3.0.6

Fixed a few more whitespace-related issues:

• Previously, fragments would be output with an extra trailing space. It is now removed.
• Previously, fragments and paragraphs would consume a newline ending them, removing it from the output. They now preserve it.

3.0.5

Fixed an issue where a single line break was not properly converted to a space when the beginning of the next line contained a formatting character. For example,

blah blah blah
~Atomics~ foo bar baz

used to become blah blah blah<emu-const>Atomics</emu-const> foo bar baz, but now becomes blah blah blah <emu-const>Atomics</emu-const> foo bar baz.

3.0.4

Fixed several bugs regarding paragraph break parsing:

• Parsing fragments that contained paragraph breaks would only emit the first paragraph. Now all paragraphs are joined with a space, i.e., the \n\n is collapsed into .
• Parsing documents with leading whitespace would cause odd behavior.
• Parsing documents that contain paragraph breaks would concatenate them all into a single paragraph. Now they properly split as separate paragraphs. (However, we still don't have full "Markdown-style" paragraph processing, in that if you start a line with a HTML tag, it will still always be wrapped in a paragraph.)

3.0.3

Fixed a bug where input with multiple consecutive format characters, like *__* or ||, would cause an exception.

3.0.2

Fix a bug (which seemed to only manifest on io.js 3.0.0, maybe?) giving "TypeError: Assignment to constant variable" when anything involving lookahead (i.e. tags and comments) was encountered.

3.0.1

Fixed comment parsing to be properly lazy. For example, foo<!-- -->bar<!-- -->baz now parses as two comments, instead of just one.

3.0.0

This release brought on @bterlson as a contributor, as he rewrote the parser to be hand-rolled instead of generated, adding much speed and flexiblity. Thanks to Brian for all of the changes listed here!

Changes to the Ecmarkdown language

• Added unordered lists.
• Added backslash escaping.
• Removed the string production (i.e. the auto-<code>-ification of strings).
• Changed the parsing of variable syntax to better match existing usage and conflict less with idiomatic underscore usage.
• Changed many edge-case behaviors, generally to be more sane.

Changes to the API

• Added a document export, which allows as input sequence of paragraphs and lists.
• Removed the list export.

Changes to the output

• The output is no longer beautified, as this was found to slow down performance by multiple orders of magnitude, and was probably unnecessary anyway since beautification is better done on a whole-document level.
• The exact positioning of HTML comments in the output has changed for a given input, e.g. they now appear inside <li>s instead of in between them.

2.0.2

Stop processing inside "HTML tags," i.e. < >-delimited runs of text.

2.0.1

Allows fragments to be empty, instead of throwing.

2.0.0

Removes support for paragraphs; use fragments at one extreme, or numeric lists at the other.

Allows HTML comments, potentially spanning multiple lines, inside of numeric lists.