Changelog for language-ecmascript-0.17.2.0

Version change log. =0.17.2.0= Issue #82: Tighter lexing of identifiers using Unicode character classes (thanks @berdario). Dependency version bumps. =0.17.1.0= Improvements to pretty-printing (Github PR #78) + fixed a bug in quasi-quotation (Github PR #77). =0.17.0.2= Fixed a bug in the Arbitrary instance: no longer generating empty list expressions. Dependency bounds update. =0.17.0.1= Dependency bounds update. =0.17= Added helper functions for code generation (see Language.ECMAScript3.Syntax.CodeGen). Bug fixes to the Arbitrary instances. =0.16.2= Build fixes for GHC 7.4.2 and 7.8.2. =0.16.1= Removed unicode-escaping of illegal identifier characters in the pretty-printer. Adjusted the isValid predicate to check for illegal characters identifiers. Adjusted the Arbitrary instance to only produce valid identifier names. =0.16= Rewritten the Arbitrary instances using 'testing-feat'. Adjusted the pretty-printer to escape invalid characters in identifier names automatically, which gives better usability in code generation use-cases. Added an 'isValid' predicate to check the validity of the AST. Note that the pretty-printer is guaranteed to produce syntactically correct code only for valid ASTs. =0.15.4= Fixed haddock build failure. =0.15.3= Ported the pretty printer to use 'wl-pprint' instead of 'pretty'. Pretty-printed code is now more compact and resembles common formatting conventions. Fixed bugs where the pretty-printer produced unparseable source representations for expression statements starting with '{' or "function". Disabled the QuickCheck pretty-printer test. =0.15.2= Fixed an error in the .cabal file that prevented some of the test modules to be packaged. =0.15.1= Added semicolons after the "return" statement in the pretty printer to avoid certain class of syntax error when printed code is parsed again. Changed the pretty-printer to put the IntLit in parenthesis if it's found on the left-side of the '.' (field access) operator. =0.15= Bug fixes in the pretty-printer, the parser and the QuickCheck arbitrary instance for the AST. Refactored tests to use test-framework. Reorganized the interface for the parser to make it consistent; see deprecation warnings for details. Added Language.ECMAScript3.SourceDiff -- a simple source-based differ for ECMAScript programs, which is now used for visualizing test failures. Revised package dependencies. Factored the analyses out in a separate package: language-ecmascript-analysis; Language.ECMAScript3.Analysis.* are now deprecated. =0.14= Refactoring of the pretty-printing module: added class Pretty with a more comprehensive coverage of AST datatypes and better documentation. Removed duplicate code and deprecated all old interfaces (everything except class Pretty). Old interfaces would be removed in 1.0 (still a long way out). =0.13= Added class PP for pretty-printing =0.12= Added a 'setAnnotation' function to the 'HasAnnotation' class. Added a 'withAnnotation' helper function to modify the top level annotation. =0.11.1= Fixed binary operator precedence in the parser. =0.11= Added a simple quasi-quoter for ECMAScript (see Language.ECMAScript3.Syntax.QuasiQuote) =0.10= The ParenExpr constructor was removed from the Expression AST (WARNING: might break some builds). The pretty printer was rewritten to automatically parenthesize sub-expressions based on operator precedence. Added a (beta quality) arbitrary instance for all the AST's to generate random JavaScript programs (useful for testing). =0.9.1= Increased the upper bound on data-default in dependencies. See https://github.com/jswebtools/language-ecmascript/pull/5