v3.7

• Added ability to disable support for inlines using dialect argument to load() and loads().
• Added get_keys(), get_value(), get_line_numbers(), and get_location().
• Deprecated get_value_from_keys(), get_lines_from_keys(), get_original_keys(), and join_keys().
• Added offset argument to Location.as_line().
• Add ability to specify source to load().
• Clarified policy on white space in inline strings.

v3.6

• De-duplicating with the on_dup argument to loads() now works well for error reporting with keymaps.
• The map_keys argument has been added to dump() and dumps().

v3.5

• Minor refinements and fixes.

v3.4

• improved the on_dup parameter to load and loads functions.
• added strict argument to join_keys functions.

v3.3

Changes to Language Definition

• Defined Minimal NestedText, a subset of NestedText.
• NestedText document files should end with a newline.

Changes to Python Implementation

• add normalize_key argument to load and loads.
• added utility functions for operating on keys and keymaps:
  • get_value_from_keys
  • get_lines_from_keys
  • get_original_keys
  • join_keys
• None passed as key is now converted to an empty string rather than "None".

v3.2

• add circular reference detection and reporting.

v3.1

• change error reporting for dump functions; culprit is now the keys rather than the value.

v3.0

• Deprecate trailing commas in inline lists and dictionaries.
• Adds keymap argument to load() and loads().
• Adds inline_level argument to dump() and dumps().
• Implement on_dup argument to load() and loads() in inline dictionaries.
• Apply convert and default arguments of dump() and dumps() to dictionary keys.

v2.0

• Deprecate quoted keys.
• Add multiline keys to replace quoted keys.
• Add inline lists and dictionaries.
• Move from renderers to converters in dump and dumps. Both allow you to support arbitrary data types. With renderers you provide functions that are responsible for directly creating the text to be inserted in the NestedText output. This can be complicated and error prone. With converters you instead convert the object to a known NestedText data type (dict, list, str, ...) and the dump function automatically formats it appropriately.
• Restructure documentation.

Be aware that aspects of this version are not backward compatible.

1. It no longer supports quoted dictionary keys.
2. The renderers argument to :func:dump and :func:dumps has been replaced by converters.
3. It no longer allows one to specify level in :func:dump and :func:dumps.

v1.3

• Move the test cases to a submodule.