* use `[]=` instead of add() to add JSON nodes to an object.
add() seems to have problems with expanding.
* renamed bench/json.nim to bench/jsonBench.nim to avoid confusion
with Nim's stdlib json.nim
* Added PresentationOptions object; merged PresentationStyle,
indentationStep and NewLineStyle in it
* Updated signatures of relevant procs
* Implemented user-defined newline style in presenter
* Added serialization.txt to doc
* Fixed rendering problem in api.txt
* Use explicit tag when ref type renders to a scalar that can be
parsed to !!null
* Added test cases for this ref type fix
* renamed serialize to serializeObject and construct to constructObject
* added new serialize() and construct() procs that handle document
start & end
* re-raise AssertionError instead of swallowing it
* Some typo fixes in code
* nim documentation now generates API doc with source links
* presenter now uses "" if a string starts with @ or `
* Removed some outdated stuff
* Added raises pragma to every proc in yaml.nim
* Properly encapsulate exceptions from YamlStreams
* Discovered a Nim compiler bug and added a workaround
* Added possibility to query line, column and line content from
YamlParser
* Updated rst documentation
* Actually call callback in fastparse if available
* Fixed parsing YAML version
* fastparse now used for everything
* private/[lexer|parser].nim gone
* Made fastparse use YamlParser object
* Got all tests working with fastparse
* Some more smaller changes to API
* Scan for implicit map key as soon as a property is encountered
* Proper handling of node properties throughout parser
* Fixed an issue where plain scalars sometimes contained garbage
when yielded
* Catch reserved chars @ and `
* Fixed some test cases
* Made parser test cases use fastparse
* Made TagLibrary a ref object
* Started writing fastparse.nim, a parser aimed to deliver
much faster speeds than the current implementation. It
currently is only able to parse block content
