Skip to content
Permalink
lit-html-1.x
Switch branches/tags
Go to file
 
 
Cannot retrieve contributors at this time
/**
* @license
* Copyright (c) 2017 The Polymer Project Authors. All rights reserved.
* This code may only be used under the BSD style license found at
* http://polymer.github.io/LICENSE.txt
* The complete set of authors may be found at
* http://polymer.github.io/AUTHORS.txt
* The complete set of contributors may be found at
* http://polymer.github.io/CONTRIBUTORS.txt
* Code distributed by Google as part of the polymer project is also
* subject to an additional IP rights grant found at
* http://polymer.github.io/PATENTS.txt
*/
/**
* @module lit-html
*/
import {reparentNodes} from './dom.js';
import {TemplateProcessor} from './template-processor.js';
import {boundAttributeSuffix, lastAttributeNameRegex, marker, nodeMarker} from './template.js';
declare const trustedTypes: typeof window.trustedTypes;
/**
* Our TrustedTypePolicy for HTML which is declared using the html template
* tag function.
*
* That HTML is a developer-authored constant, and is parsed with innerHTML
* before any untrusted expressions have been mixed in. Therefor it is
* considered safe by construction.
*/
const policy = window.trustedTypes &&
trustedTypes!.createPolicy('lit-html', {createHTML: (s) => s});
const commentMarker = ` ${marker} `;
/**
* The return type of `html`, which holds a Template and the values from
* interpolated expressions.
*/
export class TemplateResult {
readonly strings: TemplateStringsArray;
readonly values: readonly unknown[];
readonly type: string;
readonly processor: TemplateProcessor;
constructor(
strings: TemplateStringsArray, values: readonly unknown[], type: string,
processor: TemplateProcessor) {
this.strings = strings;
this.values = values;
this.type = type;
this.processor = processor;
}
/**
* Returns a string of HTML used to create a `<template>` element.
*/
getHTML(): string {
const l = this.strings.length - 1;
let html = '';
let isCommentBinding = false;
for (let i = 0; i < l; i++) {
const s = this.strings[i];
// For each binding we want to determine the kind of marker to insert
// into the template source before it's parsed by the browser's HTML
// parser. The marker type is based on whether the expression is in an
// attribute, text, or comment position.
// * For node-position bindings we insert a comment with the marker
// sentinel as its text content, like <!--{{lit-guid}}-->.
// * For attribute bindings we insert just the marker sentinel for the
// first binding, so that we support unquoted attribute bindings.
// Subsequent bindings can use a comment marker because multi-binding
// attributes must be quoted.
// * For comment bindings we insert just the marker sentinel so we don't
// close the comment.
//
// The following code scans the template source, but is *not* an HTML
// parser. We don't need to track the tree structure of the HTML, only
// whether a binding is inside a comment, and if not, if it appears to be
// the first binding in an attribute.
const commentOpen = s.lastIndexOf('<!--');
// We're in comment position if we have a comment open with no following
// comment close. Because <-- can appear in an attribute value there can
// be false positives.
isCommentBinding = (commentOpen > -1 || isCommentBinding) &&
s.indexOf('-->', commentOpen + 1) === -1;
// Check to see if we have an attribute-like sequence preceding the
// expression. This can match "name=value" like structures in text,
// comments, and attribute values, so there can be false-positives.
const attributeMatch = lastAttributeNameRegex.exec(s);
if (attributeMatch === null) {
// We're only in this branch if we don't have a attribute-like
// preceding sequence. For comments, this guards against unusual
// attribute values like <div foo="<!--${'bar'}">. Cases like
// <!-- foo=${'bar'}--> are handled correctly in the attribute branch
// below.
html += s + (isCommentBinding ? commentMarker : nodeMarker);
} else {
// For attributes we use just a marker sentinel, and also append a
// $lit$ suffix to the name to opt-out of attribute-specific parsing
// that IE and Edge do for style and certain SVG attributes.
html += s.substr(0, attributeMatch.index) + attributeMatch[1] +
attributeMatch[2] + boundAttributeSuffix + attributeMatch[3] +
marker;
}
}
html += this.strings[l];
return html;
}
getTemplateElement(): HTMLTemplateElement {
const template = document.createElement('template');
let value = this.getHTML();
if (policy !== undefined) {
// this is secure because `this.strings` is a TemplateStringsArray.
// TODO: validate this when
// https://github.com/tc39/proposal-array-is-template-object is
// implemented.
value = policy.createHTML(value) as unknown as string;
}
template.innerHTML = value;
return template;
}
}
/**
* A TemplateResult for SVG fragments.
*
* This class wraps HTML in an `<svg>` tag in order to parse its contents in the
* SVG namespace, then modifies the template to remove the `<svg>` tag so that
* clones only container the original fragment.
*/
export class SVGTemplateResult extends TemplateResult {
getHTML(): string {
return `<svg>${super.getHTML()}</svg>`;
}
getTemplateElement(): HTMLTemplateElement {
const template = super.getTemplateElement();
const content = template.content;
const svgElement = content.firstChild!;
content.removeChild(svgElement);
reparentNodes(content, svgElement.firstChild);
return template;
}
}