xmlbuilder

• Removed support for node.js 6.0. Minimum required version is now 8.0.

• Fixed a bug where importing a document into an empty parent document did not set the root node. See #213.

• Corrected typings for doctype and character data nodes. See #211.

• Rewrote all TypeSript typings to be strictly compatible with the API. This is a breaking change for TypeScript users only.

• Added option for pretty printing attributes.

• Removed support for node.js 4.0 and 5.0. Minimum required version is now 6.0.

• Added TypeScript typings. See #200.

• Calling end() with arguments no longer overwrites writer options. See #120.

• Added writer state and customizable space and endline functions to help customize writer behavior. Also added openNode and closeNode functions to writer. See #193.

• Fixed a bug where writer functions would not be called for nodes with a single child node in pretty print mode. See #195.

• Renamed elEscape to textEscape in XMLStringifier.

• Fixed a bug where empty arrays would produce child nodes. See #190.

• Removed the skipNullAttributes option. null attributes are now skipped by default. Added the keepNullAttributes option in case someone needs the old behavior.

• Removed the skipNullNodes option. null nodes are now skipped by default. Added the keepNullNodes option in case someone needs the old behavior.

• undefined values are now skipped when converting JS objects.

• Renamed stringify functions. See #194:

  • eleName -> name
  • attName -> name
  • eleText -> text

• Fixed argument order for attribute function in the writer. See #196.

• Added openAttribute and closeAttribute functions to writer. See #196.

• Added node types to node objects. Node types and writer states are exported by the module with the nodeType and writerState properties.

• Fixed a bug where array items would not be correctly converted. See #159.

• Fixed a bug where mixed-content inside JS objects with #text decorator would not be correctly converted. See #171.

• Fixed a bug where JS objects would not be expanded in callback mode. See #173.

• Fixed a bug where character validation would not obey document's XML version. Added separate validation for XML 1.0 and XML 1.1 documents. See #169.

• Fixed a bug where names would not be validated according to the spec. See #49.

• Renamed text property to value in comment and cdata nodes to unify the API.

• Removed doctype function to prevent name clash with DOM implementation. Use the dtd function instead.

• Removed dummy nodes from the XML tree (Those were created while chain-building the tree).

• Renamed attributesproperty to attribs to prevent name clash with DOM property with the same name.

• Implemented the DOM standard (read-only) to support XPath lookups. XML namespaces are not currently supported. See #122.

• Fixed an edge case where a null node at root level would be printed although skipNullNodes was set. See #187.

• Added the skipNullNodes option to skip nodes with null values. See #158.

• Added current indentation level as a parameter to the onData function when in callback mode. See #125.
• Added name of the current node and parent node to error messages where possible. See #152. This has the potential to break code depending on the content of error messages.
• Fixed an issue where objects created with Object.create(null) created an error. See #176.
• Added test builds for node.js v8 and v10.

• Simplified regex used for validating encoding.

• spacebeforeslash writer option accepts true as well as space char(s).

• spacebeforeslash writer option can now be used with XML fragments.

• Added the spacebeforeslash writer option to add a space character before closing tags of empty elements. See #157.

• Fixed character validity checks to work with node.js 4.0 and 5.0. See #161.

• Removed case conversion options.
• Removed support for node.js 4.0 and 5.0. Minimum required version is now 6.0.
• Fixed valid char filter to use XML 1.1 instead of 1.0. See #147.
• Added options for negative indentation and suppressing pretty printing of text nodes. See #145.

• Falsy values can now be used as a text node in callback mode.

• Falsy values can now be used as a text node. See #117.

• Removed lodash dependency to keep the library small and simple. See #114, #53, and #43.
• Added title case to name conversion options.

• Added the callback option to the begin export function. When used with a callback function, the XML document will be generated in chunks and each chunk will be passed to the supplied function. In this mode, begin uses a different code path and the builder should use much less memory since the entire XML tree is not kept. There are a few drawbacks though. For example, traversing the document tree or adding attributes to a node after it is written is not possible. It is also not possible to remove nodes or attributes.

var result = '';

builder.begin(function(chunk) { result += chunk; })
  .dec()
  .ele('root')
    .ele('xmlbuilder').up()
  .end();

• Replaced native Object.assign with lodash.assign to support old JS engines. See #111.

• Added the begin export function. See the wiki for details.
• Added the ability to add comments and processing instructions before and after the root element. Added commentBefore, commentAfter, instructionBefore and instructionAfter functions for this purpose.
• Dropped support for old node.js releases. Minimum required node.js version is now 4.0.

• Processing instructions are now created as regular nodes. This is a major breaking change if you are using processing instructions. Previously processing instructions were inserted before their parent node. After this change processing instructions are appended to the children of the parent node. Note that it is not currently possible to insert processing instructions before or after the root element.

  root.ele('node').ins('pi');
  // pre-v7
  <?pi?><node/>
  // v7
  <node><?pi?></node>
  

• Added custom XML writers. The default string conversion functions are now collected under the XMLStringWriter class which can be accessed by the stringWriter(options) function exported by the module. An XMLStreamWriter is also added which outputs the XML document to a writable stream. A stream writer can be created by calling the streamWriter(stream, options) function exported by the module. Both classes are heavily customizable and the details are added to the wiki. It is also possible to write an XML writer from scratch and use it when calling end() on the XML document.

• Moved require statements for text case conversion to the top of files to reduce lazy requires.

• Added text case option for element names and attribute names. Valid cases are lower, upper, camel, kebab and snake.
• Attribute and element values are escaped according to the Canonical XML 1.0 specification. See #54 and #86.
• Added the allowEmpty option to end(). When this option is set, empty elements are not self-closed.
• Added support for nested CDATA. The triad ]]> in CDATA is now automatically replaced with ]]]]><![CDATA[>.

• Updated lodash dependency to 4.0.0.

• Added the noDoubleEncoding option to create() to control whether existing html entities are encoded.

• Added the separateArrayItems option to create() to control how arrays are handled when converting from objects. e.g.

root.ele({ number: [ "one", "two"  ]});
// with separateArrayItems: true
<number>
  <one/>
  <two/>
</number>
// with separateArrayItems: false
<number>one</number>
<number>two</number>

• Removed the #list decorator. Array items are now created as child nodes by default.
• Fixed a bug where the XML encoding string was checked partially.

• #list decorator ignores empty arrays.

• Allow \r, \n and \t in attribute values without escaping. See #86.

• Use native isArray instead of lodash.
• Indentation of processing instructions are set to the parent element's.

• Updated lodash dependency to 3.5.0.

• Bumped version because previous release was not published on npm.

• Updated lodash dependency to 3.5.0.

• Updated lodash dependency to 3.3.0.

• Fixed a bug where the XMLNode constructor overwrote the super class parent.
• Removed document property from cloned nodes.
• Switched to mocha.js for testing.

• Updated lodash dependency to 3.2.0.

• Updated lodash dependency to 3.1.0.
• Support all node >= 0.8.

• Updated lodash dependency to 3.0.0.

• Show more information from attribute creation with null values.
• Added iojs as an engine.
• Self close elements with empty text.

• Fixed prepublish script to run on windows.
• Fixed bug in XMLStringifier where an undefined value was used while reporting an invalid encoding value.
• Moved require statements to the top of files to reduce lazy requires. See #62.

• Added the offset option to toString() for use in XML fragments.

• Corrected license in package description.

• Dropped performance test and memwatch dependency.

• Fixed a bug where empty indent string was omitted when pretty printing. See #59.

• Correct cases of pubID and sysID.
• Use single lodash instead of separate npm modules. See #53.
• Escape according to Canonical XML 1.0. See #54.

• Convert objects to JS primitives while sanitizing user input.
• Object builder preserves items with null values. See #44.
• Use modularized lodash functions to cut down dependencies.
• Process empty objects when converting from objects so that we don't throw on empty child objects.

• Bumped version because previous release was not published on npm.

• Removed duplicate null checks from constructors.
• Fixed node count in performance test.
• Added option for skipping null attribute values. See #26.
• Allow multiple values in att() and ins().
• Added ability to run individual performance tests.
• Added flag for ignoring decorator strings.

• Removed performance tests from npm package.

• Combined loops for speed up.
• Added support for the DTD and XML declaration.
• clone includes attributes.
• Added performance tests.
• Evaluate attribute value if function.
• Evaluate instruction value if function.

• Changed processing instruction decorator to ?.

• Added processing instructions to JS object conversion.