ChangeLog

2005-05-23  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/JavaCompilationUnit-15.sdf: new parse table that only
	has CompilationUnit as a start symbol. I was very surprised that I
	still hadn't done this for Java-front, since it is a major
	improvement of performance and error reporting :o  (JFR-32).

	* xtc/jtree-version.str: use JavaCompilationUnit-15 if possible
	(i.e. if the start symbol is CompilationUnit) (JFR-32).
	
	* syn/v1.5/languages/java-15/types/ReferenceTypes.sdf: do not
	allow TypeVars as reference types.
	
2005-05-23  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/types/TypeVariables.sdf: minor change
	in the definition of type parameters to make the structure of
	formal type parameters more close to the structure of ordinary
	type parameters. Example: the new representation is
	TypeParam(Id("T"), None). Was: TypeParam(TypeVar(Id("T")), None).

	* syn/v1.5/languages/java-15/types/ReferenceTypes.sdf: allow, but
	avoid type variables as reference types.

2005-05-23  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/types/ReferenceTypes.sdf: added the
	non-terminal TypeDecSpec, which can be a TypeName or a
	specification of a member type. The qualifier of the member type
	can be parameterized, which doesn't fit in the syntax of
	TypeNames. This fixes JFR-19, which was reported by Valentin
	David.
  
	* syn/v1.5/languages/java-15/types/TypeVariables.sdf: TypeVar is
	no longer an alias of Id. It is now a real non-terminal with a
	production Id -> TypeVar and a constructor TypeVar.

2005-05-23  Martin Bravenboer <martin.bravenboer@gmail.com>

	* test/v1.5/floating-point-literals.testsuite: rewritten testsuite
	to concrete-syntax and added many tests for floating point
	literals.  Fortunately, no bugs showed up.

	* syn/v1.5/languages/java-15/lexical/literals/*: modularized
	definition of the syntax of literals.

	* syn/v1.5/languages/java-15/lexical/literals/FloatingPointLiterals.sdf:
	added support for hexadecimal floating-point literals, which is a
	new feature of Java 1.5.
  
2005-05-20  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/statemens/Statements.sdf: replaced
	the Assert constructor of Java assert statement with
	AssertStm. Reported by Valentin David.

2005-03-05  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/Main.sdf: new main module that
	imports all Java modules and does not declare context-free
	start-symbols.

	* syn/v1.5/Java-15.sdf: declare all context-free start-symbols in
	the main module for creating the parse table. If you import this
	module in an extension, then you will inherit all contxt-free
	symbols. Import 'languages/java-15/Main' if you don't want this.

	* syn/v1.5/languages/java-15/*/*.sdf: removed all context-free
	start-symbols from the specific SDF modules.

2005-01-05  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/stratego-v1.5/Embedded-Java-15.sdf: 'lhs' is rather confusing
	for the left-hand side. Added 'e' as a meta-variable for the
	left-hand side. This is not ambiguous.

2005-01-04  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/stratego-v1.5/Embedded-Java-15.sdf: added a variable for the
	LeftHandSide of an assignment.
	
2004-11-01  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/types/ParameterizedTypes.sdf: allow
	any type as actual type argument. I hate semantic limitations that
	are enforced by the syntax. The make the syntax definition unclear
	and make meta-programming more difficult.
	
2004-10-17  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/stratego-v1.5/Stratego-Java-15.sdf: another interesting
	problem related to heuristic filters and ambiguities. We are all
	writing variables in the following way:

	variables
	  "e" [0-9]* -> JavaExpr {prefer}

	Unfortunately, this can cause many ambiguities! There is no follow
	restriction for the variable name, which means that identifiers or
	keywords might be split. For example, I was hit by 'transient'
	being parsed as 't' 'ransient'. Since variables are often
	preferred, this option was chosen instead of the correct parse:
	the keyword transient.

	So, always put follow restrictions on meta-variables. Usually,
	meta-variables are not allowed to be followed by the identifiers
	of the host language. Maybe we could generate this stuff by
	analyzing the follow restrictions of the object language.

2004-10-16  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/stratego-v1.5/Stratego-Java-15.sdf: removed the quotation
	with explicit disambiguation for a list of block statement. This
	is always ambiguous with a single block statement, which only
	shows up if you disable the heuristic filters of sglr.
	
2004-10-15  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/lexical/Modifiers.sdf: added a
	general sort Modifier.

	* syn/stratego-v1.5/Stratego-Java-15.sdf: added variables and
	quotations for names, modifiers. Several other improvements of the
	Java embedding in Stratego. This embedding really needs to be
	parameterized with the symbols of the host language, since you
	definitely do not want to define this for every host language.

	* java-front.m4: use AS_HELP_STRING instead of AC_HELP_STRING.

2004-09-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* xtc/parse-java.str: now supports the --preserve-comments and
	--preserve-positions options. These options can be used to
	annotate the resulting abstract syntax tree with information on
	the comments and positions in the source file. pp-java supports
	the comment annotations and will include them in the resulting
	Java source file.

2004-08-09  Martin Bravenboer <martin.bravenboer@gmail.com>

	* xtc/jtree2text.str: gone.

	* xtc/pp-java.str: now takes the same release options as pp-java.

	* pp/basic: restored the old pretty-printer for use with the old
	syntax definition.

	* syn/basic/Basic-Java.sdf: new name for Java.sdf. Java.sdf
	suggests that this is the primary syntax definition, which is not
	the case.
	
2004-08-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* Unattractive fix to an ambiguity: the declaration "Foo<Bar>
	fred;" can be parsed as an expression of < and >, or a variable
	declaration. It should be a variable declaration, since relational
	expressions are in fact not allowed as expression
	statements. However, the SDF syntax definition allows any
	expression as statement. This not correct.

	However, I can see no concise way to define the expressions that
	are allowed. Using specific non-terminal names (such as
	Assignment) will break the priority mechanism, since injections to
	expressions have to be defined.

	Unfortunately, this is something that cannot be expressed
	concisely in SDF. It makes using just one sort for Expression less
	attractive.

	For now, I've added an avoid to the expression -> statement
	production. This solves the ambiguity, but I would prefer a
	solution without prefer/avoid. Suggestions are welcome ...
	
2004-08-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/expressions/ArrayCreationExpressions.sdf:
	added support for the unbounded wildcard array base type, which is
	going to be part of 1.5, but is not yet supported by the proposed
	final draft of JSR 14.

	* pp/pp-Java-Expressions.str: pretty-printing of the unbounded
	wildcard array base type.

	* pp/pp-Java.str: fixed pretty-printing of parameterized classes
	and interfaces.
	
2004-08-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* pp/pp-Java.str: ad-hoc fix for an interesting problem:
	Minus(Minus(e)) was pretty-printed as --e and therefore parsed as
	PreDecr afterwards. This cannot be derived from the priorities,
	but maybe the lexical restrictions could be used ... scary!
	
2004-08-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/expressions/UnaryOperators.sdf: the
	priority of a primitive cast "(int) e", is different from the cast
	to a reference type "(Foo) e". In the 'basic' SDF syntax
	definition there was just one production rule for casts and the
	priority of all casts was equal, which is not correct.

	In the 1.5 syntax definition, this has been resolved by
	introducing two different productions (CastPrim and CastRef) and a
	more complex definition of priorities.

	* tests/v1.5/expression.testsuite: many parse-unit tests for
	priorities of unary operators, in particular cast expressions.

	* syn/basic/Java-Expressions.sdf: restored the priority of the
	cast to the unary operator level. This is incorrect, but I think I
	should leave this legacy syntax definition untouched. That is, it
	is obsolete.

	If there is a need for a separate 1.4 syntax definition, then this
	should be a developed by removing the 1.5 constructs from the
	syntax definition. The 'basic' syntax definition is too poor.

	So, don't use Java and Generic-Java. Use Java-15.

	* test/Makefile.parse-unit: invoke parse-unit with the new
	"--no-heuristic-filters" flag. Should be removed if java-front is
	going to be released before StrategoXT 0.12.

	* xtc/parse-java.str: sglr is now invoked with the -fi and -fe
	flags to disable the heuristic filters of SGLR (injection count
	and eagerness).
	
2004-08-08  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/languages/java-15/expressions/Priorities.sdf: parsing
	the jdk1.5.0 beta sources revealed an error in the priority of
	cast expressions. The cast expression should not be in the same
	group as the unary operators: it has higher priority.

	* syn/basic/Java-Expressions.sdf: same fix in the legacy Java
	syntax definition.
	
2004-08-05  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/*: Java 1.5 syntax definition is more or less
	finished. Currently performing tests on the GNU classpath and
	J2SDK sources to find bugs. Interesting detail is that the GNU
	Classpath uses quite some variables with the name 'enum', which is
	forbidden in Java 1.5.

	* test/gnu-classpath/*: upgrade to GNU Classpath 0.10.
	
	* test/gnu-classpath/v1.4,v1.5: test both Java versions (currently
	parsing only: I still need to update the pretty-printer for 1.5).
	
2004-08-03  Martin Bravenboer <martin.bravenboer@gmail.com>
	
	* syn/v1.5/*: JavaFront will use hierarchical SDF module names for
	the new syntax definition of Java 1.5 (Java 2 version 5.0). These
	hierarchical module names are nice, and it reduces the number of
	changes that the guys at CWI need to make if they are going to use
	the JavaFront syntax definitions.

	The structure will be an exact copy of the Java Language
	Specification 3rd edition, which is available in the JSR 14 final
	draft.

	Maybe the version scheme of Java in JavaFront should be based in
	the JLS, since that version scheme appears to be the one and only
	consistent version scheme invented by Sun (1-2-3, that makes
	sense!).
	
2004-08-02  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/v1.5/*: working on full support for Java 1.5. This
	will not be an extension of the syntax definition Java 1.4, since
	the changes in Java 1.5 affect many exisiting production rules in
	1.4.

	* syn/generic/*: the current definition of Generic Java in SDF is
	still available but will not be extended to cover 1.5.
	
2004-07-26  Martin Bravenboer <martin.bravenboer@gmail.com>

	* test/gnu-classpath/Makefile.am: check diff of `parse | pp |
	parse' and `parse' abstract syntax tree. This diff should be a
	very good test of the pretty-printer. The check succeeds, so the
	pretty-printer is safe to use on large codebases.

2004-07-25  Martin Bravenboer <martin.bravenboer@gmail.com>

	* test/gnu-classpath/Makefile.am: Pretty-print and parse all Java
	source files in GN Classpath. All files can now be pretty-printed
	(and parsed afterwards) with java-front. The result is not yet
	compared to the original AST. Hope to add this check (and fix the
	problems that will probably show up) next week.

	* pp/pp-Java-Statements.str: fixes in break, continue and labeled
	statements.

	* syn/basic/Java-Literals.sdf: improved representation of class
	literals. A class literal is now Lit(Class(type)).

	* pp/pp-Java-Literals.str: fixed pretty-printing of class
	literals.
	
	* pp/jtree-parenthesize.str: parenthesize requires latest
	StrategoXT sources (2004-07-25) to handle all priority problems in
	pretty-printing.

	* pp/pp-Java-Statements.str: fixed pretty-printing of final local
	variable declarations.
	
2004-07-24  Martin Bravenboer <martin.bravenboer@gmail.com>

	* pp/Makefile.am: jtree-parenthesize is now a separate tool.

	* pp/pp-java.str: invoke jtree-parenthesize.

	* syn/basic/*: declare all sorts.

	* pp/pp-Java-Literals.str: implemented pretty-printing of octal
	escapes.

	* pp/pp-Java-Literals.str: implemented pretty-printing of Unicode
	escapes.

	* pp/pp-Java-Names.str: fixed pretty-printing of ClassName.

	* pp/pp-Java-Statements.str: fixed the pretty-printing of
	if-then-else. All cases were implemented in separate rules, and as
	a consequence one case (block block) was missing. Cleaner rules
	now.

	* pp/pp-Java.str: improved pretty-printing of all constructs
	separated by commas: use a, b instead of a , b.

	* pp/pp-Java.str: fix in the pretty-printing of multiple variable
	declarations in a single local variable declaration statement.
	
	* pp/pp-Java-Statements.str: fix in the pretty-printing of switch
	statements.

	* pp/pp-Java-Expressions.str: fix in the pretty-printing of
	fields.

	* pp/pp-Java.str: fix in the pretty-printing of
	constant declarations in interfaces.
		
	* syn/basic/Java-Statements.sdf: improved the representation of
	the for statement. This first subterm is now a list of
	expressions or a single local variable declaration. The second
	subterm is an optional expression, the third subterm is a list of
	expressions.

	* test/basic/statements.testsuite: parse-unit testsuite rewritten
	to concrete syntax. All test input is now inline and wildcards are
	used to abstract over the output structure.

	* pp/pp-Java-Statements.str: pretty-printing a for statement no
	longer uses a block if the statement is not a block. Enforcing
	this is a job for a separate tool.

	* pp/pp-Java-Statements.str: finally fixed several known problems
	in the pretty-printing of for statements.
	
2004-07-21  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Stratego-Java.sdf: added the stm meta variable, stm |[
	... ]| quotation and ~stm: anti-quotation, and ~bstm:
	anti-quotation. Reported by Raihan.
	
2004-07-10  Martin Bravenboer <martin.bravenboer@gmail.com>

	* xtc/parse-java.str: added an -s|--start-symbol option to parse
	different start symbols then a full CompilationUnit.
	
2004-07-07  Martin Bravenboer <martin.bravenboer@gmail.com>

	* xtc/parse-java.str: added a --release flag to the parse-java
	util. Use parse-java --release 1.4 to parse classic Java. Default
	is 1.5.

	* xtc/parse-java.str: added some documentation in the form of
	tool-doc.
	
2004-06-30  Martin Bravenboer <martin.bravenboer@gmail.com>

	* pp/pp-Java.str: now uses a generated java-parenthesize
	module. This module places parenthesis only where really
	necessary.

	* pp/pp-Java-Statements.str: improved pretty-printing of
	if-statements.
	
2004-04-20  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Java-Identifiers.sdf: context-free to lexical rejects
	are not allowed (BoolLiteral). Replaced by the token true, false
	and null. Reported by Jurgen Vinju.

	* test/basic/identifiers.testsuite: parse-unit testsuite for
        Java identifiers.	
	
2004-04-07  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/generic/Java-StaticImport.sdf: added production rules for
	the static import facility of 1.5.

	* pp/pp-Java.str: added pretty-print rules for static import.
	
2004-03-16  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/eblock/*: implemented an extension of Generic-Java with
	expression blocks. An Expression block a list of block-level
	statements, followed or preceeded by an expression. This
	expression is the value of the expression block. The block-level
	statements are to be lifted to statements before (PreEBlock) or
	after (PostEBlock) the context expression.

	This feature is extremely useful in code generators.

	* trans/core-lift-eblocks.str: first implementation of EBlock
	lifting. It more or less works, but probably not for nested
	blocks, and anonymous inner classes.

2004-02-24  Martin Bravenboer <martin.bravenboer@gmail.com>

	* pp/*: implemented NamedEscapes in the pretty printer.

2004-02-16  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Makefile.am
	* syn/generic/Makefile.am
	* sig/Makefile.am: produce rtg-nf, not rhg-nf.

2004-02-05  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Java-Literals.sdf: character and string literals are
	no longer lexical syntax.

	Lexical syntax is imploded to Strings in an abstract syntax tree
	and that is not very useful because the rather complex structure
	is completely lost then. When using a separate lexical analysis
	like in conventional parsing techniques this cannot really be
	prevented because the parser just operates on tokens.

	I think we should forget about tokens at all: lexical syntax as it
	used right now in SDF is just syntax without layout and you don't
	want to implode this to a String in all cases. I hope the new
	approach to defining layout for context-free syntax sections in
	SDF3 will solve these problems in a more clean way.

	Currently, by defining literals in syntax sections constructors
	can be attached to them and layout is not inserted by the SDF
	normalizer in pgen. This is technique is also applied in the XML
	syntax definition.

	Why is this useful?
	- AST is now much easier to 'desugar'
	
	- clearer and more correct pretty printing
	
	- improved handling of 'lexical syntax' in the AST: no need to
	escape stuff or add double quotes or whatever.

	- better support for concrete object syntax where meta variables
	are in 'lexical syntax'.

	The change to syntax sections revead an ambigiuity bug in octal
	escapes, which wasn't reported before because of the 'handy'
	filters in SGLR. parse-unit should invoke sglr with all filters
	disabled.

	* pp/pp-Java-Literals.str: added the simple cases for the new
	literal constructs. Still need to implement all constructs.

	* sig/Makefile.am: Stratego signatures for Java and Generic-Java
	are now generated from RHG using stratego-regular tools. sdf2rhg
	can handle the more complex SDF constructs because it operates on
	partially normalized SDF modules. sig2rhg just generates one big
	module, so there are no separate modules anymore.

2004-02-05  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Java-Prefixed.sdf: renamings for some (not all yet)
	Java sorts by prefixing them with Java. This module should be used
	for embedding concrete object syntax in Java.

2004-02-04  Martin Bravenboer <martin.bravenboer@gmail.com>

	* syn/basic/Java-Comments.sdf: added follow restriction on
	optional layout: LAYOUT?  -/- [\/].[\/]. This restriciton is
	related to the one for /* and solves this problem:
	
	+ * {// aa
	}
	  <ClassBody-CF>
	    {
	    <LAYOUT?-CF>
	    <ClassBodyDec*-CF>
	    <LAYOUT?-CF>
	      <LAYOUT-CF>
	        // aa
	
	    }
	  <ClassBody-CF>
	    {
	    <LAYOUT?-CF>
	      <LAYOUT-CF>
	        // aa
	
	    <ClassBodyDec*-CF>
	    <LAYOUT?-CF>
	    }
	
2004-01-31  Martin Bravenboer <martin.bravenboer@gmail.com>

	* test/j2sdk: separate directory for j2sdk testsuites.

	* Version to 0.4pre.

	* Bugs to stratego-bugs@cs.uu.nl

2004-01-31  Martin Bravenboer <martin.bravenboer@gmail.com>

	* Follow restriction for /* */ and LAYOUT?. All of GNU Classpath
	java.* is now accepted by the Java grammar of java-front :) .

	* Version to 0.4.

2004-01-31  Martin Bravenboer <martin.bravenboer@gmail.com>

	* Fixed typo causing unicode escape to be rejected.

	* Added final local variabele declarations.

	* Still 3 of 1035 GNU Classpath java.* don't parse.
	
2004-01-31  Martin Bravenboer <martin.bravenboer@gmail.com>

	* Major performance improvements by removing a horrible reject
	construct on 'InputChar'. This appeared to have an amazing bad
	influence on the performance in modules with newline terminated
	single line comments ( // ... )

	* xtc/files2parse-unit-testsuite: generate a parse-unit testsuite
	from a newline separated list of files.

	* test/gnu-classpath/Makefile.am: test the Java grammar by parsing
	the full GNU Classpath. There are still some minor problems and
	parse-unit causes a segmentation fault after parsing about 500
	files. 

2003-08-29  Martin Bravenboer <mbravenb@cs.uu.nl>

	* Improving the ast of java-front:
	- syn/basic/Java-Identifiers.sdf: Id is now a context-free sort
	 and uses the constructor Id.
	 - ayn/basic/Java-Names.sdf: Name now has a constructor Name.

	 * pp/pp-Java-Names.str, pp/pp-Java.str: using the new ast
	structure.

2003-08-28  Martin Bravenboer <mbravenb@cs.uu.nl>

	* configure.in: java-front requires stratego-regular from now (who
	doesn't? ;) )

	* Makefile.regular: copied from xml-tools. Should not be modified
	(subversion should support links)

	* syn/basic/Makefile.am: install Java.rtg-nf

2003-02-25  Martin Bravenboer <martin@mbravenboer.org>

	* pp/Makefile.am: Include sdf/gpp, not sdf/box-tools.

2002-11-23  Martin Bravenboer <martin@mbravenboer.org>

	* Small changes to configure.in to profit from new autotools.

2002-11-20  Martin Bravenboer <martin@mbravenboer.org>

	* Ad-hoc solution for parentheses around expressions in
	for-loop. A more powerfull solution for priorities must fix this
	later.

2002-11-20  Martin Bravenboer <martin@mbravenboer.org>

	* Added the last missing pretty printing constructs.  Thanks to
	Stefan Reich for mentioning this problem.

2002-11-18  Martin Bravenboer <martin@mbravenboer.org>

	* Added priority for array access. I have to figure out if this
	according to the JLS.

2002-09-20  Martin Bravenboer <martin@mbravenboer.org>

	* Removed the priorities of yesterday. Eelco Visser pointed out
	that because of the interpretation of priorities a while in the
	then branch of an if-else is no longer allowed. A prefer on the
	if-then solves everything.
     
2002-09-19  Martin Bravenboer <martin@mbravenboer.org>

	* Added context-free priorities for while, for, and the labeled
	statement. if-then-else > if-then doesn't solve all dangling else
	problems. Now it should work (unit-tests available in de statement
	test-suite).

2002-09-14  Martin Bravenboer <martin@mbravenboer.org>

	* Added a production for generic methods in interfaces. I forgot
	this one.

2002-08-29  Martin Bravenboer <martin@mbravenboer.org>

	* Implemented pretty-printing of while.

	* Implemented pretty-printing of (Q)NewInstance with class-bodies
	for Generic Java and classic Java

2002-08-27  Martin Bravenboer <martin@mbravenboer.org>

	* configure.in improved after some tips of Merijn de Jonge.

	* Added package definition (created by Merijn de Jonge) for the
	Online Package Base. java front and the various required tools can
	now be downloaded as a bundle from the Online Package Base.

2002-08-26  Martin Bravenboer <martin@mbravenboer.org>

	* Added a script "parse-java" to parse Java or Generic Java code.

	* Moved all scripts to "scripts"

2002-08-25  Martin Bravenboer <martin@mbravenboer.org>

	* Fixed huge priority bug with method invocation and field access:
	Expr "." Id must have highest priority

2002-08-24  Martin Bravenboer <martin@mbravenboer.org>

	* Added anti quotation for literals to Stratego-Java.sdf

2002-08-22  Martin Bravenboer <martin@mbravenboer.org>

	* Modified pp-Java-Literals to workaround concrete syntax bug in
	Stratego.

	* Implemented pretty-printing for the if construct and throw in
	pp-Java-Statement.

2002-08-21  Martin Bravenboer <martin@mbravenboer.org>

	* Implemented pretty-printing in pp-Generic-Java.

	* Created Generic-Java.sdf, a testsuite and Stratego signature
	generation for Generic-Java.

	* Restructured method invocation

	* Removed distinction TopLevel/Member/Local TypeDecs

2002-08-20  Martin Bravenboer <martin@mbravenboer.org>

	* Added LocalClassDec -> BlockStm

2002-08-19  Martin Bravenboer <martin@mbravenboer.org>

	* Implemented pretty-printer

2002-08-05  Martin Bravenboer <martin@mbravenboer.org>

	* Implemented parse-unit and started a decent testsuite

2002-08-04  Martin Bravenboer <martin@mbravenboer.org>

	* Solved the array creation <-> array access ambiguity with a
	reject.

2002-07-30  Martin Bravenboer <martin@mbravenboer.org>

	* Added priorities for binary expressions.

2002-07-29  Martin Bravenboer <martin@mbravenboer.org>

	* Initial version of Java-Expressions module.

2002-07-28  Martin Bravenboer <martin@mbravenboer.org>

	* Work on JLS 2.0: inner classes and primaries.

2002-07-27  Martin Bravenboer <martin@mbravenboer.org>

	* Initial version of the GNU build system powered version of Java
	Front for XT.