says "not for directives". Is there ant other way to do this?

typescript - How to access a child component class from the parent com...

typescript angular
Rectangle 27 115

In a Nutshell

Similar topic is discussed here Equivalent of $compile in Angular 2. We need to use JitCompiler and NgModule. Read more about NgModule in Angular2 here:

JitCompiler

The principal is:1) create Template2) find ComponentFactory in cache - go to 7) 3) - create Component 4) - create Module 5) - compile Module 6) - return (and cache for later use) ComponentFactory7) use Target and ComponentFactory to create an Instance of dynamic Component

Here is a code snippet (more of it here) - Our custom Builder is returning just built/cached ComponentFactory and the view Target placeholder consume to create an instance of the DynamicComponent

// here we get a TEMPLATE with dynamic content === TODO
  var template = this.templateBuilder.prepareTemplate(this.entity, useTextarea);

  // here we get Factory (just compiled or from cache)
  this.typeBuilder
      .createComponentFactory(template)
      .then((factory: ComponentFactory<IHaveDynamicData>) =>
    {
        // Target will instantiate and inject component (we'll keep reference to it)
        this.componentRef = this
            .dynamicComponentTarget
            .createComponent(factory);

        // let's inject @Inputs to component instance
        let component = this.componentRef.instance;

        component.entity = this.entity;
        //...
    });
PartsModule:NgModule
DynamicModule:NgModule
PartsModule
  • create new RuntimeModule:NgModule. This module will contain the previously created Component type
JitCompiler.compileModuleAndAllComponentsAsync(runtimeModule)
ComponentFactory
  • create an Instance of the DynamicComponent - job of the View Target placeholder and ComponentFactory
@Inputs
INPUT
TEXTAREA
@Outputs
NgModule

While I would like to show a very simple example, in this case, I would need three modules (in fact 4 - but I do not count the AppModule). Please, take this rather than a simple snippet as a basis for a really solid dynamic component generator.

string-editor
text-editor
date-editor
number-editor
@NgModule({
  imports:      [ 
      CommonModule,
      FormsModule
  ],
  declarations: [
      DYNAMIC_DIRECTIVES
  ],
  exports: [
      DYNAMIC_DIRECTIVES,
      CommonModule,
      FormsModule
  ]
})
export class PartsModule { }

Where DYNAMIC_DIRECTIVES are extensible and are intended to hold all small parts used for our dynamic Component template/type. Check app/parts/parts.module.ts

The second will be module for our Dynamic stuff handling. It will contain hosting components and some providers.. which will be singletons. Therefor we will publish them standard way - with forRoot()

import { DynamicDetail }          from './detail.view';
import { DynamicTypeBuilder }     from './type.builder';
import { DynamicTemplateBuilder } from './template.builder';

@NgModule({
  imports:      [ PartsModule ],
  declarations: [ DynamicDetail ],
  exports:      [ DynamicDetail],
})

export class DynamicModule {

    static forRoot()
    {
        return {
            ngModule: DynamicModule,
            providers: [ // singletons accross the whole app
              DynamicTemplateBuilder,
              DynamicTypeBuilder
            ], 
        };
    }
}

Check the usage of the forRoot() in the AppModule

Finally, we will need an adhoc, runtime module.. but that will be created later, as a part of DynamicTypeBuilder job.

The forth module, application module, is the one who keeps declares compiler providers:

...
import { COMPILER_PROVIDERS } from '@angular/compiler';    
import { AppComponent }   from './app.component';
import { DynamicModule }    from './dynamic/dynamic.module';

@NgModule({
  imports:      [ 
    BrowserModule,
    DynamicModule.forRoot() // singletons
  ],
  declarations: [ AppComponent],
  providers: [
    COMPILER_PROVIDERS // this is an app singleton declaration
  ],

In our example we will process detail of this kind of entity

entity = { 
    code: "ABC123",
    description: "A description of this Entity" 
};

To create a template, in this plunker we use this simple/naive builder.

// plunker - app/dynamic/template.builder.ts
import {Injectable} from "@angular/core";

@Injectable()
export class DynamicTemplateBuilder {

    public prepareTemplate(entity: any, useTextarea: boolean){

      let properties = Object.keys(entity);
      let template = "<form >";
      let editorName = useTextarea 
        ? "text-editor"
        : "string-editor";

      properties.forEach((propertyName) =>{
        template += `
          <${editorName}
              [propertyName]="'${propertyName}'"
              [entity]="entity"
          ></${editorName}>`;
      });

      return template + "</form>";
    }
}

A trick here is - it builds a template which uses some set of known properties, e.g. entity. Such property(-ies) must be part of dynamic component, which we will create next.

To make it a bit more easier, we can use an interface to define properties, which our Template builder can use. This will be implemented by our dynamic Component type.

export interface IHaveDynamicData { 
    public entity: any;
    ...
}
DynamicTypeBuilder

So, we are touching the core of our solution. The Builder, will 1) create ComponentType 2) create its NgModule 3) compile ComponentFactory 4) cache it for later reuse.

// plunker - app/dynamic/type.builder.ts
import { JitCompiler } from '@angular/compiler';

@Injectable()
export class DynamicTypeBuilder {

  // wee need Dynamic component builder
  constructor(
    protected compiler: JitCompiler
  ) {}

And here is a snippet how to get a ComponentFactory:

// plunker - app/dynamic/type.builder.ts
// this object is singleton - so we can use this as a cache
private _cacheOfFactories:
     {[templateKey: string]: ComponentFactory<IHaveDynamicData>} = {};

public createComponentFactory(template: string)
    : Promise<ComponentFactory<IHaveDynamicData>> {    
    let factory = this._cacheOfFactories[template];

    if (factory) {
        console.log("Module and Type are returned from cache")

        return new Promise((resolve) => {
            resolve(factory);
        });
    }

    // unknown template ... let's create a Type for it
    let type   = this.createNewComponent(template);
    let module = this.createComponentModule(type);

    return new Promise((resolve) => {
        this.compiler
            .compileModuleAndAllComponentsAsync(module)
            .then((moduleWithFactories) =>
            {
                factory = _.find(moduleWithFactories.componentFactories
                                , { componentType: type });

                this._cacheOfFactories[template] = factory;

                resolve(factory);
            });
    });
}
Module

And here are two methods, which represent the really cool way how to create a decorated classes/types in runtime. Not only @Component but also the @NgModule

protected createNewComponent (tmpl:string) {
  @Component({
      selector: 'dynamic-component',
      template: tmpl,
  })
  class CustomDynamicComponent  implements IHaveDynamicData {
      @Input()  public entity: any;
  };
  // a component for this particular template
  return CustomDynamicComponent;
}
protected createComponentModule (componentType: any) {
  @NgModule({
    imports: [
      PartsModule, // there are 'text-editor', 'string-editor'...
    ],
    declarations: [
      componentType
    ],
  })
  class RuntimeComponentModule
  {
  }
  // a module for just this Type
  return RuntimeComponentModule;
}

our component dynamic types differ, but just by template. So we use that fact to cache them. This is really very important. Angular2 will also cache these.. by the type. And if we would recreate for the same template strings new types... we will start to generate memory leaks.

Final piece is a component, which hosts the target for our dynamic component, e.g. <div #dynamicContentPlaceHolder></div>. We get a reference to it and use ComponentFactory to create a component. That is in a nutshell, and here are all the pieces of that component (if needed, open plunker here)

import {Component, ComponentRef,ViewChild,ViewContainerRef}   from '@angular/core';
import {AfterViewInit,OnInit,OnDestroy,OnChanges,SimpleChange} from '@angular/core';

import { IHaveDynamicData, DynamicTypeBuilder } from './type.builder';
import { DynamicTemplateBuilder }               from './template.builder';

@Component({
  selector: 'dynamic-detail',
  template: `
<div>
  check/uncheck to use INPUT vs TEXTAREA:
  <input type="checkbox" #val (click)="refreshContent(val.checked)" /><hr />
  <div #dynamicContentPlaceHolder></div>  <hr />
  entity: <pre>{{entity | json}}</pre>
</div>
`,
})
export class DynamicDetail implements AfterViewInit, OnChanges, OnDestroy, OnInit
{ 
    // wee need Dynamic component builder
    constructor(
        protected typeBuilder: DynamicTypeBuilder,
        protected templateBuilder: DynamicTemplateBuilder
    ) {}
    ...

We just receive, template and component builders. Next are properties which are needed for our example (more in comments)

// reference for a <div> with #dynamicContentPlaceHolder
@ViewChild('dynamicContentPlaceHolder', {read: ViewContainerRef}) 
protected dynamicComponentTarget: ViewContainerRef;
// this will be reference to dynamic content - to be able to destroy it
protected componentRef: ComponentRef<IHaveDynamicData>;

// until ngAfterViewInit, we cannot start (firstly) to process dynamic stuff
protected wasViewInitialized = false;

// example entity ... to be recieved from other app parts
// this is kind of candiate for @Input
protected entity = { 
    code: "ABC123",
    description: "A description of this Entity" 
  };

In this simple scenario, our hosting component does not have any @Input. So it does not have to react to changes. But despite of that fact (and to be ready for coming changes) - we need to introduce some flag if the component was already (firstly) initiated. And only then we can start the magic.

ComponentFacotry
Component
protected refreshContent(useTextarea: boolean = false){

  if (this.componentRef) {
      this.componentRef.destroy();
  }

  // here we get a TEMPLATE with dynamic content === TODO
  var template = this.templateBuilder.prepareTemplate(this.entity, useTextarea);

  // here we get Factory (just compiled or from cache)
  this.typeBuilder
      .createComponentFactory(template)
      .then((factory: ComponentFactory<IHaveDynamicData>) =>
    {
        // Target will instantiate and inject component (we'll keep reference to it)
        this.componentRef = this
            .dynamicComponentTarget
            .createComponent(factory);

        // let's inject @Inputs to component instance
        let component = this.componentRef.instance;

        component.entity = this.entity;
        //...
    });
}
destroy()
// this is the best moment where to start to process dynamic stuff
public ngAfterViewInit(): void
{
    this.wasViewInitialized = true;
    this.refreshContent();
}
// wasViewInitialized is an IMPORTANT switch 
// when this component would have its own changing @Input()
// - then we have to wait till view is intialized - first OnChange is too soon
public ngOnChanges(changes: {[key: string]: SimpleChange}): void
{
    if (this.wasViewInitialized) {
        return;
    }
    this.refreshContent();
}

public ngOnDestroy(){
  if (this.componentRef) {
      this.componentRef.destroy();
      this.componentRef = null;
  }
}

That is pretty much it. Do not forget to Destroy anything what was built dynamically (ngOnDestroy). Also, be sure to cache dynamic types and modules if the only difference is their template.

Check it all in action here

to see previous versions (e.g. RC5 related) of this post, check the history

this look like such a complicated solution, the deprecated one was very simple and clear, is there other way to do this ?

I think the same way as @tibbus: this got way more complicated than it used to be with the deprecated code. Thanks for your answer, though.

@ribsies thanks for your note. Let me clarify something. Many other answers try to make it simple. But I am trying to explain it and show it in a scenario, closed to real usage. We would need to cache stuff, we would have to call destroy on re-creation etc. So, while the magic of dynamic building is really in type.builder.ts as you've pointed, I wish, that any user would understand how to place that all into context... Hope it could be useful ;)

@Radim Khler - I have tried this example. it's working without AOT. But when I tried to do run this with AOT it shows error "No NgModule metadata found for RuntimeComponentModule". can you plz help me to solve this error .

The answer itself is perfect! But for real life applications not practicable. The angular team should provide a solution for this in the framework, as this is common requirement in business applications. If not, it has to be asked if Angular 2 is the right platform for business applications.

typescript - How can I use/create dynamic template to compile dynamic ...

angular typescript compilation angular2-compiler
Rectangle 27 52

This assumption is thus wrong. Unspecified component attributes are ignored by the JSF renderers.

You have basically the following options to get it to work:

Html5RenderKit

@Sam: it's not JSF impl specific.

How to add placeholder attribute to JSF input component? - Stack Overf...

jsf placeholder
Rectangle 27 240

Look in HTML output for actual client ID

You need to look in the generated HTML output to find out the right client ID. Open the page in browser, do a rightclick and View Source. Locate the HTML representation of the JSF component of interest and take its id as client ID. You can use it in an absolute or relative way depending on the current naming container. See following chapter.

Note: if it happens to contain iteration index like :0:, :1:, etc (because it's inside an iterating component), then you need to realize that updating a specific iteration round is not always supported. See bottom of answer for more detail on that.

If a component which you'd like to reference by ajax process/execute/update/render is inside the same NamingContainer parent, then just reference its own ID.

If it's not inside the same NamingContainer, then you need to reference it using an absolute client ID. An absolute client ID starts with the NamingContainer separator character, which is by default :.

<h:form id="form">
    <p:commandLink update=":otherform:result"> <!-- OK! -->
</h:form>
<h:form id="otherform">
    <h:panelGroup id="result" />
</h:form>

NamingContainer components are for example <h:form>, <h:dataTable>, <p:tabView>, <cc:implementation> (thus, all composite components), etc. You recognize them easily by looking at the generated HTML output, their ID will be prepended to the generated client ID of all child components. Note that when they don't have a fixed ID, then JSF will use an autogenerated ID in j_idXXX format. You should absolutely avoid that by giving them a fixed ID. The OmniFaces NoAutoGeneratedIdViewHandler may be helpful in this during development.

If you know to find the javadoc of the UIComponent in question, then you can also just check in there whether it implements the NamingContainer interface or not. For example, the HtmlForm (the UIComponent behind <h:form> tag) shows it implements NamingContainer, but the HtmlPanelGroup (the UIComponent behind <h:panelGroup> tag) does not show it, so it does not implement NamingContainer. Here is the javadoc of all standard components and here is the javadoc of PrimeFaces.

So in your case of:

<p:tabView id="tabs"><!-- This is a NamingContainer -->
    <p:tab id="search"><!-- This is NOT a NamingContainer -->
        <h:form id="insTable"><!-- This is a NamingContainer -->
            <p:dialog id="dlg"><!-- This is NOT a NamingContainer -->
                <h:panelGrid id="display">

The generated HTML output of <h:panelGrid id="display"> looks like this:

<table id="tabs:insTable:display">

You need to take exactly that id as client ID and then prefix with : for usage in update:

<p:commandLink update=":tabs:insTable:display">

If this command link is inside an include/tagfile, and the target is outside it, and thus you don't necessarily know the ID of the naming container parent of the current naming container, then you can dynamically reference it via UIComponent#getNamingContainer() like so:

<p:commandLink update=":#{component.namingContainer.parent.namingContainer.clientId}:display">

Or, if this command link is inside a composite component and the target is outside it:

<p:commandLink update=":#{cc.parent.namingContainer.clientId}:display">

Or, if both the command link and target are inside same composite component:

<p:commandLink update=":#{cc.clientId}:display">
UIComponent#findComponent()

A search expression consists of either an identifier (which is matched exactly against the id property of a UIComponent, or a series of such identifiers linked by the UINamingContainer#getSeparatorChar character value. The search algorithm should operates as follows, though alternate alogrithms may be used as long as the end result is the same:

UIComponent
  • If the search expression begins with the the separator character (called an "absolute" search expression), the base will be the root UIComponent of the component tree. The leading separator character will be stripped off, and the remainder of the search expression will be treated as a "relative" search expression as described below.
  • Otherwise, if this UIComponent is a NamingContainer it will serve as the basis.
  • Otherwise, search up the parents of this component. If a NamingContainer is encountered, it will be the base.
  • Otherwise (if no NamingContainer is encountered) the root UIComponent will be the base.
  • If the search expression is a simple identifier, this value is compared to the id property, and then recursively through the facets and children of the base UIComponent (except that if a descendant NamingContainer is found, its own facets and children are not searched).
  • If the search expression includes more than one identifier separated by the separator character, the first identifier is used to locate a NamingContainer by the rules in the previous bullet point. Then, the findComponent() method of this NamingContainer will be called, passing the remainder of the search expression.

Note that PrimeFaces also adheres the JSF spec, but RichFaces uses "some additional exceptions".

"reRender" uses UIComponent.findComponent() algorithm (with some additional exceptions) to find the component in the component tree.

Those additional exceptions are nowhere in detail described, but it's known that relative component IDs (i.e. those not starting with :) are not only searched in the context of the closest parent NamingContainer, but also in all other NamingContainer components in the same view (which is a relatively expensive job by the way).

If this all still doesn't work, then verify if you aren't using <h:form prependId="false">. This will fail during processing the ajax submit and render. See also this related question: UIForm with prependId="false" breaks <f:ajax render>.

It was for long time not possible to reference a specific iterated item in iterating components like <ui:repeat> and <h:dataTable> like so:

<h:form id="form">
    <ui:repeat id="list" value="#{['one','two','three']}" var="item">
        <h:outputText id="item" value="#{item}" /><br/>
    </ui:repeat>

    <h:commandButton value="Update second item">
        <f:ajax render=":form:list:1:item" />
    </h:commandButton>
</h:form>

However, since Mojarra 2.2.5 the <f:ajax> started to support it (it simply stopped validating it; thus you would never face the in the question mentioned exception anymore; another enhancement fix is planned for that later).

This only doesn't work yet in current MyFaces 2.2.7 and PrimeFaces 5.2 versions. The support might come in the future versions. In the meanwhile, your best bet is to update the iterating component itself, or a parent in case it doesn't render HTML, like <ui:repeat>.

@this
@form
UIForm
@all
@none

PrimeFaces has enhanced this with new keywords and composite expression support:

@parent
@namingcontainer
UINamingContainer
@widgetVar(name)
widgetVar

You can also mix those keywords in composite expressions such as @form:@parent, @this:@parent:@parent, etc.

PrimeFaces Selectors (PFS) as in @(.someclass) allows you to reference components via jQuery CSS selector syntax. E.g. referencing components having all a common style class in the HTML output. This is particularly helpful in case you need to reference "a lot of" components. This only prerequires that the target components have all a client ID in the HTML output (fixed or autogenerated, doesn't matter). See also How do PrimeFaces Selectors as in update="@(.myClass)" work?

@jack: Just read javadoc: docs.oracle.com/javaee/6/api/javax/faces/component/ Since JSF 2.0 it has become configurable instead of a constant.

context.getViewRoot().findComponent(":inputform"         + UINamingContainer.getSeparatorChar(context) + "inputtext" );

Thank you, note about failing ajax rerender inside of form with prependId="false" saved my day.

what is the exact meaning of client ID as outlined in your explaination? Is it the same as in JSF -> "The client-side identifier for this component". regards+thanks for your work.

@antonu17: As stated in the answer, it's only supported in Mojarra's f:ajax.

Sign up for our newsletter and get our top new questions delivered to your inbox (see an example).

jsf - How to find out client ID of component for ajax update/render? C...

ajax jsf jsf-2 primefaces clientid
Rectangle 27 10

Spring Boot 1.4 introduced testing annotation called @MockBean. So now mocking and spying on Spring beans is natively supported by Spring Boot.

java - Spring JUnit: How to Mock autowired component in autowired comp...

java spring junit dependency-injection
Rectangle 27 459

Kickoff example returning String as plain text

Indeed, the keyword is "ajax": Asynchronous JavaScript and XML. However, last years it's more than often Asynchronous JavaScript and JSON. Basically, you let JS execute an asynchronous HTTP request and update the HTML DOM tree based on the response data.

Since it's pretty a tedious work to make it to work across all browsers (especially Internet Explorer versus others), there are plenty of JavaScript libraries out which simplifies this in single functions and covers as many as possible browser-specific bugs/quirks under the hoods, such as jQuery, Prototype, Mootools. Since jQuery is most popular these days, I'll use it in the below examples.

Create a /some.jsp like below (note: the code doesn't expect the JSP file being placed in a subfolder, if you do so, alter servlet URL accordingly):

<!DOCTYPE html>
<html lang="en">
    <head>
        <title>SO question 4112686</title>
        <script src="http://code.jquery.com/jquery-latest.min.js"></script>
        <script>
            $(document).on("click", "#somebutton", function() { // When HTML DOM "click" event is invoked on element with ID "somebutton", execute the following function...
                $.get("someservlet", function(responseText) {   // Execute Ajax GET request on URL of "someservlet" and execute the following function with Ajax response text...
                    $("#somediv").text(responseText);           // Locate HTML DOM element with ID "somediv" and set its text content with the response text.
                });
            });
        </script>
    </head>
    <body>
        <button id="somebutton">press here</button>
        <div id="somediv"></div>
    </body>
</html>

Create a servlet with a doGet() method which look like this:

@Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
    String text = "some text";

    response.setContentType("text/plain");  // Set content type of the response so that jQuery knows what it can expect.
    response.setCharacterEncoding("UTF-8"); // You want world domination, huh?
    response.getWriter().write(text);       // Write response body.
}

Map this servlet on an URL pattern of /someservlet or /someservlet/* as below (obviously, the URL pattern is free to your choice, but you'd need to alter the someservlet URL in JS code examples over all place accordingly):

@WebServlet("/someservlet/*")
public class SomeServlet extends HttpServlet {
    // ...
}

Or, when you're not on a Servlet 3.0 compatible container yet (Tomcat 7, Glassfish 3, JBoss AS 6, etc or newer), then map it in web.xml the old fashioned way (see also our Servlets wiki page):

<servlet>
    <servlet-name>someservlet</servlet-name>
    <servlet-class>com.example.SomeServlet</servlet-class>
</servlet>
<servlet-mapping>
    <servlet-name>someservlet</servlet-name>
    <url-pattern>/someservlet/*</url-pattern>
</servlet-mapping>

Now open the http://localhost:8080/context/test.jsp in the browser and press the button. You'll see that the content of the div get updated with the servlet response.

With JSON instead of plaintext as response format you can even get some steps further. It allows for more dynamics. First, you'd like to have a tool to convert between Java objects and JSON strings. There are plenty of them as well (see the bottom of this page for an overview). My personal favourite is Google Gson. Download and put its JAR file in /WEB-INF/lib folder of your webapplication.

Here's an example which displays List<String> as <ul><li>. The servlet:

Do note that jQuery automatically parses the response as JSON and gives you directly a JSON object (responseJson) as function argument when you set the response content type to application/json. If you forget to set it or rely on a default of text/plain or text/html, then the responseJson argument wouldn't give you a JSON object, but a plain vanilla string and you'd need to manually fiddle around with JSON.parse() afterwards, which is thus totally unnecessary if you set the content type right in first place.

Map<String, String>
<option>

with

<select id="someselect"></select>
List<Product>
<table>
Product
Long id
String name
BigDecimal price

Here's an example which does effectively the same as previous example, but then with XML instead of JSON. When using JSP as XML output generator you'll see that it's less tedious to code the table and all. JSTL is this way much more helpful as you can actually use it to iterate over the results and perform server side data formatting. The servlet:

@Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
    List<Product> products = someProductService.list();

    request.setAttribute("products", products);
    request.getRequestDispatcher("/WEB-INF/xml/products.jsp").forward(request, response);
}

The JSP code (note: if you put the <table> in a <jsp:include>, it may be reusable elsewhere in a non-ajax response):

<?xml version="1.0" encoding="UTF-8"?>
<%@page contentType="application/xml" pageEncoding="UTF-8"%>
<%@taglib prefix="c" uri="http://java.sun.com/jsp/jstl/core" %>
<%@taglib prefix="fmt" uri="http://java.sun.com/jsp/jstl/fmt" %>
<data>
    <table>
        <c:forEach items="${products}" var="product">
            <tr>
                <td>${product.id}</td>
                <td><c:out value="${product.name}" /></td>
                <td><fmt:formatNumber value="${product.price}" type="currency" currencyCode="USD" /></td>
            </tr>
        </c:forEach>
    </table>
</data>
$(document).on("click", "#somebutton", function() {             // When HTML DOM "click" event is invoked on element with ID "somebutton", execute the following function...
    $.get("someservlet", function(responseXml) {                // Execute Ajax GET request on URL of "someservlet" and execute the following function with Ajax response XML...
        $("#somediv").html($(responseXml).find("data").html()); // Parse XML, find <data> element and append its HTML to HTML DOM element with ID "somediv".
    });
});

You'll by now probably realize why XML is so much more powerful than JSON for the particular purpose of updating a HTML document using Ajax. JSON is funny, but after all generally only useful for so-called "public web services". MVC frameworks like JSF use XML under the covers for their ajax magic.

You can use jQuery $.serialize() to easily ajaxify existing POST forms without fiddling around with collecting and passing the individual form input parameters. Assuming an existing form which works perfectly fine without JavaScript/jQuery (and thus degrades gracefully when enduser has JavaScript disabled):

<form id="someform" action="someservlet" method="post">
    <input type="text" name="foo" />
    <input type="text" name="bar" />
    <input type="text" name="baz" />
    <input type="submit" name="submit" value="Submit" />
</form>

You can progressively enhance it with ajax as below:

$(document).on("submit", "#someform", function(event) {
    var $form = $(this);

    $.post($form.attr("action"), $form.serialize(), function(response) {
        // ...
    });

    event.preventDefault(); // Important! Prevents submitting the form.
});

You can in the servlet distinguish between normal requests and ajax requests as below:

@Override
protected void doPost(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
    String foo = request.getParameter("foo");
    String bar = request.getParameter("bar");
    String baz = request.getParameter("baz");

    boolean ajax = "XMLHttpRequest".equals(request.getHeader("X-Requested-With"));

    // ...

    if (ajax) {
        // Handle ajax (JSON or XML) response.
    } else {
        // Handle regular (JSP) response.
    }
}

The jQuery Form plugin does less or more the same as above jQuery example, but it has additional transparent support for multipart/form-data forms as required by file uploads.

If you don't have a form at all, but just wanted to interact with the servlet "in the background" whereby you'd like to POST some data, then you can use jQuery $.param() to easily convert a JSON object to an URL-encoded query string.

var params = {
    foo: "fooValue",
    bar: "barValue",
    baz: "bazValue"
};

$.post("someservlet", $.param(params), function(response) {
    // ...
});

The same doPost() method as shown here above can be reused. Do note that above syntax also works with $.get() in jQuery and doGet() in servlet.

If you however intend to send the JSON object as a whole instead of as individual request parameters for some reason, then you'd need to serialize it to a string using JSON.stringify() (not part of jQuery) and instruct jQuery to set request content type to application/json instead of (default) application/x-www-form-urlencoded. This can't be done via $.post() convenience function, but needs to be done via $.ajax() as below.

var data = {
    foo: "fooValue",
    bar: "barValue",
    baz: "bazValue"
};

$.ajax({
    type: "POST",
    url: "someservlet",
    contentType: "application/json", // NOT dataType!
    data: JSON.stringify(data),
    success: function(response) {
        // ...
    }
});

Do note that a lot of starters mix contentType with dataType. The contentType represents the type of the request body. The dataType represents the (expected) type of the response body, which is usually unnecessary as jQuery already autodetects it based on response's Content-Type header.

Then, in order to process the JSON object in the servlet which isn't being sent as individual request parameters but as a whole JSON string the above way, you only need to manually parse the request body using a JSON tool instead of using getParameter() the usual way. Namely, servlets don't support application/json formatted requests, but only application/x-www-form-urlencoded or multipart/form-data formatted requests. Gson also supports parsing a JSON string into a JSON object.

JsonObject data = new Gson().fromJson(request.getReader(), JsonObject.class);
String foo = data.get("foo").getAsString();
String bar = data.get("bar").getAsString();
String baz = data.get("baz").getAsString();
// ...

Do note that this all is more clumsy than just using $.param(). Normally, you want to use JSON.stringify() only if the target service is e.g. a JAX-RS (RESTful) service which is for some reason only capable of consuming JSON strings and not regular request parameters.

Important to realize and understand is that any sendRedirect() and forward() call by the servlet on an ajax request would only forward or redirect the ajax request itself and not the main document/window where the ajax request originated. JavaScript/jQuery would in such case only retrieve the redirected/forwarded response as responseText variable in the callback function. If it represents a whole HTML page and not an ajax-specific XML or JSON response, then all you could do is to replace the current document with it.

document.open();
document.write(responseText);
document.close();

Note that this doesn't change the URL as enduser sees in browser's address bar. So there are issues with bookmarkability. Therefore, it's much better to just return an "instruction" for JavaScript/jQuery to perform a redirect instead of returning the whole content of the redirected page. E.g. by returning a boolean, or an URL.

String redirectURL = "http://example.com";

Map<String, String> data = new HashMap<>();
data.put("redirect", redirectURL);
String json = new Gson().toJson(data);

response.setContentType("application/json");
response.setCharacterEncoding("UTF-8");
response.getWriter().write(json);
function(responseJson) {
    if (responseJson.redirect) {
        window.location = responseJson.redirect;
        return;
    }

    // ...
}

need to parse the json on the last example.

@kuhaku: nope. If you read post from top to bottom, you'll learn why.

This answer has been my lifeline for the last month or so lol. Learning a bunch from it. I LOVE the XML example. Thanks for putting this together! One noob question though if you have time. Is there a reason for putting the xml folder in WEB-INF ?

java - How to use Servlets and Ajax? - Stack Overflow

java ajax jsp servlets