1  ## 2.1.1
  2  
  3  Bugfixes:
  4  
  5   * Fixed a bug that prevented float values to be serialized (#141)
  6  
  7  ## 2.1.0
  8  
  9  Features:
 10  
 11   * New pragmas ``scalar`` and ``collection`` allow you to modify the
 12     presentation style used for certain types and object fields.
 13     Defined in new module ``yaml/style``.
 14   * The presenter now honors the node style set in the events it presents,
 15     if possible. So if a scalar is set to be a literal block scalar, it is
 16     presented as such unless impossible or presenter options specifically
 17     prevent this.
 18   * The presenter can now output single-quoted scalars. It only does so when
 19     this scalar style is explicitly set on an event.
 20  
 21  Changes:
 22  
 23   * renamed ``canonicalDumper`` / ``setCanonicalStyle`` to
 24     ``explanatoryDumper`` / ``setExplanatoryStyle`` because it was
 25     a misnomer and there is nothing canonical about this output style.
 26     The terminology *canonical* was carried over from PyYAML, but the
 27     YAML specification uses that term for different things.
 28     The old names are kept with a ``deprecated`` pragma.
 29   * The ``explanatoryDumper`` now automatically enables the
 30     tag shorthand ``!n!``, because in this style you want that for readability.
 31  
 32  Bugfixes:
 33  
 34   * Fixed a bug that prevented instances of generic types to be used in ``Option``
 35     fields (e.g. ``Option[seq[string]]``) (#101)
 36   * Fixed a bug that caused invalid indentation when dumping with certain
 37     settings (#140)
 38   * Fixed parsing errors for verbatim tags in flow style (#140)
 39   * Fixed a bug that caused presentation of block scalars in
 40     flow collections (#140)
 41   * Fixed a bug that sometimes caused the last word of a folded block scalar
 42     not to be presented.
 43   * Fixed maximum line length not properly implemented in presenter in a number
 44     of cases.
 45   * Fixed a bug that prevented the presenter from outputting compact
 46     flow mappings in cMixed mode.
 47   * Fixed block scalars as mapping keys not being presented properly.
 48   * Added workaround for a regression bug in Nim 2.0.2
 49     (https://github.com/nim-lang/Nim/issues/23112)
 50  
 51  ## 2.0.0
 52  
 53  Breaking Changes:
 54  
 55   * Requires Nim 2.x
 56   * ``yaml/serialization`` has been split into
 57     ``yaml/native`` (low-level API to load a ``YamlStream`` into Nim vars),
 58     ``yaml/loading`` (high-level loading API) and
 59     ``yaml/dumping`` (high-level dumping API).
 60   * Dumping API now has a ``Dumper`` object.
 61     All dumping procs require an instance of this object.
 62     Previous parameters ``tagStyle``, ``handles`` and ``options`` have been
 63     moved into this object.
 64   * Constants with default values have been removed in favor of default values
 65     for object fields.
 66   * Low-level native API for loading and dumping has undergone changes to
 67     parameter order. Serialization and deserialization context is now first
 68     parameter, to be able to use prefix notation.
 69   * Removed ``PresentationStyle``; instead ``Dumper`` can be initialized with
 70     presets mirroring the former values of ``PresentationStyle``.
 71   * Removed deprecated ``YamlDocument`` type from DOM API.
 72     Use ``YamlNode`` instead.
 73  
 74  Features:
 75  
 76   * Can now load and dump fields of the parent type(s) of used types (#131)
 77   * Updated type guessing to use regexes from YAML 1.2 instead of old YAML 1.1.
 78     Type guessing is used primarily for heterogeneous items
 79     (implicit variant objects).
 80   * Presenter no longer outputs trailing spaces
 81   * More presentation options
 82   * Simplified default presentation style: Don't output ``%YAML 1.2``,
 83     ``%TAG !n! ...`` or ``---`` by default.
 84     Output compact notation for collections.
 85   * Loading now works at compile time (#70, #91).
 86     Dumping doesn't work at compile time currently.
 87  
 88  Bugfixes:
 89  
 90   * Don't crash on invalid input, instead raise a catchable
 91     ``YamlParserError`` (#129)
 92   * Fixed some parser errors in exotic edge cases
 93   * Several fixes to make the library work with Nim 2.x
 94  
 95  ## 1.1.0
 96  
 97  Features:
 98  
 99   * ``YamlNode`` now contains node styles and preserves them
100     when serializing to YAML again
101   * Added ``maxLineLength`` to ``PresentationOptions``. (#119)
102   * Added ``loadFlattened`` to resolve aliases while loading,
103     instead of deserializing them into pointers. (#117)
104  
105  Bugfixes:
106  
107   * Fixed problems with ARC/ORC (#120)
108   * Fixes some edge cases around whitespace while parsing
109   * Fixed a problem that made ``{.ignore: [].}`` fail when trying
110     to ignore collection values. (#127)
111   * Always write a newline character at the end of output, as required
112     by POSIX for text files.
113   * Fixed an error with loading recursive nodes into YamlNodes.
114   * Fixed an error where ``0`` could not be loaded into an unsigned
115     integer lvalue. (#123)
116   * Fixed an error where `float32` values could not properly
117     be deserialized. (#124)
118   * Fixed a compiler error concerning stricteffects. (#125)
119  
120  ## 1.0.0
121  
122  Features:
123  
124   * ``YamlNode`` can now be used with the serialization API (``load`` / ``dump``)
125     and can be used to hold substructures that should not be deserialized to
126     native types (#48).
127  
128  Bugfixes:
129  
130   * Raise a proper exception when a stream contains no documents but one is
131     expected (#108)
132   * Comments after a block scalar do not lead to a crash anymore (#106)
133   * Fixed an error with parsing document end markers (#115)
134   * Fixed an error when serializing block scalars (#105)
135  
136  ## 0.16.0
137  
138  Features:
139  
140   * dumping ``sparse`` objects now omits empty ``Option`` fields (#100).
141  
142  Bugfixes:
143  
144   * Fixed several parser errors that emerged from updates on the test suite.
145   * Fixed ``raises`` annotations which could lead to compilation errors (#99).
146  
147  ## 0.15.0
148  
149  Features:
150  
151   * Compiles with --gc:arc and --gc:orc
152  
153  Bugfixes:
154  
155   * Parser rewrite: Fixes some test suite errors (including #83)
156   * Fixed problems where a syntax error lead to an invalid state (#39, #90)
157   * Serialize boolean values as ``true`` / ``false`` instead of ``y`` / ``y``
158     to conform to YAML 1.2 spec.
159  
160  ## 0.14.0
161  
162  Features:
163  
164   * **Breaking change**:
165     transient, defaultVal, ignore and implicit are now annotations.
166   * Added ``sparse`` annotation to treat all ``Option`` fields as optional.
167  
168  Bugfixes:
169  
170   * can now use default values with ref objects (#66)
171  
172  ## 0.13.1
173  
174  Bugfixes:
175  
176   * Changed `nim tests` to `nim test` to make nim ci happy.
177  
178  ## 0.13.0
179  
180  Bugfixes:
181  
182   * Fixed submodule link to yaml-test-suite.
183  
184  Features:
185  
186   * Added support for `Option` type.
187  
188  ## 0.12.0
189  
190  Bugfixes:
191  
192   * Made it work with Nim 0.20.2
193  
194  ### 0.11.0
195  
196  Bugfixes:
197  
198   * Made it work with Nim 0.19.0
199  
200  NimYAML 0.11.0 is unlikely to work with older Nim versions.
201  
202  ## 0.10.4
203  
204  Bugfixes:
205  
206   * Made it work with Nim 0.18.0
207  
208  ### 0.10.3
209  
210  Bugfixes:
211  
212   * Fixed a nimble error when installing the package.
213  
214  Features:
215  
216   * Added `ignoreUnknownKeys` macro to ignore all mapping keys that do not map
217     to a field of an object / tuple (#43).
218  
219  ### 0.10.2
220  
221  Bugfixes:
222  
223   * Fixed a nimble warning (#42)
224   * Make sure special strings (e.g. "null") are properly quoted when dumping JSON
225     (#44)
226  
227  ### 0.10.1
228  
229  Bugfixes:
230  
231   * Made it *actually* work with Nim 0.17.0.
232  
233  ### 0.10.0
234  
235  Features:
236  
237   * Compatibility with Nim 0.17.0 (#40).
238     **Important:** This fix breaks compatibility with previous
239     Nim versions!
240  
241  ### 0.9.1
242  
243  Features:
244  
245   * Added `YamlParser.display()` which is mainly used by tests
246   * NimYAML now builds for JS target (but does not work properly yet)
247  
248  Bugfixes:
249  
250   * Correctly present empty collections in block-only style (#33)
251   * Correctly handle `{1}` (#34)
252   * Recognize empty plain scalar as possible `!!null` value
253   * Require colons before subsequent keys in a flow mapping (#35)
254   * Allow stream end after block scalar indicators
255   * Fixed regression bugs introduced with timestamp parsing (#37)
256  
257  ### 0.9.0
258  
259  Features:
260  
261   * Better DOM API:
262     - yMapping is now a Table
263     - field names have changed to imitate those of Nim's json API
264     - Better getter and setter procs
265   * Added ability to resolve non-specific tags in presenter.transform
266  
267  Bugfixes:
268  
269   * Fixed parsing floating point literals (#30)
270   * Fixed a bug with variant records (#31)
271   * Empty documents now always contain an empty scalar
272   * Block scalars with indentation indicator now have correct whitespace on first
273     line.
274  
275  ### 0.8.0
276  
277  Features:
278  
279   * NimYAML now has a global tag URI prefix for Nim types,
280     `tag:nimyaml.org,2016:`. This prefix is denoted by the custom tag handle
281     `!n!`.
282   * Support arbitrary tag handles.
283   * Added ability to mark object and tuple fields as transient.
284   * Added ability to set a default value for object fields.
285   * Added ability to ignore key-value pairs in the input when loading object
286     values.
287   * Support `!!timestamp` by parsing it to `Time` from module `times`.
288  
289  Bugfixes:
290  
291   * Fixed a bug concerning duplicate TagIds for different tags in the
292     `serializationTagLibrary`
293   * Convert commas in tag URIs to semicolons when using a tag URI as generic
294     parameter to another one, because commas after the tag handle are interpreted
295     as flow separators.
296  
297  ### 0.7.0
298  
299  Features:
300  
301   * Better handling of internal error messages
302   * Refactoring of low-level API:
303     * No more usage of first-class iterators (not supported for JS target)
304     * Added ability to directly use strings as input without stdlib's streams
305       (which are not available for JS)
306   * Added ability to parse octal and hexadecimal numbers
307   * Restructuring of API: now available as submodules of yaml. For backwards
308     compatibility, it is still possible to `import yaml`, which will import all
309     submodules
310   * Check for missing, duplicate and unknown fields when deserializing tuples and
311     objects
312  
313  Bugfixes:
314  
315   * Fixed double quotes inside plain scalar (#25)
316   * Return correct line content for errors if possible (#23)
317   * Some smaller lexer/parser fixes
318  
319  ### 0.6.3
320  
321  Bugfixes:
322  
323   * Can load floats from integer literals (without decimal point) (#22)
324  
325  ### 0.6.2
326  
327  Bugfixes:
328  
329   * Fixed problem when serializing a type that overloads the `==` operator (#19)
330   * Fixed type hints for floats (`0` digit was not processed properly) (#21)
331  
332  ### 0.6.1
333  
334  Bugfixes:
335  
336   * Fixed deserialization of floats (#17)
337   * Handle IndexError from queues properly
338  
339  ### 0.6.0
340  
341  Features:
342  
343   * Properly support variant object types
344   * First version that works with a released Nim version (0.14.0)
345  
346  Bugfixes:
347  
348   * Fixed a crash in presenter when outputting JSON or canonical YAML
349   * Raise an exception when trying to output multiple documents in JSON style
350  
351  ### 0.5.1
352  
353  Bugfixes:
354  
355   * Fixed a problem that was introduced by a change in Nim devel
356  
357  ### 0.5.0
358  
359  Features:
360  
361   * Support variant object types (experimental)
362   * Added ability to use variant object types to process
363     heterogeneous data
364   * Support `set` type
365   * Support `array` type
366   * Support `int`, `uint` and `float` types
367     (previously, the precision must be specified)
368   * Check for duplicate tag URIs at compile time
369   * Renamed `setTagUriForType` to `setTagUri`
370  
371  Bugfixes:
372  
373   * None, but fastparse.nim has seen heavy refactoring
374  
375  ### 0.4.0
376  
377  Features:
378  
379   * Added option to output YAML 1.1
380   * Added benchmark for processing YAML input
381   * Serialization for OrderedMap
382   * Use !nim:field for object field names (#12)
383  
384  Bugfixes:
385  
386   * Code refactoring (#9, #10, #11, #13)
387   * Some small improvements parsing and presenting
388  
389  ### 0.3.0
390  
391  Features:
392  
393   * Renamed some symbols to improve consistency (#6):
394     - `yamlStartSequence` -> `yamlStartSeq`
395     - `yamlEndSequence` -> `yamlEndSeq`
396     - `yamlStartDocument` -> `yamlStartDoc`
397     - `yamlEndDocument` -> `yamlEndDoc`
398     - `yTagMap` -> `yTagMapping`
399   * Introduced `PresentationOptions`:
400     - Let user specify newline style
401     - Merged old presentation options `indentationStep` and `presentationStyle`
402       into it
403   * Use YAML test suite from `yaml-dev-kit` to test parser.
404  
405  Bugfixes:
406  
407   * Fixed various parser bugs discovered with YAML test suite:
408     - Block scalar as root node no longer leads to a parser error
409     - Fixed a bug that caused incorrect handling of comments after plain scalars
410     - Fixed bugs with newline handling of block scalars
411     - Fixed a bug related to block sequence indentation
412     - Skip content in tag and anchor names and single-quoted scalars when
413       scanning for possible implicit map key
414     - Properly handle more indented lines in folded block scalars
415     - Fixed a problem with handling ':' after whitespace
416     - Fixed indentation handling after block scalar
417  
418  ### 0.2.0
419  
420  Features:
421  
422   * Added DOM API
423   * Output block scalars in presenter if scalar is long and block scalar output
424     is feasible. Else, use multiple lines for long scalars in double quotes.
425  
426  Bugfixes:
427  
428   * Improved parser (#1, #3, #5)
429   * Made parser correctly handle block sequences that have the same indentation
430     as their parent node (#2)
431   * Fixed problems with outputting double quoted strings (#4)
432  
433  ### 0.1.0
434  
435   * Initial release