commonmark.js 0.19.0

• Fixed underscore emphasis to conform to spec changes (commonmark/commonmark-spec#317, #12).
• Fixed variable shadowing (linter warnings) in test.js.
• Makefile: Moved lint first to avoid regenerating dist by accident.
• Dingus: Fixed word wrapping in text area (commonmark/commonmark-spec#319).
• Updated spec.txt in test/.
• Added release_checklist.md.

commonmark.js 0.18.2

• Fixed walker so that it stops at this.root, even when this is
  not the document root.
• Fixed typo (Ed Adams).
• Removed unused argument from processEmphasis (Robin Stocker).
• Shortened code for removing a sequence of delimiters, changing the
  links on the edges rather than dropping them one by one
  (Eli Barzilay).
• Code layout improvements for readability (Eli Barzilay).
• Dingus: rewrote without ACE editor for better accessibility
  and to fix layout in smal windows (#11).
• Optimize checking for final spaces in parseNewline (Robin
  Stocker). This yields a 5-10% speed improvement in make bench.
• Consolidated regex definitions (Robin Stocker).
• Return boolean from spnl and parseBackticks for consistency
  (Robin Stocker).
• Removed unnecessary check for hrule in parseListMarker (Robin
  Stocker).
• Return error status if tests fail.
• Use encode/decode from markdown-it/mdurl for URL
  normalization. This fixes cases like [link](http://google.com/?q=%3f)
  which was formerly wrongly converted to
  <a href="http://google.com/?q=%253f">link</a> (#9).
• Refactored test.js, adding smart_punct tests from cmark, and
  factoring out specTest function.
• Make unmatched double quotes left quotes with smart_punct (#14).
• Code cleanup in finalize.

commonmark.js 0.18.1

• Added nextNonspace, blank, indent properties to Parser.
• Factored out advance-to-nonspace code into findNextNonspace.
• Put block starts into blockStarts property of Parser. Now
  almost all block-specific material has been moved out of
  incorporateLine.
• Removed parameter from addLine.
• Simplified code, removed matchAt.
• Small optimization in entityToChar
• Use charCodeAt in a few places where charAt was used.
• Added peek() function in blocks. This ensures we check bounds
  before calling charCodeAt.
• Use peek instead of raw charCodeAt in inlines.js.
• Dingus improvements:
  • Use ACE editor in dingus. Implemented L->R sync scroll. Also,
    the block containing the cursor is highlighted on the right.
  • Split code from dingus.html into separate file dingus.js.
  • Split CSS into separate file, dingus.css.
  • Added debounce.
  • Moved dingus code to dingus/ directory. dingus/Makefile
    builds commonmark.js in the dingus directory. So to deploy,
    you just need to copy the dingus directory.
  • Use local copies of js dependencies. Instead of including these in
    the repository, though, we have the Makefile install them with bower.
  • Fixed one-off error in line number calculation.
  • XML and HTML are now same font size.
  • Allow line wrapping in editor.
  • AST window now scrolls like HTML window.
• Updated reNonSpace for new whitespace definition in spec.
• Added smart punctuation parsing option.
  • Added options param to InlineParser. options.smart triggers
    smart punctuation parsing.
  • Added --smart option to bin/commonmark.
  • Implemented "smart" parsing of dashes, ellipses, quotes. The
    algorithm is the same as in cmark.
  • Renamed handleEmphasis -> handleDelims.
  • Added benchmark with smart.
  • Added checkbox for smart to dingus.
• Fixed code example for walk.
• Made isContainer a getter to conform to README (#7).
• Allow list start number of 0 in HTML writer (#10).
• Fixed use of isContainer in XML writer.
  isContainer is now a getter, not a function.

commonmark 0.17.1

• Reorganized block parsing in a more modular way. There is now a blocks property of the parser that contains information about each type of block, which is used in parsing. Ultimately this will make it easier to extend the library, but the project is still only partially completed.
• Code cleanup and simplification, with some performance optimizations.
• Removed version from bower.json. Bower takes version from tags.
• Initialize some properties at beginning of 'parse'. This fixes some mistakes in source position when the same Parser object was used to parse multiple times (#3).
• Made parsing of backslash escapes a bit more efficient.
• Removed refmap parameter of InlineParser.parse(). Instead, set the refmap property before running the parser.
• Set _string_content to null after using, allowing it to be GCd.
• Removed _strings; just append to _string_content. This gives better performance with v8.
• Format benchmarks so that samples are linked.
• Added in-browser benchmark.
• Added API documentation to README.
• xml renderer: use sourcepos attribute, not data-sourcepos.
• Changed license to 2-clause BSD. Added clause for spec.

commonmark 0.17

• Renamed DocParser -> Parser. Note: library users should update their code or it will break.
• Added normalize-reference.js. This does a proper unicode case fold instead of just using toUpperCase. It is also faster, partly because we can do one pass for space and case normalization.
• Removed artificial distinction btw FencedCode, IndentedCode in blocks.js.
• Removed vestigial ReferenceDef node type.
• Added getters and (in some cases) setters for "public" properties of Nodes. Renamed non-public properties to start with underscore. This will allow us to keep the API stable while changing the underlying data structure. And it will avoid exposing properties that have only an instrumental value in parsing.
• Removed Node.toObject().
• Rename bullet_char -> bulletChar.
• Check for blank line before checking indent in Item.
• Removed unnecessary setting of default tight=true in finalize. We do that when the listData object is initialized.
• Performance optimization - avoid repeating scan for nonspace.
• Moved check for closing fence to close-block-check section. This is a more logical arrangement and addresses commonmark/commonmark-spec#285.
• Added offset property to DocParser. Use this in addLine, instead of offset parameter, which has been removed.
• Implemented new spec for emphasis and strong emphasis with _.
• html.js - explicitly specify second parameter of escapeXml.
• Fixed escaping error in CDATA regex.
• Removed some dead code and fixed incorrect call to addChild with three arguments (Robin Stocker).
• Adjust lastLineLength before returning after close fence.
• Propagate lastLineBlank up through parents. Previously we just kept it set on the bottom child. But this will give a quicker determination of lastLineBlank.
• Moved continuation checks & finalizers into blocks property of Parser. This is a first step towards keeping the code for each kind of block in a central place, rather than spread all over the code base. This is preparatory for a more modular structure, where each type of block has a record describing how it is parsed and finalized. Eventually this will also contain functions for checking for a block start, and metadata that determines how line data should be handled.
• Added currentLine property to Parser.
• Renamed first_nonspace -> next_nonspace.
• Put generated commonmark.js in dist/ rather than js/.
• Miscellaneous code cleanup.
• Split JS code into (this) independent repository.
• Added detailed benchmark with samples (make bench-detailed).
• Added dist/commonmark.js to repo (for bower).
• Added bower.json (commonmark/commonmark-spec#288).
• Updated test suite. Now shows how performance depends on length in pathological cases.
• Don't use -1 as second param for .slice. This seems to cause a deoptimization, as reported by node --trace-deopt.
• Added CONTRIBUTING.md.
• Added .travis.yml to test against various node versions.
• Renamed changelog.js.txt -> changelog.txt.