ChangeLog for kdevelop-pg-qt v0.9.5
===================================

* Milian Wolff: use -std=c++0x in hope for more portability
* Milian Wolff: only try to add gnuism when using gnu compiler
* Jonathan Schmidt-Dominé: Fixed old try/recover at end of file bug. BUG: 269832
* Jonathan Schmidt-Dominé: Fixed build error (inactive without FOLLOW_CHECKER_DEBUG)
* Jonathan Schmidt-Dominé: Some autos instead of typeof for better readability
* Jonathan Schmidt-Dominé: SIGSEGV issues now fixed.
* Jonathan Schmidt-Dominé: Revert "Revert "FIRST/FOLLOW-output like before…", it leads to segfaults when trying to"
* Jonathan Schmidt-Dominé: Fixed mismatched calloc/delete (free instead of delete)
* Milian Wolff: Revert "FIRST/FOLLOW-output like before…", it leads to segfaults when trying to build PHP on 64bit.
* Jonathan Schmidt-Dominé: FIRST/FOLLOW-output like before…
* Jonathan Schmidt-Dominé: Fixed infinite recursion in output of follow-dep-checker TODO: fix ‘,’s in output
* Jonathan Schmidt-Dominé: start-rule-detection
* Jonathan Schmidt-Dominé: Fixed follow-set-generation. Mapping in bnf-visitor is now global…
* Jonathan Schmidt-Dominé: Cleaned up kdev-pg-follow
* Jonathan Schmidt-Dominé: Added --no-parser to test.sh and changed expected5
* Jonathan Schmidt-Dominé: Fixed all the license-headers (they talked aboud kdev-pg, some where missing)
* Jonathan Schmidt-Dominé: Cleaned up useless stuff in kdev-pg-first (mItem, mMergeBlocked etc.)
* Andreas Pakulat: Also add the *ast.h to the source list.
* Jonathan Schmidt-Dominé: Easy-Algo works, even bugs from the old implementation are now fixed (too much EOFs)
* Jonathan Schmidt-Dominé: Simple NextFollow-implementation, next step: having a look at the initializers
* Jonathan Schmidt-Dominé: BnfVisitor works, NextFirst ported, wow, that is short!!
* Jonathan Schmidt-Dominé: Relevant parts of the BnfVisitor are now imptemented, warning: visiting will _not_ be acyclic, so there is no need for extra-symbols for kleene-stars.
* Jonathan Schmidt-Dominé: Slarted easy-algo...
* Milian Wolff: fix argument order and offset
* Jonathan Schmidt-Dominé: Removed the last occurences of qDebug... Removed the for-loop for '^'-cursor-positioning in yyerror
* Jonathan Schmidt-Dominé: Fixed error in prefix-operator-code.
* Milian Wolff: break has no arguments
* Jonathan Schmidt-Dominé: Small modification in serialize-visitor.
* Jonathan Schmidt-Dominé: Small modification in serialize-visitor.
* Jonathan Schmidt-Dominé: Fixed Clone-Tree-visitInlinedNonTerminal
* Jonathan Schmidt-Dominé: Fixed GenerateAstRule::visitInlinedNonTerminal
* Jonathan Schmidt-Dominé: Inlining in cool-example.
* Jonathan Schmidt-Dominé: Fixed PrettyPrinter::visitInlinedNonTerminal.
* Jonathan Schmidt-Dominé: Fixed various visitInlinedNonTerminal-methods.
* Jonathan Schmidt-Dominé: Fixed first-set-creation with inlined non-terminals.
* Jonathan Schmidt-Dominé: Added second inline-checker-test
* Jonathan Schmidt-Dominé: Avoid inlining of operator expressions
* Jonathan Schmidt-Dominé: Better DefaultVisitor::visitInlinedNonTerminal (affects the generated code)
* Jonathan Schmidt-Dominé: Use a new token (T_INLINE) for ".=" (instead of '.' '=')
* Jonathan Schmidt-Dominé: Updated case3 (empty-first)
* Jonathan Schmidt-Dominé: Added test for cyclic-inlining-detection.
* Jonathan Schmidt-Dominé: Parsing of scoped-names commented out, because it conflicts with the ".=".
* Jonathan Schmidt-Dominé: Cyclic-inlining-checker added
* Jonathan Schmidt-Dominé: Fixed whitespace
* Jonathan Schmidt-Dominé: Started inlining-support.
* Jonathan Schmidt-Dominé: First step towards more flexibility in operator-expressions.
* Jonathan Schmidt-Dominé: Fixed spelling-error.
* Jonathan Schmidt-Dominé: Better cmake-module.
* Jonathan Schmidt-Dominé: Fixed problem with COPY_CODE_TO_YYLVAL (cstring-functions related, as usual O_o)
* Jonathan Schmidt-Dominé: Better line-numbers in compiler-messages, thank you Victor Vicente for the negative feedback. ;)
* Jonathan Schmidt-Dominé: Fixed message in test/expected(1|2)
* Jonathan Schmidt-Dominé: Added missing reinterpret_cast to the generated Visitor::visitNode-method.
* Jonathan Schmidt-Dominé: Now you can use a switch-statement in Visitor::visitNode instead of a lookup-table.
* Jonathan Schmidt-Dominé: Fixed messages.
* Jonathan Schmidt-Dominé: Test-Cases for all kinds of problems (FIRST/FIRST, empty first, empty operator, undefined token undefined non-terminal additionally to the existing FIRST/FOLLOW-test) Fixed: Missing EmptyOperator-checking Fixed: Ugly output in EmptyFirst (symbols got visited multiple times)
* Jonathan Schmidt-Dominé: Updated TODO and README (there was kdev-pg-related stuff!!)
* Jonathan Schmidt-Dominé: Better output and better testing ({FIRST/FIRST, FIRST/FOLLOW} x {operator-based, bnf-based})
* Jonathan Schmidt-Dominé: A central QTextStream instead of qDebug, there where missing flushes etc., now problem-reporting works fine.
* Jonathan Schmidt-Dominé: Support for --beautiful-code (that is the old-style) and --error-aware-code (when compiling the generated source the compiler messages will contain references to the grammar-file)
* Jonathan Schmidt-Dominé: Use [:#include "bla.h":] instead of [:#include <bla.h>:] in code-generation for user-supplied headers.
* Jonathan Schmidt-Dominé: Op2: A very complex test-case covering a lot of features, could be used as unit-test.
* Jonathan Schmidt-Dominé: Fixed wrong strcpy-stuff, ihgitt!
* Jonathan Schmidt-Dominé: A more complex example for operator-expression-parsing which may be used for a unit-test.
* Jonathan Schmidt-Dominé: Operator-expression-parsing improved -better error handling -more flexibility
* Jonathan Schmidt-Dominé: Removed unnecessary KDE-dependency
* Jonathan Schmidt-Dominé: Added enable_testing
* Jonathan Schmidt-Dominé: Fixed ASBeautifier (++ instead of --) Added kFatal when trying to use the expression-parser with --no-ast.
* Jonathan Schmidt-Dominé: It is krazy:excludeall and not krazy:exclude-all!!
* Jonathan Schmidt-Dominé: Fixed a lot of krazy-issues.
* Jonathan Schmidt-Dominé: AUTHORS...
* Jonathan Schmidt-Dominé: Produce extra-errors in the expression-parser.
* Jonathan Schmidt-Dominé: -Support for ternary expressions -Simplified parentheses
* Jonathan Schmidt-Dominé: Added TernaryExpression to AST-gen and Visitor-gens
* Jonathan Schmidt-Dominé: Fixed the lexer (rulePosition has to be set for operator-rules, too)
* Jonathan Schmidt-Dominé: 1. Support for parentheses 2. AST-definitions get sorted topologically (ast_base etc.) 3. Operator-Expressions without any operators work
* Jonathan Schmidt-Dominé: Added support for prefix-expressions...
* Jonathan Schmidt-Dominé: Fixed start- and endToken
* Jonathan Schmidt-Dominé: Object-layout computation... (experiment, no real code)
* Jonathan Schmidt-Dominé: new-visitor for operator-expressions
* Jonathan Schmidt-Dominé: Some tests: 1-+2*3^5 startToken and endToken are not yet correctly set...
* Jonathan Schmidt-Dominé: startToken and endToken for operator-expressions
* Jonathan Schmidt-Dominé: Removed the "if (!m_indent) " to get good dumps.
* Jonathan Schmidt-Dominé: Binary/Postifx: proper ASTs and proper visiting
* Jonathan Schmidt-Dominé: Visitor+DefaultVisitor
* Jonathan Schmidt-Dominé: -First test-cases for bottom-up-parsing work! -Default visitor
* Fathi Boudra: On Qt only build, set cmake minimum requirement. Use KDE 4 defaults.
* Fathi Boudra: Check KDE4 before adding tests directory. The tests use kde4_add_unit_test.
* Fathi Boudra: find Qt 4 as a fallback when we build with Qt only