Following the question I would like to ask about the appropriate way to check the initial render status of a component (not the update status) in shadow DOM. Is there any similar to document.readyState or a promise?
I have also tried to do:
getItems() {
this.updateComplete
.then(() => {
this.nodesLists = this.shadowRoot.querySelectorAll(".name");
})
.then(...)
}
which also failed.
Tia
await this.updateComplete (or this.updateComplete.then(...)) is the correct way to wait until the element has no pending render work before e.g. querying the state of the element's rendering, so your code should generally work as long as the element is connected to the document before running getItems.
Example: https://jsbin.com/jiquhez/edit?html,console,output
Note however, that if you await updateComplete before the element is connected and the element has no properties set that would trigger a render, then updateComplete currently resolves before the first render. This may be considered an unintended bug, filed at lit-element/#594.
Note you may also want to look into using the firstUpdated lifecycle method, depending on your use case. This is a method you can implement on your class to perform one-time work following the first update/render cycle for the element (useful for e.g. selecting static nodes that won't change based on rendering).
Example: https://jsbin.com/limikas/edit?html,console,output
Related
I am attempting to create a search bar using a custom HTML component for predictive text input. The way this component is built, it generates several plain HTML children that I need to act on to get full features. Specifically, I need to execute a blur action on one of the generated elements when the user presses escape or enter.
I got it to work using a ref on the custom component and calling getElementsByClassName on the ref, but using getElementsByClassName does not seem like the best solution. It pierces through the virtual and has odd side effects when testing.
This is a snippet of the component being rendered:
<predictive-input id='header-search-bar-input' type='search'
value={this.state.keywords}
ref={(ref: any) => this.predictiveInput = ref}
onKeyDown={(e: React.KeyboardEvent<any>) => this.handleKeyDown(e)}>
</predictive-input>
and the keyDown handler:
private handleKeyDown(e: React.KeyboardEvent<any>) {
// must access the underlying input element of the kat-predictive-input
let input: HTMLElement = this.predictiveInput.getElementsByClassName('header-row-text value')[0] as HTMLElement;
if (e.key === 'Escape') {
// blur the predictive input when the user presses escape
input.blur();
} else if (e.key === 'Enter') {
// commit the search when user presses enter
input.blur();
// handles action of making actual search, using search bar contents
this.commitSearch();
}
}
The element renders two children, one for the bar itself and one for the predictive dropdown. The classes of the underlying in the first are 'header-row-text' and 'value', so the element is correctly selected, but I am worried that this is violating proper React style.
I am using React 16.2, so only callback refs are available. I would rather avoid upgrading, but if a 16.3+ solution is compelling enough, I could consider it.
If you don't have any control over the input then this is the best approach in my opinion.
It's not ideal, but as you're stuck with a 3rd party component you can only choose from the methods that are available to you. In this case, your only real options are to find the element based on its class, or its position in the hierarchy. Both might change if the package is updated, but if I had to choose which would be more stable, I'd go for className.
With Polymer 1.* and WCT, when testing my element <sp-veteran></sp-veteran> I am not able to stub out the methods ._getSpComboBox() and ._getItems() in the ready function. I get Error thrown outside of test function: this._getSpComboBox(...)._getItems is not a function.
Since it is in the ready function, I need to use the WCT api stub instead of sinon.stub since the later requires me to grab the element which I can not do before fixture().
Any suggestions?
original code:
_getSpComboBox: function() {
return Polymer.dom(this.$.veteran.root).querySelector('sp-combo-box');
},
ready: function() {
if (this.editMode) {
this._getSpComboBox()._getItems();
}
this.$.veteranNoAjax.read();
this._setStyle();
}
test:
<test-fixture id="sp-veteran">
<template>
<h2>edit veteran in edit mode</h2>
<sp-app>
<sp-toast></sp-toast>
<sp-veteran edit-mode></sp-veteran>
</sp-app>
</template>
</test-fixture>
before(() => {
replace('sp-app').with('fake-sp-app');
stub('sp-ajax', {read: ()=> entitiesMock});
const _getItems = ()=> entitiesMock;
stub('sp-veteran', {_getSpComboBox: ()=> _getItems});
Unfortunately testing ready in Polymer1 is kind of a pain, or at least I haven't found an easy way that doesn't have odd side-effects. Calling the ready method after you've attached your stubs/spies is always an option but as I mentioned it can cause some odd issues. This was alleviated in Polymer2 as ready is called by the first call of connectedCallback for your element, so you can create the element then bind your spies and manually add to trigger it, just don't forget to remove after.
In the case of DOM manipulation in a Polymer element, you should be using the attached lifecycle instead, this will solve your issue as I mentioned above for testing, but it also saves you a weird potential usage case in the future. Since ready only runs once for an instance of an element, any logic in your ready statement won't get re-run if that element is re-used later, instead if you put the logic in your attached lifecycle if that element is removed from the DOM then added again later in another location it will rerun it's logic to fetch it's new children.
I need to setup the behaviour of a polymer web-compontent at runtime. I tried to change the "behaviours" array by pushing the new behaviour, but it didn't work. Is there a proper way to do it?
I'm trying to create a table web-component with a pager at bottom. It should be extensible allowing the loading of data from a javascript array, a restful service or a custom source. Thus, I decided to create a behaviour for each one of these source and change it when the source changes. Is it a correct way to design it?
Here as example the source code of the behaviour to load data from an array. It has the following function:
itemsLoad: function(page, itemsPerPage, callback) {...
which is called from the web-component to load data of a specific page. My idea is that each behaviour based on the type of data source (e.g. CSV, JSON, etc.) will implement this method in a different way. Then, the behaviour will be registered at run-time, because is at run-time that the developers knows which is the source to use.
I don't think you will be able to change behaviours at run-time, because they are mixed into the element prototype.
What you can do is create a separate element for each of your cases (csv, json, etc) and create nodes dynamically as required. You could than place that element inside your grid
<table-component>
<json-data-source></json-data-source>
</table-component>
The <table-component> would look for a child element which implements itemsLoad to get the data.
EDIT
To work with child nodes you would use Polymer's DOM API. For example you could listen to added child nodes and select one that implements the itemsLoad method.
Polymer({
attached: function() {
Polymer.dom(this).observeNodes(function(info) {
var newNodes = info.addedNodes;
for(var i=0; i<newNodes.length; i++) {
var dataSource = newNodes[i];
if(dataSource.itemsLoad && typeof dataSource.itemsLoad === 'function') {
this.loadItems(dataSource);
break;
}
}
});
}
loadItems: function(dataSource) {
dataSource.itemsLoad().then(...);
}
});
You could replace Polymer.dom(this).observeNodes with simply iteration over Polymer.dom(this).children. Whichever works best for you.
Overview
DOM elements that are dynamically-rendered within dom-if, dom-repeat <templates> seem to be rendered asynchronously thus making unit-testing a bit of a pain.
The Polymer Component
template(is='dom-if', if='{{!defaultPrintAll}}')
template(is='dom-repeat', items='{{_pageBucketItems}}')
button(type='button', class$='{{_computeDefaultClass(item)}}', on-tap='_togglePageClick') {{item}}
The Test
test("Clicking 'Export All' to off, reveals board-selection tiles", function() {
$("#export-pdf-checkbox-all").siblings(".checkbox").trigger("click");
Polymer.dom.flush()
expect($(".board-range__button")).to.be.visible;
});
Why it seems to fail:
When clicking a button which triggers the dom-if/dom-repeat the elements don't render in a synchronous order.
The dom-if and it's subsequent/nested dom-repeat render asynchronously.
To make matters worse, the button itself get's it's class in a computed/binded manner (mind the class$= on the button).
So the question boils down to this:
Is it possible to force render the dom-if, dom-repeat, and the computed-binding of the class in a synchronous order after I simulate the click to the button which activates all 3 of those conditions?
Notes:
I'm using Polymer's official WCT as the test harness.
I'm also using chai-jquery.
I've also used Polymer.dom.flush() but it still doesn't, ahem.. flush.
I'm aware that I can use chai-as-promised.js instead but it adds unnecessary complexity to my tests for a trivial matter such as this, so I'd like to avoid it.
Rather than using Polymer.dom.flush(), try using the flush function that WCT puts on the window. This will enqueue a callback function to be executed, in theory, after the template has rendered.
test("Clicking 'Export All' to off, reveals board-selection tiles", function(done) {
$("#export-pdf-checkbox-all").siblings(".checkbox").trigger("click");
flush(function () {
expect($(".board-range__button")).to.be.visible;
done();
}
});
Important to notice: Asynchronous tests require the done function to be passed into the test callback, and require done to be called after your conditions have been evaluated.
bangin' my head against this and it's starting to hurt.
I'm having trouble with adding an event to an element.
I'm able to add the event, and then call it immediately with element.fireEvent('click'), but once the element is attached to the DOM, it does not react to the click.
example code:
var el = new Element('strong').setStyle('cursor','pointer');
el.addEvent('click',function () { alert('hi!'); });
el.replaces(old_element); // you can assume old_element exists
el.fireEvent('click'); // alert fires
however, once I attach this to the DOM, the element is not reactive to the click. styles stick (cursor is pointer when I mouseover), but no event fires. tried mouseover as well, to no avail.
any clues here? am I missing something basic? I am doing this all over the place, but in this one instance it doesn't work.
EDIT----------------
ok here's some more code. unfortunately I can't expose the real code, as it's for a project that is still under tight wraps.
basically, the nodes all get picked up as "replaceable", then the json found in the rel="" attribute sets the stage for what it should be replaced by. In this particular instance, the replaced element is a user name that should pop up some info when clicked.
again, if I fire the event directly after attaching it, all is good, but the element does not react to the click once it's attached.
HTML-----------
<p>Example: <span class='_mootpl_' rel="{'text':'foo','tag':'strong','event':'click','action':'MyAction','params':{'var1': 'val1','var2': 'val2'}}"></span></p>
JAVASCRIPT-----
assumptions:
1. below two functions are part of a larger class
2. ROOTELEMENT is set at initialize()
3. MyAction is defined before any parsing takes place (and is properly handled on the .fireEvent() test)
parseTemplate: function() {
this.ROOTELEMENT.getElements('span._mootpl_').each(function(el) {
var _c = JSON.decode(el.get('rel'));
var new_el = this.get_replace_element(_c); // sets up the base element
if (_c.hasOwnProperty('event')) {
new_el = this.attach_event(new_el, _c);
}
});
},
attach_event: function(el, _c) {
el.store(_c.event+'-action',_c.action);
el.store('params',_c.params);
el.addEvent(_c.event, function() {
eval(this.retrieve('click-action') + '(this);');
}).setStyle('cursor','pointer');
return el;
},
Works just fine. Test case: http://jsfiddle.net/2GX66/
debugging this is not easy when you lack content / DOM.
first - do you use event delegation or have event handlers on a parent / the parent element that do event.stop()?
if so, replace with event.preventDefault()
second thing to do. do not replace an element but put it somewhere else in the DOM - like document.body's first node and see if it works there.
if it does work elsewhere, see #1
though I realsie you said 'example code', you should write this as:
new Element('strong', {
styles: {
cursor: "pointer"
},
events: {
click: function(event) {
console.log("hi");
}
}
}).replaces(old_element);
no point in doing 3 separate statements and saving a reference if you are not going to reuse it. you really ought to show the ACTUAL code if you need advice, though. in this snippet you don't even set content text so the element won't show if it's inline. could it be a styling issue, what is the display on the element, inline? inline-block?
can you assign it a class that changes it on a :hover pseudo and see it do it? mind you, you say the cursor sticks which means you can mouseover it - hence css works. this also eliminates the possibility of having any element shims above it / transparent els that can prevent the event from bubbling.
finally. assign it an id in the making. assign the event to a parent element via:
parentEl.addEvent("click:relay(strong#idhere)", fn);
and see if it works that way (you need Element.delegate from mootools-more)
good luck, gotta love the weird problems - makes our job worth doing. it wouldn't be the worst thing to post a url or JSFIDDLE too...