Fast, tiny, standards-compliant XML DOM implementation for node and the browser
npm install slimdombat
npm install --save slimdom
`
or
`bat
yarn add slimdom
`
The package includes both a commonJS-compatible UMD bundle (dist/slimdom.umd.js) and an ES6 module (dist/slimdom.esm.js). This means it should work in most JavaScript environments that support the ES2017 standard or newer.
Usage
Create documents by parsing XML or start from scratch using the slimdom.Document constructor, and manipulate them using the [standard DOM API][domstandard].
`javascript
import * as slimdom from 'slimdom';
// alternatively, in node and other commonJS environments:
// const slimdom = require('slimdom');
// Parse from a string:
const document2 = slimdom.parseXmlDocument('Hello! ');
document2.documentElement.setAttribute('attr', 'new value');
const xml2 = slimdom.serializeToWellFormedString(document2);
// -> 'Hello! '
// Or start with an empty document:
const document = new slimdom.Document();
document.appendChild(document.createElementNS('http://www.example.com', 'root'));
const xml = slimdom.serializeToWellFormedString(document);
// -> '
`
Some DOM API's, such as the DocumentFragment constructor, require the presence of a global document, for instance to set their initial ownerDocument property. In these cases, slimdom will use the instance exposed through slimdom.document. Although you could mutate this document, it is recommended to always create your own documents (using the Document constructor) to avoid conflicts with other code using slimdom in your application.
When using a Range, make sure to call detach when you don't need it anymore. Unless you are only targeting environments that implement the WeakRef proposal, we do not have a way to detect when we can stop updating the range for mutations to the surrounding nodes. In environments that support WeakRef, calling detach is optional.
Features and limitations
This library implements:
- All node types: Attr, CDATASection, Comment, Document, DocumentFragment, DocumentType, Element, ProcessingInstruction, Text and XMLDocument.
- Range, which correctly updates under mutations.
- MutationObserver
- XMLSerializer, and read-only versions of innerHTML / outerHTML on Element.
- DOMParser, for XML parsing only.
This library is aimed at providing a lightweight and consistent experience for dealing with XML and XML-like data. For simplicity and efficiency, this implementation deviates from the spec in a few minor ways. Most notably, normal JavaScript arrays are used instead of HTMLCollection / NodeList and NamedNodeMap, HTML documents are treated no different from other documents and a number of features from in the DOM spec are missing. In most cases, this is because alternatives are available that can be used together with slimdom with minimal effort.
Do not rely on the behavior or presence of any methods and properties not specified in the DOM standard. For example, do not use JavaScript array methods exposed on properties that should expose a NodeList and do not use Element as a constructor. This behavior is _not_ considered public API and may change without warning in a future release.
This library implements the changes from [whatwg/dom#819][dom-adopt-pr], as the DOM specification as currently described has known bugs around adoption. It also deviates from the [DOM serialization algorithms][domparsing] that deal with assigning and resolving conflicts in namespace prefixes, as the specification has a number of bugs that currently remain unaddressed.
As serializing XML using the XMLSerializer does not enforce well-formedness, you may instead want to use the serializeToWellFormedString function which does perform such checks.
$3
The DOMParser interface is implemented, but this can only be used to parse XML. You may want to use the parseXmlDocument function instead, which throws when parsing fails - the spec requires DOMParser to generate and return an error document in those cases.
The XML parser is non-validating, but does check for well-formedness. It does not support an external DTD or external parsed entities, but does check any internal DTD for syntactic errors. During parsing, any referenced entities are included, default attribute values are materialized and the DTD internal subset is discarded. References to external entities are replaced with nothing. References to parameter entities are ignored.
The parseXmlFragment function may be used to parse fragments of XML. This function accepts the same format as specified for external parsed entities, except that it does not support parameter entities. That means it accepts an optional text declaration (similar to the XML version declaration) followed by any content that may be found between an element's start and end tags. That does not include doctype nodes. An optional resolver function may be provided to resolve any namespace prefixes not defined in the fragment itself.
`javascript
// Parse a fragment of XML with missing namespace declarations
const fragment = slimdom.parseXmlFragment(
,
{ resolveNamespacePrefix: (prefix) => resolved-${prefix} }
);
const xml3 = slimdom.serializeToWellFormedString(fragment);
// ->
`
This library does not implement HTML parsing, which means no insertAdjacentHTML on Element, nor createContextualFragment on Range. The innerHTML and outerHTML properties are read-only. If you need to parse HTML, see [this example][parse5-example] which shows how to connect the [parse5][parse5] HTML parser with the help of the [dom-treeadapter][dom-treeadapter] library.
To guard against [entity expansion attacks][billion-laughs], the parser by default limits how much entity expansion is allowed to increase the document size. To avoid blocking legitimate uses of entity expansion, this limit is only enforced after a certain output size threshold is reached. You can adjust these values by setting the entityExpansionMaxAmplification and entityExpansionThreshold options when calling parseXmlDocument. Please [open an issue][new-issue] if you ever need to increase these values for a non-attack input.
$3
This library does not implement CSS selectors, which means no querySelector / querySelectorAll on ParentNode and no closest / matches / webkitMatchesSelector on Element. This library also does not implement XPath, which means no XPathResult / XPathExpression / XPathEvaluator interfaces and no createExpression / createNSResolver / evaluate on Document.
To query a slimdom document using XPath or XQuery, use [FontoXPath][fontoxpath].
To query a slimdom document using CSS, see [this example][sizzle-example] which shows how to use [sizzle][sizzle] to run queries using CSS selectors.
$3
Emulating a full browser environment is not the goal of this library. Consider using [jsdom][jsdom] instead if you need that.
This implementation offers no special treatment of HTML documents, which means there are no implementations of HTMLElement and its subclasses. This also affects HTML-specific casing behavior for attributes and tagNames. The id / className / classList properties on Element and compatMode / contentType on Document have not been implemented. HTML-specific query methods (getElementById for interface NonElementParentNode, getElementsByClassName on Document and Element) are also missing.
This library does not currently implement events, including the Event / EventTarget interfaces. It also currently does not contain an implementation of AbortController / AbortSignal. As these may have wider applications than browser-specific use cases, please file an issue if you have a use for these in your application and would like support for them to be added.
There is currently no support for shadow DOM, so no Slottable / ShadowRoot interfaces and no slot / attachShadow / shadowRoot on Element. Slimdom also does not support the APIs for custom elements using the is option on createElement / createElementNS.
This library has no notion of URLs (baseURI on Node, and URL / documentURI on Document), nor of encodings (characterSet / charset / inputEncoding on Document). This library only deals with JavaScript strings, not raw byte streams.
This library omits properties and methods that exist mainly for web compatibility reasons (insertAdjacentElement / insertAdjacentText on Element, hasFeature on DOMImplementation, specified on Attr, the XSLTProcessor interface). This also includes all interfaces and interface members listed as historical / removed in the [DOM living standard][domstandard].
$3
The following features are missing simply because I have not yet had, or heard of, a need for them. If you do need one of these, feel free to create a feature request issue or even submit a pull request.
- Iteration helpers (NodeIterator / TreeWalker / NodeFilter, and the createNodeIterator / createTreeWalker methods on Document).
- attributeFilter for mutation observers.
- isConnected / getRootNode / isEqualNode / isSameNode on Node
[billion-laughs]: https://en.wikipedia.org/wiki/Billion_laughs_attack
[dom-adopt-pr]: https://github.com/whatwg/dom/pull/819
[dom-treeadapter]: https://github.com/RReverser/dom-treeadapter
[fontoxpath]: https://github.com/FontoXML/fontoxpath/
[jsdom]: https://github.com/jsdom/jsdom
[new-issue]: https://github.com/bwrrp/slimdom.js/issues/new
[parse5-example]: https://github.com/bwrrp/slimdom.js/tree/main/test/examples/parse5
[parse5]: https://github.com/inikulin/parse5
[sizzle-example]: https://github.com/bwrrp/slimdom.js/tree/master/test/examples/sizzle
[sizzle]: https://github.com/jquery/sizzle
Contributing
Pull requests for missing features or tests, bug reports, questions and other feedback are always welcome! Just [open an issue][new-issue] on the github repo, and provide as much detail as you can.
To work on the slimdom library itself, clone the repository and run npm install to install its dependencies.
The slimdom library and tests are developed in TypeScript, using prettier to automate formatting.
This repository includes a full suite of tests based on jest. Run npm test to run the tests, or npm run test:debug to debug the tests and code by disabling coverage and enabling the node inspector (see chrome://inspect in Chrome).
A runner for the W3C XML Conformance Test Suites is included in the test/dom-parsing/xmlConformance.tests.ts file. These tests are used to check that the parser conforms to the specificiations. To run these tests, first execute npm run download-xmlconf to download the necessary files to the temp/xmlconf directory. The tests will then be included automatically when running npm test. Tests for documents that are supposed to be rejected use Jest's snapshots feature to guard against unintentional changes to the errors produced. You may need to update these snapshots when the format of errors changes by running npm test -- -u.
An experimental runner for the W3C web platform tests is included in the test/web-platform-tests directory temporarily unavailable due to the migration to jest. To use it (when re-enabled), clone the web platform tests repository somewhere and set the WEB_PLATFORM_TESTS_PATH environment variable to the corresponding path. Then run npm test as normal. The webPlatform.tests.ts` file contains a blacklist of tests that don't currently run due to missing features.