Release 0.9.x

[karfau]

All Commits

0.9.2

Feature

• add Element.getElementsByClassName #722

Fixed

• add missing types for Document.documentElement and Element.tagName #721 #720

Thank you, @censujiang, @Mathias-S, for your contributions

0.9.1

Fixed

• DOMParser.parseFromString requires mimeType as second argument #713
• correct spelling of isHTMLMimeType in type definition #715 / #712
• sync types with exports #717 / #285 / #695

Other

• minimum tested node version is 14 #710

Thank you, @krystofwoldrich, @marvinruder,
@amacneil, @defunctzombie,
@tjhorner, @danon, for your contributions

0.9.0

Summary on dev.to

Features

• feat: expose all DOM level 2 element prototypes #637 / #40
• feat: add iterator function to NodeList and NamedNodeMap #634 / #633

Fixed

• parse empty/whitspace only doctype internal subset #692
• avoid prototype clash in namespace prefix #554
• report fatalError when doctype is inside elements #550

Other

• test: add fuzz target and regression tests #556
• chore: improve .gitignore and provide .envrc.template #697
• chore: Apply security best practices #546
• ci: check test coverage in PRs #524
• docs: add missing commas to readme #566
• docs: click to copy install command in readme #644
• docs: enhance jsdoc comments #511

Thank you, @kboshold, @edi9999, @apupier, @shunkica, @homer0, @jhauga, @UdayKharatmol, for your contributions

0.9.0-beta.11

Fixed

• report more non well-formed cases #519 / #45 / #125 / #467
  BREAKING-CHANGE: Reports more not well-formed documents as fatalError
  and drop broken support for optional and unclosed tags in HTML.

Other

• Translate/drop non English comments #518
• use node v16 for development #517

Thank you, @brodybits, @cbettinger, @josecarlosrx, for your contributions

0.9.0-beta.10

Fixed

• dom: prevent iteration over deleted items #514/ #499

Chore

• use prettier plugin for jsdoc #513

Thank you, @qtow, @shunkica, @homer0, for your contributions

0.9.0-beta.9

Fixed

• Set nodeName property in ProcessingInstruction #509 / #505
• preserve DOCTYPE internal subset #498 / #497 / #117
  BREAKING CHANGES: Many documents that were previously accepted by xmldom, esecially non well-formed ones are no longer accepted. Some issues that were formerly reported as errors are now a fatalError.
• DOMParser: Align parseFromString errors with specs #454

Chore

• stop running mutation tests using stryker #496
• make toErrorSnapshot windows compatible #503

Thank you, @cjbarth, @shunkica, @pmahend1, @niklasl, for your contributions

0.9.0-beta.8

Fixed

• Throw DOMException when calling removeChild with invalid parameter #494 / #135

BREAKING CHANGE: Previously it was possible (but not documented) to call Node.removeChild with any node in the tree,
and with certain exceptions, it would work. This is no longer the case: calling Node.removeChild with an argument that is not a direct child of the node that it is called from, will throw a NotFoundError DOMException, as it is described by the specs.

Thank you, @noseworthy, @davidmc24, for your contributions

0.9.0-beta.7

Feature

• Add compareDocumentPosition method from level 3 spec. #488

Fixed

• getAttribute and getAttributeNS should return null (fix: getAttribute and getAttributeNS should return null #477) #46
• several issues in NamedNodeMap and Element (fix: several issues in NamedNodeMap and Element #482) #46
• properly parse closing where the last attribute has no value #485 / #486
• extend list of HTML entities #489

BREAKING CHANGE: Iteration over attributes now happens in the right order and non-existing attributes now return null instead of undefined. THe same is true for the namepsaceURI and prefix of Attr nodes.
All of the changes are fixing misalignment with the DOM specs, so if you expected it to work as specified,
nothing should break for you.

Chore

• update multiple devDependencies
• Configure jest (correctly) and wallaby #481 / #483

Thank you, @bulandent, @zorkow, for your contributions

0.9.0-beta.6

Fixed

• Properly check nodes before replacement #457 / #455 / #456

Thank you, @edemaine, @pedro-l9, for your contributions

0.9.0-beta.5

Fixed

• fix: Restore ES5 compatibility #452 / #453

Thank you, @fengxinming, for your contributions

0.9.0-beta.4

Fixed

• Security: Prevent inserting DOM nodes when they are not well-formed CVE-2022-39353
  In case such a DOM would be created, the part that is not well-formed will be transformed into text nodes, in which xml specific characters like < and > are encoded accordingly.
  In the upcoming version 0.9.0 those text nodes will no longer be added and an error will be thrown instead.
  This change can break your code, if you relied on this behavior, e.g. multiple root elements in the past. We consider it more important to align with the specs that we want to be aligned with, considering the potential security issues that might derive from people not being aware of the difference in behavior.
  Related Spec: https://dom.spec.whatwg.org/#concept-node-ensure-pre-insertion-validity

Chore

• update multiple devDependencies
• Add eslint-plugin-node for lib #448 / #190
• style: Apply prettier to all code #447 / #29 / #130

Thank you, @XhmikosR, @awwright, @frumioj, @cjbarth, @markgollnick for your contributions

0.9.0-beta.3

Fixed

• fix: Stop adding tags after incomplete closing tag #445 / #416
  BREAKING CHANGE: It no longer reports an error when parsing HTML containing incomplete closing tags, to align the behavior with the one in the browser.
  BREAKING CHANGE: If your code relied on not well-formed XML to be parsed and include subsequent tags, this will no longer work.
• fix: Avoid bidirectional characters in source code #440

Other

• ci: Add CodeQL scan #444

Thank you, @ACN-kck, @mgerlach for your contributions

0.9.0-beta.2

Fixed

• Avoid iterating over prototype properties #437 / #436

Thank you, @Supraja9726 for your contributions

0.9.0-beta.1

Fixed

Only use HTML rules if mimeType matches #338, fixes #203

In the living specs for parsing XML and HTML, that this library is trying to implement,
there is a distinction between the different types of documents being parsed:
There are quite some rules that are different for parsing, constructing and serializing XML vs HTML documents.

So far xmldom was always "detecting" whether "the HTML rules should be applied" by looking at the current namespace. So from the first time an the HTML default namespace (http://www.w3.org/1999/xhtml) was found, every node was treated as being part of an HTML document. This misconception is the root cause for quite some reported bugs.

BREAKING CHANGE: HTML rules are no longer applied just because of the namespace, but require the mimeType argument passed to DOMParser.parseFromString(source, mimeType) to match 'text/html'. Doing so implies all rules for handling casing for tag and attribute names when parsing, creation of nodes and searching nodes.

BREAKING CHANGE: Correct the return type of DOMParser.parseFromString to Document | undefined. In case of parsing errors it was always possible that "the returned Document" has not been created. In case you are using Typescript you now need to handle those cases.

BREAKING CHANGE: The instance property DOMParser.options is no longer available, instead use the individual readonly property per option (assign, domHandler, errorHandler, normalizeLineEndings, locator, xmlns). Those also provides the default value if the option was not passed. The 'locator' option is now just a boolean (default remains true).

BREAKING CHANGE: The following methods no longer allow a (non spec compliant) boolean argument to toggle "HTML rules":

• XMLSerializer.serializeToString
• Node.toString
• Document.toString

The following interfaces have been implemented:
DOMImplementation now implements all methods defined in the DOM spec, but not all of the behavior is implemented (see docstring):

• createDocument creates an "XML Document" (prototype: Document, property type is 'xml')
• createHTMLDocument creates an "HTML Document" (type/prototype: Document, property type is 'html').
  • when no argument is passed or the first argument is a string, the basic nodes for an HTML structure are created, as specified
  • when the first argument is false no child nodes are created

Document now has two new readonly properties as specified in the DOM spec:

• contentType which is the mime-type that was used to create the document
• type which is either the string literal 'xml' or 'html'

MIME_TYPE (/lib/conventions.js):

• hasDefaultHTMLNamespace test if the provided string is one of the miem types that implies the default HTML namespace: text/html or application/xhtml+xml

Thank you @weiwu-zhang for your contributions

Chore

• update multiple devDependencies

[karfau]

All tickets related to the current milestone 0.9.0 have been closed.
There are plenty of breaking changes in this release and I will write a summary, but the main "topic" is that the code running controls the rules for parsing using the now mandatory mimeType argument for DOMParser.parseFromString and preserving that information when serializing, along with plenty of fixes for more correctly detecting (not) well-formed XML and throw a ParseError for way more cases. There are also simplifications towards error handling.

If nothing else comes up the current version will be released as 0.9.0 in the near future.