yaml

• Add a "./": "./" export target
• Add fallback export targets for Node 13.0-13.6

New Features

• Add ES module exports + set type in package.json (Fixes #138)
• Add custom identifiers for core tags
• Add sortMapEntries option (#44)
• Use ES modules in browser build (supported by all modern bundlers and other packagers)

Bugfixes

• Create missing intermediate collections in addIn & setIn (Fixes #131)
• Account for initial indentation when stringifying values (Fixes #133)
• Use a Map for prevObjects in createNode (#137)
• YAML.createNode sets tag in the result, if explicitly given
• Drop Node 6 tests due to incompatibility with Jest 25 (transpilation remains as Node 6)
• cst: Minor refactoring to improve legibility, add some comments
• cst: Drop Node.atCollectionItem as unused since 93a8a440
• cst: Drop Directive.endOfDirective as unused since 307df17e
• Clean up YAML 1.1 & JSON boolean tag objects
• Add tests to improve code coverage
• Update dependencies

• During schema.createNode, accept Nodes as input (#128)
• Update dependencies

• Empty collection items only grab inline comments (#125, #126)
• Set CST Node#context as non-enumerable, to simplify logging

New Features

• Add optional id, customTags arguments to Document#setSchema()
• Keep explicit keys with block scalar values as such
• Add simpleKeys option (#122)

Improved Errors & Warnings

• Add more context to pretty error messages
• Improve error for missing flow collection terminator char
• Use offset when prettifying flow collection character errors
• Add warning when stringifying map/seq-valued keys for JS Object
• Use process.emitWarning rather than console.warn in YAML.parse if possible

Bugfixes

• Fix aliased anchors on null-value nodes (6KGN)
• Fix stringification of timestamp-like strings in a YAML 1.1 context (fixes #119)
• Consider input ending after implicit map key : as valid (Fixes #120)

• Allow same-column end for flow collections (#108, #113, #114)
• Add Document#directivesEndMarker (#112)
• Add directives-end and document-end marker ranges to CST Document during parse

• Apply previous document directives in YAML 1.1 streams (#109)
• Fix export of Type from "yaml/util"
• Update dependencies

New Features

• Refactor exports (#99, #107)
  • In addition to main export, use only yaml/parse-cst, yaml/types and yaml/util
  • Drop public export of custom tag objects
  • Rename tags option as customTags, and allow/prefer string identifiers for built-in custom tags
  • Add warnings to deprecated endpoints (set _YAML_SILENCE_DEPRECATION_WARNINGS to disable)
  • Drop babel-plugin-add-module-exports, using instead custom untranspiled CommonJS files for public exports
• Add protection for exponential entity expansion attacks, such as Billion laughs and quadratic expansion attacks (#104)
  • Add maxAliasCount option, defaulting to max 100 aliases for any anchor
  • For Object, stringify mapping key as YAML rather than JSON
• Simplify custom tag creation
  • Add default stringifier (item, ...) => item.toString(...) for collections
  • Add Schema#createPair(key, value, ctx); drop create{Map,Seq} exports
  • Improve & expand documentation on custom tags
• Preserve float formatting when parsed to Document (#97)
  • Add optional minFractionDigits property to numbers
  • Add EXP format handling for float numbers
  • Allow scalar resolvers to return a Scalar instance
• Add prettyErrors option, which makes errors prettier and drops their source reference (#96)
• Add common ancestor YAMLError for error classes
• Reorganise files, e.g. separating tags/ from schema/ and adding constants.js

Bugfixes

• Tag presence in AST node now corresponds to explicit tags in YAML source (#97)
• Add Pair#addToJSMap(ctx, map), fixing merge of anchor seq with mapAsMap: true

New Features

• Add @babel/runtime as a dependency for use via "browser" (#83)
• Support circular objects, representing circular references using YAML alias nodes when encountered during YAML.stringify() (#84)
• Deprecate the class attribute of tags, replacing it with the more generic identify(value) (#86)
• Turn the string stringifier into a generic scalar stringifier, fallback to it by default, and export it as stringify() from yaml/schema (#86)
• Add yaml-playground as a git submodule, extending ESLint rules
• Add start script, providing an interactive node env with YAML defined

Bugfixes

• Drop overeager browser polyfill generation; use playground/src/polyfill.js instead
• Add BrowserStack tests to Travis CI
• Accept arrays of tag objects as values of tags option (#90)
• Allow for blank lines after less-indented comments (#91)

• Prevent clashes between the namespaces of top-level Nodes and their values (82ac5a8)
• Stringify pairs with single-line flow collection values on a single line (#85)
• Fix stringification of objects set in collection values (#87)

• Use one- rather than zero-based indexing for line/col positions (f117880)

New Features

• Add context.root and rangeAsLinePos to CST nodes (#67, #68)
• Add collection accessor methods (#74)
  • New methods: add/delete/get/has/set and addIn/deleteIn/getIn/hasIn/setIn
  • Available on all collections as well as documents containing a collection
• Include a separate browser build in the release (#73)
  • Used by Webpack, Rollup, etc. via the "browser" value in package.json
  • Uses a Browserslist config of > 0.5%, not dead to (currently) support at least IE 11
  • Has a dependency on @babel/runtime that is not included in the package.json dependencies
• Optimise the Node.js release to leave out unnecessary polyfills (87beb89)
• Run CI tests on latest, LTS & 6.0.0 versions of Node.js (#70, #73)
• Move & reorganise test files under tests/ (12876df)
• Add .editorconfig (#69)

Bugfixes

• Add missing tag value to !!set (#78)
• Improve tag matcher test to allow supporting Symbols (#79)

• Keep empty line in comment lost during parsing (#75, #76)

New Features

• Support blank lines (#64)
  • Add BLANK_LINE nodes to the CST
  • Add a boolean property spaceBefore for all AST nodes
  • cst: Assign comments & blank lines at end of collection to parent unless sufficiently indented
  • Always separate document comments from the rest by a blank line
• Add missing YAML 1.1 collections: !!omap, !!pairs, and !!set (#65)
• Add tag argument for createNode(), along with its refactored schema-specific implementation (29f399b)
• Add optional createNode and nodeClass properties to tag objects (bcb8d9c, 3ae63bc)
• Add property based tests using fast-check, thanks to @dubzzz (#62)

Bugfixes

• Fix parsing of minimal strings: '', '-', '?', ':' (#61)
• Revert earlier fix for #48, which is now properly handled thanks to the fix for #57 (#59)
• cst/ParseContext: Stop advancing node.range.end past newline (08571ff)
• cst/PlainValue: Limit valueRange.end to actual value end (0001333)
• cst: Stop breaking the spec by allowing tab indents for map values (572f540)
• Stop aliasing Pair comment to the comment of its value (9b49ddb, 6d3cb36):

  key: # this is now a pair comment
    value
  

• Render maps with all null values as ? a\n? b\n or { a, b } (40b7a2a)

New Features

• Support for JavaScript Set, Map and other iterable objects. When stringifying or creating nodes, this is automatic; Map and its descendants turn into YAML mappings, while other iterables become sequences. When parsing, a new option mapAsMap creates Maps rather than Objects from YAML mappings. (#46, #49, #50, #51)
• Add a new public export at 'yaml/schema', providing easier-to-use stringifier options (#58)

Bugfixes

• Stop including end comments in collections if next content is not sufficiently indented (#41, #54)
• Stop folding more-indented lines at start of folded block scalars (#55)
• cst/PlainValue: Allow :, to start a plain value in flow-out (#56)
• Only advance to end of escape during double-quoted scalar folding (#57)

• Stop splitting \ escapes in double-quoted strings across line endings (#48)
• Quote plain strings consisting of just - or ? (#52)

• fix(setOrigRanges): add missing props (#45)

• Bugfix: Doesn't quote colons #43
• Update dev dependencies

Whoa. It's like, ready? Just one minor bug has been reported in the last month, so now's as good a time as any to officially release this thing. A few small changes compared to the last RC:

• Drop the need for .default when using require('yaml')
• Fix for the bug part of #28, now always putting collection comments on their own line
• Fix setOrigRanges() for flow collections, adding origOffset to FlowChar objects (Fixes #40)

More eyes reveal more bugs. Many of which have been fixed, thanks in particular to @ikatyang:

• schema/parseMap: Don't drop comments on empty value nodes (#19)
• cst/CollectionItem: Don't reset offset at end of document (#19)
• fix(document): add missing cstNode (#26)
• fix(document): add missing type (#27)
• fix(merge): no error when key is null (#29)
• fix(cst): correct document.valueRange.end (#30)
• Add setOrigRanges() to YAML.parseCST() output (#31)
• fix(flowSeq): explicit key with empty value (#33)
• fix: empty blockValue as empty string (#35)
• fix: checkKeyLength false positive (#37)
• cst/Collection: Set atLineStart during lineStart reset (#38)

In addition to the above bugfixes, the opinions of the string stringifier have been updated (2ab3eab):

• Quote "x" as '"x"'
• Prefer quoted over block for single line values
• Prefer block style for multiline
• Prefer literal over folded block style for short lines

Finally, dev dependencies have been updated, including an update to Babel 7 rc.2 that drops @babel/preset-stage-3 in favor of the specific required plugins.

As the library is now passing practically all of the yaml-test-suite tests, along with any others I've been able to find, and there are no more required features that are missing, this might be the last release before 1.0.0. Therefore, this version has been released on npm with the latest tag, so from now on the install command is just:

npm install yaml

In terms of features, rc.7 has rather little to offer on top of rc.6:

• parseCST is now also included in the default YAML export
• Add a test harness for yaml-test-suite event tests, yaml/dist/test-events
• Add a new default-false option, keepCstNodes
• Bugfix: Encode & decode URI escapes in tag suffixes

• Add support for YAML 1.0 and YAML 1.1, renaming extended schema as yaml-1.1 (note: YAML 1.0 prefix tags are parsed and stringified, but are not resolved)
• Refactor the API a bit again, adding parseDocument & renaming parseDocuments as parseAllDocuments. This more closely matches the naming pattern used by other libraries.
• Add type and format to content nodes, allowing for stringification control.
• Add keepBlobsInJSON and keepNodeTypes options, both defaulting to true.
• Split default options in YAML.defaultOptions and YAML.Document.defaults, depending on YAML version dependence
• Update to latest yaml-test-suite (now only four tests are broken in the upstream repo: yaml/yaml-test-suite#25, yaml/yaml-test-suite#27, yaml/yaml-test-suite#38, yaml/yaml-test-suite#39)
• Various minor bugfixes

• Add doc.anchors
• Add Alias and Merge nodes to AST (#12)

• Rename lowest API level more appropriately as a CST parser rather than an AST parser, so you'll now need to import yaml/parse-cst rather than yaml/parse-ast to use it.
• Simplify the object-level API, dropping Node#origTag, skip setting empty values, and letting Pair extend Node.
• Document the YAML.Document contents more clearly as an AST.

• Refactor Document#resolveValue() into YAML.createNode().
• Drop the internal doc reference from collections, and otherwise simplify the Map and Seq objects by separating out the parsing functions -- this should also make it easier to move or copy a collection between documents. Subsequently, directly calling .toString() on a Map or Seq will now get you a JSON representation of the object; for YAML output you'll need to stringify the whole document.
• Bug fixes, including in particular #18 (reported by @zoroaster00).

Switched from beta to rc because this really is nearly ready to publish.

• Documentation is now at https://eemeli.org/yaml/, rather than partly in README files and partly missing.
• Rethink semantic/syntax error boundary: Instead of classifying based on source API level, syntax errors now result in broken data while semantic errors "only" affect metadata.
• Add public re-exports: map, pair, parse-ast, scalar, seq, types/binary, types/timestamp
• Specify node ≥ 6 dependency (PR #11)
• Please ignore rc.1, it was missing a minor package.json update

Various bugfixes, in particular #7, #8, #9 and #10 as reported by @ikatyang.

Getting close to a v1 release...

• Actual documentation is developing at http://eemeli.org/yaml/
• Finalise (read: refactor) much of the public API
  • Document#contents can be set to anything
  • Add Document#resolveValue() to deeply wrap values in YAML objects
  • Move warnings from Document#errors to Document#warnings
  • Add YAML.stringify() and YAML.defaultOptions
  • Replace "stream" with "documents" in API
  • Add YAMLSemanticError class for Document#parse() errors
• String(Document) output improvements
  • tags are properly included, including prefixes
  • output is prettily folded where possible & appropriate
• Fix various corner-case syntax bugs, with thanks in particular to @ikatyang

• Comments mostly work
• Tags are included in resolved nodes
• Consider comments at start without directives-end indicator to be part of the document contents, not directives

• Merge raw-yaml@0.2.5 as src/ast
• Fix remaining error-reporting mismatches

• Fix remaining stringifier mismatches with yaml-test-suite in.json
• Fix mismatches with yaml-test-suite out.yaml