CHANGELOG

Version 1.3.1 (2019-06-24)
--------------------------
- Upgraded to ASM 7.1
- Fixed class loader in parser generation
- Added cross-build for Scala 2.13.0
- Dropped support for Scala 2.10


Version 1.3.0 (2019-01-23)
--------------------------
- Upgraded to ASM 7.0
- Added Scala 2.13.0-M5 build


Version 1.2.0 (2018-09-14)
--------------------------
- Added support for `~~>` and friends in ReductionRules
- Added support for PushRules to chain after ReductionRules
- Upgraded to ASM 6.2.1
- Added Scala 2.13.0-M4 build
- Dropped support for Java 1.5


Version 1.1.8 (2017-01-11)
--------------------------
- Cross build for Scala 2.10, 2.11 and 2.12
- Upgrade to ASM 5.2
- Fixed "Utils.findConstructor doesn't match boolean parameter" (#94)
- Added method for clearing class cache in AsmUtil


Version 1.1.7 (2015-02-09)
--------------------------
- Upgrade to Scala 2.11.4, cross build for Scala 2.9.2, 2.9.3, 2.10 and 2.11
- Upgrade to ASM 5.0.3, closed #76
- Make parboiled-java relocatable, closed #80


Version 1.1.6 (2013-08-22)
--------------------------
- Core:
  - Improved error msg for unexpected end of input
  - Fixed error reporting for rules underneath testNot rules
- Scala
  - Added support for ReductionRule1 inside oneOrMore/zeroOrMore while returning a subtype
  - Upgraded to Scala 2.10.2


Version 1.1.5 (2013-05-02)
--------------------------
- General: Added missing OSGi Bundle metadata
- Core: Fixed MemoMismatcher producing erroneous parsing results
        for subsequent parsing runs with the same parser instance
- Java: Fixed problem AsmUtils
- Scala: Upgraded to Scala 2.9.3 and 2.10.1


Version 1.1.4 (2012-11-27)
--------------------------
Core:
- Added parser automatic parser termination after matching 100K EOI, closed #52
- Extended IndentDedentBuffer with option to not swallow empty lines, closed #56
Java:
- Upgraded to ASM 4.1
Scala:
- Upgraded to Scala 2.10.0-RC3


Version 1.1.3 (2012-10-16)
--------------------------
Core & Java:
- Removed superfluous dependency on Scala library


Version 1.1.2 (2012-10-15)
--------------------------
Scala:
- Fix cross-building path generation to include the full scala-version
for unstable versions (milestones, RC)
- Upgrade to Scala 2.10-RC1


Version 1.1.1 (2012-09-21)
--------------------------
General:
- Fixed incorrect scope for scalatest dependency
- Upgraded to SBT 0.12.0
- Changed deployment target to sonatype (i.e. maven central)
Scala:
- Enabled proper scala-version cross-building
- Upgraded to Scala 2.10-M7


Version 1.1 (2012-07-28)
------------------------
Java:
- Added `NTimes` rule construction helper
- Upgraded to ASM 4.0, closes #41
- Improved ABC example
Scala:
- Upgraded to Scala 2.10-M6


Version 1.0.2 (2011-09-30)
--------------------------
Core:
- Fixed #32 (ReportingParseRunner assumes only one parse error)
Java:
- Added 'NoneOf' rule constructor
- Updated Java parser example
- Fixed #22 (problem with parse tree creation for skipNode() matchers)
- Fixed #24 (IllegalStateException in certain cases involving parser inheritance)
- Fixed #30 (problem in action class naming resulting in NoSuchFieldError in certain cases)
- Fixed #31 (incorrect handling of 'switch' statements in rule method extension)
Scala:
- Upgraded to Scala 2.9.1
- Fixed #29 (nTimes(1, "a") will throw ClassCastException)


Version 1.0.1 (2011-07-05)
--------------------------
Scala:
- Upgraded to Scala 2.9.0-1


Version 1.0.0 (2011-05-13)
--------------------------
Scala:
- Upgraded to Scala 2.9.0.final 


Version 0.11.2 (2011-05-13)
---------------------------
- Changed build system from Apache Buildr to SBT 0.7.7 (Scala Build Tool)
Core:
- Added convenience methods to Context and BaseActions returning the current location in the InputBuffer as a Position
- Fixed another potential problem with ActionMatcher subcontext initialization
- Fixed MissingFormatArgumentException in ParseRuntimeException when args array is empty
Java:
- Added convenience 'currentIndex()' method to BaseActions
Scala:
- Added '~:?' operator
- Fixed 'withContext' not usable on '~%' operator problem
- Turned all remaining references to Context[_] to Context[Any]
(this is the last version built against Scala 2.8.1)


Version 0.11.1 (2011-04-06)
---------------------------
Core:
- Fixed bug causing parser actions at the first position in sequences to sometimes inadvertently
  advance the match cursor, leading to hard-to-find input mismatches
- Fix bug in TracingParseRunner.withLog
- Added support for timeouts to RecoveringParseRunner


Version 0.11.0 (2011-03-21)
---------------------------
Core:
- Major overhaul of RecoveryParseRunner, fixed numerous bugs, improved performance
- Changed RecoveringParseRunner to not create additional ParseTreeNodes underneath error nodes
- Added Context.inErrorRecovery method
- Added "strict" flag to IndentDedentBuffer for enabling/disabling exception throwing on illegal indentations
- Fixed bug in IndentDedentBuffer leaving blank lines on indented comments before EOI
- Smaller fixes, cleanups and improvements 
Java:
- Switched internal class name resolving from using the current Threads ContextClassLoader to using the respective
  class' own ClassLoader in an attempt to increase compatibility with a broader range of application environments
Scala:
- Simplified ParsingResult to be more scala idiomatic, removed "result", renamed "resultOption" to "result",
  removed "hasErrors" method
- Renamed POPx rules to DROP, DROP2 and DROP3
- Removed Parser.withParseTreeBuilding method, changed Parser.buildParseTree from var to a def, which needs to be
  overridden in order to activate parse tree building
- Removed empty parens on a number of rule marker methods
- Added overloads with Rule2 parameter to optional, zeroOrMore, oneOrMore and nTimes


Version 0.10.1 (2011-02-11)
---------------------------
Core:
- Added Context.getMatchLength() method and helpers
- Added support for IndexRanges
- Completely overhauled ParseRunner implementations architecture in preparation for coming extensions
- Deprecated static run methods of main ParseRunners
- Improved logging architecture for DebuggingValueStack and TracingParseRunner
- Moved class InputBuffer.Position into package org.parboiled.support
- Fixed small memory leak in DefaultValueStack
- Completely rewrote IndentDedentBuffer implementation (the previous one had serious limitations)
- Many smaller fixes, cleanups and improvements 
Java:
- Added StringBuilderVar helper for parser actions 
- Fixed bug causing StringIndexOutOfBoundsException during parser extension for parsers without package prefix
- Fixed NPE in AsmUtils caused by actions involving methods defined in super-interfaces
- Upgraded to ASM 3.3.1
Scala:
- Fixed bug in Rule0 causing ClassCastException in certain cases involving ReductionRule2s and PopRule2s
- Moved Input.collectContents from scala module into core


Version 0.10.0 (2011-01-13)
---------------------------
- Switched build system from Ant to Apache Buildr
- Completed mavenization, parboiled artifacts are now also available from the http://scala-tools.org/repo-releases/ maven repository
- Split source into several modules creating distinct artifacts
- Changed to 3 part version numbering in preparation of the coming semantic versioning compatibility (http://semver.org)
- Fixed VerifyError with certain Var<...> constructs
- Fixed IllegalStateException during error recovery involving AnyMatchers
- Fixed bug in creation of expected string error message for AnyOf matchers
- Fixed bug in DebuggingValueStack (thx to Matthieu Baechler)
- Fixed GH issue #10 (ParseTreeUtils.findNodeByLabel() methods are non static)
- Fixed bug in ErrorUtils
- Fixed Java parser example to support non-ascii chars in identifiers
- Added pushFromContext rule creator to Scala facade
- Improved Scala Input wrapper
- Removed internal dependency on jetbrains annotations
- Removed internal dependency on google collections
- Upgraded to ASM 3.3
- Upgraded to Scala 2.8.1
- Many smaller cleanup and improvements


Version 0.9.9.0 (2010-10-04)
----------------------------
- BREAKING CHANGES:
  - Moved all named MatcherVisitor implementations from org.parboiled.support to new package org.parboiled.matchervisitors
  - Moved InputBuffer implementations from org.parboiled.support to to new package org.parboiled.buffers
  - Moved special character definitions from org.parboiled.Characters class into dedicated class org.parboiled.Characters
  - Decapitalized not-extended rule construction helper methods in BaseParser (e.g. ToRule -> toRule) for better compliance with the style guide
  - New MatcherPath implementation, not more "lastMatch" property in InvalidInputError
- Significantly extended Scala facade:
  - Cleaner structure, faster rule building, faster rules (e.g. enabled fast-string-matching for FirstOf operator)
  - More flexible action functions (more action operators, 'withContext' actions)
  - Added missing variance annotations to rule types and aliases
  - Added org.parboiled.scala.testing.ParboiledTest trait for simplified parser testing
  - More and better scala wrappers (ParseRunners, ParsingResult, parsing Input)
- Added TracingParseRunner as another important debugging tool
- Added IndentDedentInputBuffer supporting line indentation based grammars
- Added Context.getFirstMatchChar and helpers in Java and Scala
- Added NothingMatcher as counterpart to the EmptyMatcher, along with respective rule constants
- Added missing 'getStarterChar' functionality to CustomMatcher base class
- Improved filtering of Node and Matcher tree printing with simple predicate logic
- Fixed incorrect parse error reporting in some cases
- Fixed in Scala facade: Some standard rule builders don't group their sub rules correctly
- Fixed bug in RecoveringParseRunner causing NPE if very first char was not matched
- Fixed bug in RecoveryParseRunner incorrectly executing pre-error sub rules of resync sequences
- Fixed bug preventing the creation of 'action-only' rules
- Fixed bug in FirstOfStringsMatcher preventing the matching of alternatives that are prefixes of other alternatives
- Fixed matcher labelling problem in Scala facade
- Many smaller cleanups and improvements


Version 0.9.8.2 (2010-08-13)
----------------------------
- BREAKING CHANGES:
  - Renamed BaseParser.FirstOf(string) overload to BaseParser.AnyOf
  - Renamed CharSetMatcher to AnyOfMatcher
- Fixed bug in MemoMismatchesMatcher causing erroneous rule mismatches at input location 0


Version 0.9.8.1 (2010-08-12)
----------------------------
- BREAKING CHANGE:
  - Renamed BaseParser.CharSet methods to BaseParser.FirstOf overloads
- Fixed bug in RecoveringParseRunner producing inconsistenst value stack states during rule resynchronization
- Added BaseParser.matchOrDefault() helper
- Added ability to set ClassLoader for auxiliary classes (for parser extension) via the threads ContextClassLoader
- Added a few more convenience overloads to Parser.zeroOrMore(...) and Parser.oneOrMore(...) in Scala facade
- Added MemoMismatches rule option to Scala facade
- Improved Javadoc and Scaladoc documentation as well as Scala examples


Version 0.9.8.0 (2010-08-10)
----------------------------
- BREAKING CHANGES:
  - Added a parser "Value Stack", retired parser value handling via parse tree decoration
  - Removed all action expression helpers for the old parse-tree-based value handling
  - Changed default parsing behavior from parse-tree-building to parse-tree-less, added @BuildParseTree annotation
  - Changed Eoi(), Any() and Empty() to public static final fields on the BaseParser
  - Renamed BaseParser.CharIgnoreCase and BaseParser.StringIgnoreCase to BaseParser.IgnoreCase
- Fixed bug provoking an exception at the 65536th input character
- Fixed bug in Test-/TestNotMatchers
- Fixed some problems in Java Parser example
- Added Scala facade for efficient parser building without byte code manipulation,
  featuring DSL elements for value stack operations with complete static type checking
- Added a complete JsonParser as an example for using the Scala facade
- Added a matcher specialization for fast string matching also in FirstOf-Rules (~10% speedup in JavaParser example)
- Added overloads to standard rule builders OneOrMore(...), Optional(...), Test(...), TestNot(...) and ZeroOrMore(...)
  directly taking several sub rules as a sequence
- Added @MemoMismatches and respective implementation (which is a first step towards "packratting" support)
- Added ProfilingParseRunner and JavaParserProfiler example
- Adapted build process to build two independent libraries, one for Java, one for Scala
- Changed source directory structure to "Maven-style"


Version 0.9.7.3 (2010-07-09)
----------------------------
- Fixed several problems in parse error recovery
- Added fast string matching (significant speed increase for languages relying on string matching)
- Added additional error recovery strategy: Single character replacement
- Added support for custom InputBuffer implementations
- Removed superfluous CAPTURE constructs along with all related code


Version 0.9.7.2 (2010-06-14)
----------------------------
- BREAKING CHANGE: Tracking of locations in input text is now purely done via integer indices (was: instances of
  class org.parboiled.common.InputLocation) for a much reduced memory footprint and a significant performance increase,
  especially for large input files and rather simple grammars
  (most of the functionality of org.parboiled.common.InputLocation has been moved to org.parboiled.common.InputBuffer)
- Added smaller extensions in a few support classes


Version 0.9.7.1 (2010-06-08)
----------------------------
- BREAKING CHANGE: Rule methods w/ parameters are now also automatically labelled, @Label now requires parameter!
- Fixed action variable initializers not working as expected in certain scenarios
- Fixed several bugs related to overriding of Rule methods in multi-level parser class hierarchies
- Fixed "Constructor with descriptor '()V' not found" exception in certain action expressions
- Fixed displaying of empty "expected" part in parse error messages caused by TestNot(...) failures
- Fixed small problems with building against Java 6
- Added missing field and method private access verification for non-action instructions
- Added Context.getPrevStartLocation() and Context.getPrevEndLocation() and respective helpers
- Added convenience overload ErrorUtils.printParseErrors(ParsingResult)
- Improved error reporting of parse errors caused by TestNot(...) failures
- Improved parsing performance of Reporting- and RecoveringParseRunner by 10+ % at the expense of one more parsing run on invalid input
- Improved calculator examples, added new CalculatorParser6 example demonstrating parse-tree-less parsing and action variables
- Changed java example to exercise parboileds parse-tree-less instead of the parse-tree-based parsing performance


Version 0.9.7.0 (2010-04-28)
----------------------------          
- Fixed exception during parser creation when parser contains action/capture expressions with constructor calls
- Fixed NoSuchFieldError occuring in some scenarios with similar actions in derived parser classes
- Fixed "Couldn't get field" exception during parser extension occuring in parsers accessing parent fields
- Fixed bug in parser extension causing java.lang.VerifyError in some scenarios
- Fixed Context.inPredicate() not working underneath TestNotMatchers
- Fixed illegal context issue in action expressions recursing into themselves
- Fixed problem in MutableTreeNodeImpl preventing the addition of new child nodes
- Fixed unintended @SuppressNode on String/StringIgnoreCase rules (now:@SuppressSubnodes)
- Added support for action variables (org.parboiled.support.Var<T>)
- Added prevValue(), prevText() and prevChar() functionality for parse tree less AST node building
- Added support for parser statistics (class org.parboiled.ParserStatistics)
- Added @SkipActionsInPredicates and @DontSkipActionsInPredicates
- Added BaseParser.newInstance()


Version 0.9.6.0 (2010-04-14)
----------------------------
- Fixed serious performance problem in construction of parse tree nodes with many subnodes
- Fixed ClassCastException on caching of rule creating methods with primitive array parameters
- Fixed smaller problems in Java parser example
- MAJOR BREAKING CHANGES:
  - Renamed all methods returning Rule instances to begin with an uppercase character
  - Renamed all helper method defined in BaseActions class from all uppercase to "regular" Java method names
  - Renamed @Leaf to @SuppressSubnodes
- Significantly improved parser extension logic for speed and memory requirements
- Added support for action expressions in rule methods with parameters
- Added support for accessing local variables and methods parameters from within action expressions
- Added new overloads taking char[] for CharSet, String and StringIgnoreCase rule creators
- Added BaseParser.FromCharArray(...)
- Added support for explicit action expressions via BaseParser.ACTION(...)
- Added UP2, UP3, UP4, UP5, UP6 and DOWN2, DOWN3, DOWN4, DOWN5, DOWN6 context switching helpers
- Added @DontLabel annotation preventing the automatic labelling of rules created by parameterless rule methods
- Added @SuppressNode and @SkipNode annotations and logic
- Added CustomMatcher base class for custom matcher implementations
- Added Context.lookAhead(int)
- Added support for Captures (special closure-like constructs simplifying certain rule creation scenarios)
- Improved Calculator examples
- Smaller bug fixes and other improvements


Version 0.9.5.1 (2010-03-16)
----------------------------
- Fixed classLoader issue triggering an IOException(Class not found) exception in some environments
- Added org.parboiled.examples.pegtranslator.PegTranslator example, thanks to Radu Vlasov for the contribution
- Replaced BaseActions.NO_ERROR() with the inversed BaseActions.HAS_ERROR()
- Small Javadoc improvements and other cosmetic changes


Version 0.9.5.0 (2010-03-02)
----------------------------
- Major update of basic matching architecture (some breaking changes) 
- Removed BaseParser.parse(...), introduced ParseRunner interface with 4 standard implementations
- Implemented a completely new and much improved parse error reporting and recovery logic
- Removed all, now obsolete, rule enforcement constructs like enforcedSequence(...)
- Replaced a number of utility classes (like ImmutableList) with implementations from Google Collections
- Added Proguard-based build step to perform unused class/member removal for more flexible use of external libraries
- Smaller performance optimization (ca. 10% faster parsing of error free input over version 0.9.2.0)
- Corrected and improved API documentation
- Many smaller bug fixes, refactorings and other improvements


Version 0.9.2.0 (2010-02-02)
----------------------------
- Added @Leaf annotation for marking rules as "creating parse tree leaf nodes"
- Added @Label annotation
- Significantly improved parsing speed (with some @Leaf usage: speedup factor of ca. 3)
- Changed Action.run() to Action.run(Context), removed AbstractAction
- Changed ParseError.context to ParseError.matcherPath
- Removed Parboiled.RecoverFromErrors flag logic
- Removed Failed-Rule-Memoization option due to limited effectiveness
- Improved Java parser example, added performance test
- Smaller bug fixes and architectural improvements


Version 0.9.1.1 (2010-01-29)
----------------------------
- Fixed @Cached problem with overloaded rule creation methods
- Added BaseParser.charSet(...) and a few other standard rule definition overloads


Version 0.9.1.0 (2010-01-28)
----------------------------
- Added Failed-Rule-Memoization option
- Added @Cached annotation logic for automatic caching of rule creating methods with parameters
- Added parser extension caching for not recreating extension class upon repeated Parboiled.createParser(...)
- Added ENFORCED() action expression helper
- Improved character matchers for increased parsing speed
- Renamed @DontExtend to @KeepAsIs
- Removed BaseParser.fromUserObject(...), overwrite toRule(...) for the same effect
- Smaller bug fixes and architectural improvements


Version 0.9.0.2 (2010-01-26)
----------------------------
Major rewrite of action and parser extension infrastructure, now based on bytecode analysis,
action instructions separation and automatic action class generation/parser class rewriting.
Many, many thanks to Ken Wenzel for the original idea as well as a sample implementation!
- Removed ActionResult interface, actions can now be arbitrary boolean expressions
- Removed now obsolete EQUALS, AND, OR, NOT as well as all CONVERT_XXX helpers
- Removed dependency on cglib, parboiled JAR now has no non-SDK dependencies
- Simplified running parboiled examples (now run via ANT) 


Version 0.8.5.0 (2010-01-18)
----------------------------
- Fixed IllegalStateException on trying to label a rule from within a recursion
- Fixed rare bug occurring in certain scenarios involving recursive test/testNot rules
- Improved reporting of parse errors, also renamed ParserConstructionException to GrammarException and
  ParsingException to ActionException, introduced ParserRuntimeException
- Enabled writing access to ParseError list returned by Context.getParseErrors()
- Inlined MatcherContext.runMatcher(boolean) for reduced stack footprint and cleaner stack traces


Version 0.8.4.0 (2010-01-07)
----------------------------
- Fixed incorrect ordering of action calls when several calls are present in one rule
- Fixed test/testNot rules throwing ParserConstructionException at EOI
- Added action helpers: EQUALS, AND, OR, NOT, NEXT_CHAR, NODE_BY_LABEL and NODES_BY_LABEL
- Added Time and RPN example (thanks to Reinier Zwitserloot from http://projectlombok.org for the RPN code) 
- Added Action interface enabling stand-alone action objects within rule creation expressions
- Added SplitParserTest, testing splitting of one grammar into several parser objects
- Simplified ActionParameter infrastructure, now allows casting of action parameters
- Changed access of BaseParser.toRule(s) from private to protected to allow for easy conversion of custom objects


Version 0.8.3.0 (2009-12-15)
----------------------------
- Improved parsing speed by factor 4 after first basic CPU profiling
- Added @SkipInPredicates annotation for action methods
- Added option for convenient rule creation from user objects (suggested by Manuel Brotz)


Version 0.8.2.0 (2009-12-11)
----------------------------
- Added option for manual rule label change
- Added error filter to only report the first parse error at a specific input location
- Added string matching optimization for one-letter strings
- Added some more parser structure validation checks
- Improved customizability of parse tree printouts
- Improved naming of firstOf-rules in parse errors and parse tree printing
- Updated SparqlParser example


Version 0.8.1.0 (2009-12-07)
----------------------------
- Fixed definition of special characters to be fully compatible with the unicode standard
- Fixed IllegalStateException during parse error recovery of testNot() rules
- Added support for case independent matching of characters and strings
- Added SPARQL parser example (incl. test), thanks to Ken Wenzel from http://www.iwu.fraunhofer.de


Version 0.8.0.0 (2009-11-12)
----------------------------
first public release