JParseCode 0.23.2

Added

• classSignature.comments parsing and AST output added

JParseCode 0.23.1

Changed

• Expanded C# and Java field parsing tests

JParseCode 0.23.0

Added

• Lots of additional unit tests
• TypeExtractor.isSimpleLiteral() used to determine whether a FieldDef.initializer should be serialized as a literal value "initializer" or an "initializerExpression"

Changed

• FieldExtractor now parses field default value assignments and returns FieldDef instead of FieldSig (FieldDef contains initializer field)
• FieldExtractor constructor requires a new 3rd parameter, an OperatorUtil<>
• ClassAst remove T_ENUM type parameter, enum members in a parsed enum will be represented by T_FIELD
• FieldDef.initializerToJson() signature changed significantly to support new multiple token initializer expressions
• AccessModifierParser and AccessModifierEnum removed parseFromSrc(), cleaned up code, added documentation
• AccessModifierParser tryParseFromSrc(String) and defaultAccessModifier(String, ...) modified to takes List<String> for correctness and to simplify call sites
• AstExtractor.createFieldParser() and implementations (CsBlock and JavaBlock) now return AstParser<List<FieldDef>> instead of ...FieldSig
• BlockUtil parseKeyword() and tryParseKeyword() replaced with tryToBlock()
• CsBlockParser._extractBlocksFromTree() renamed extractBlocksFromTree() (which was already public)
• AstParser added blockComplete() with empty default implementation, will be called by BlockExtractor

Removed

• Merged FieldSig and FieldSigResolved into FieldDef and FieldDefResolved

Fixed

• C# annotation parsing fixed to handle multiple annotations in the same block, i.e. '[Annotation(One), Annotation(Two), ...]'
• An issue with nested classes/interfaces getting assigned the incorrect access modifiers
• An issue with complex generic type parsing

JParseCode 0.22.0

Removed

• ProjectClassSet resolveSimpleName() in favor of moving the two lines of code to the single calling location

Fixed

• C# full type name resolution fixed to resolve against the parent namespaces the class resides in (affects types in class signatures, method signatures, and fields)
• Java class signature parsing fixed to support both extends and implements in the same signature implements (can't believe I overlooked this and didn't have a unit test)

JParseCode 0.21.0

Parsing performance optimizations (~15-20%)

Changed

• Update dependencies [email protected] and [email protected]
• ParserActionLogger and ParseTimes optimizations and changes to work with jtext-tokenizer new TokenizationLogger
• JavaFileTokenizer, CsFileTokenizer, CommentTokenizer, IdentifierTokenizer, and NumberTokenizer reusable/reusableCharParsers flag implementation to control which CharParserMatchableFactory implementation is used
• CodeTokenizer.tokenizeDocument() fragmentConstructor parameter optimization to skip creating fragment source string twice, major performance improvement ~5-10%
• New aggregate performance logging output for text tokenization related counters

JParseCode 0.20.0

Method generic type parameters parsing support (i.e. 'public T Create()' in C#).

Added

• Method generic type parameters added - MethodSig and sub-classes have a new typeParameters field. Note: if these 'types' contain lower/upper type bounds these are included in the typeName property in Java and are not yet included in C# parsing
  • Added basic unit tests for this new method generic type parameters parsing support
• AstFragType.isBlock(CodeToken, char) overload for attempted performance optimization

Changed

• Renamed DataTypeExtractor -> TypeExtractor
• Added EnhancedListIterator as a replacement for TokenListIterable to try and simplify fragment iteration since it is a fairly hot section of code in most parsing scenarios
• ParseTimes.TrackerAction enums renamed:
  • LOAD -> READ
  • PARSE -> EXTRACT_AST
• Update dependency [email protected] and code to match
  • Adjustments to new CharParserMatchableFactory() and CharConditions.Identifier.newInstance() calls in IdentifierTokenizer and NumberTokenizer
• JSON output now excludes empty annotations arrays on fields, methods, and parameters
• Additional performance log counters added to several methods and printed when performance info is enabled via -debug CLI flag

Removed

• AstFragType.isType() static method and overloads since they were unused
• Unused CommentTokenizer.createCommentTokenizerForJava()

Fixed

• An exception being thrown when resolving class names if a file imported a namespace and also imported a child namespace of the first namespace

JParseCode 0.19.2

Parameter varargs parsing support (i.e. 'int...' in Java).

Changed

• Update dependency [email protected] and [email protected]
  • Classes combined, class names simplified, and unused classes and methods removed from libraries
  • Code identifier parser now provided by jtext-tokenizer
  • Several bug fixes around compound optional parser conditions
• Added char[] src, int srcOff, int srcLen parameters to CodeTokenizer.tokenizeDocument()
• Renamed IdentifierTokenizer newIdentifierTokenizer() to createIdentifierTokenizer()
• Improved unit tests

Removed

• IdentifierTokenizer.createIdentifierTokenizer()

JParseCode 0.19.1

Changed

• Finish CommentAndWhitespaceExtractor and tests for it
• TextToken interface now includes hashCode() and equals(Object)
  • TextFragmentRefToken now implements hashCode() and equals(Object)
• Minor code cleanup, use StringSplit.split() with char instead of String where possible

JParseCode 0.19.0

0.19.0 - 2019-07-04

Changed

• IdentifierTokenizer.createIdentifierWithGenericTypeTokenizer() now takes one parameter int maxGenericTypeDepth
• Changed CsFileTokenizer.createFileParser() -> createCsTokenizers() and JavaFileTokenizer.createFileParser() -> createJavaTokenizers()
• Added some private constructors that throw AssertionError to static classes

Removed

• Changed GenericTypeTokenizer._createGenericTypeTokenizer() from public to private
• Removed IdentifierTokenizer field static int genericTypeDepth in favor of callers explicitly passing the argument to createIdentifierWithGenericTypeTokenizer() which now takes one parameter int maxGenericTypeDepth
• Removed CodeTokenizerBuilder in favor of CodeTokenizer static methods
  • Manually build a tokenizer list of type PairList<CharParserFactory, TextTransformer<CodeTokenType>>
  • Call CodeTokenizer.createTokenizer() with the language you used to pass to the CodeTokenizerBuilder constructor and the list of tokenizers you manually created

Fixed

• AnnotationExtractor to handle all C# keyword-followed-by-a-block annotation arguments like default(T), nameof(T), and typeof(T)

JParseCode 0.18.1

Fixed

• AnnotationExtractor to handle C# typeof(T) annotation arguments