$Id: ChangeLog 21317 2007-02-14 08:37:12Z jurgenv $

Wed Feb 14 08:50:33 CET 2007 <jurgenv@cwi.nl>
 * sdf-checker/src/Sdf-Checker.c: regenerated after fixing a weird bug (import
   relation was inverted by an equation at post-check time). Also, I improved
   a number of error messages, by explaining in the error message why an attribute is 'illegal'.


Tue Jan 30 07:52:14 CET 2007 <jurgenv@cwi.nl>
 * src/Sdf2-Normalization.c: forgot to implement renaming for non-transitive
   priorities.

Fri Jan 26 09:28:10 CET 2007 <jurgenv@cwi.nl>
 * src/parsetablegen.c: adapted to new interface of toasfix function.
 * removevarsyntax/src/removevarssyntax.c: adapted to new interface of toasfix function.
 * sdf-checker/src/sdfchecker.c: adapted to new interface of toasfix function.

Thu Jan 25 19:07:17 CET 2007 <jurgenv@cwi.nl>
 * tests/grammars: added Backquote test grammar

Tue 2007-01-23    <storm@cwi.nl>
  * src/Sdf2-Normalization.c: incorporated fixes wrt lifting of 
    sorts (only completely uppercase sorts with no dashes were 
    correctly lifted to syntax). (closes #694)

Thu Jan 18 18:12:59 CET 2007 <martin.bravenboer@gmail.com>

  Fixes for Bug 685: Associatvity no longer applied to parse table

  * src/ksdf2table.c: the elements of 'nonTransitivePrioEntries'
    should be added to the priority_table, which should happen after
    determining the transitive closure.
  * grammar/Sdf2.baf: regenerated for the bug fixed by this commit.
  * test/tables/*.tbl: regenerated the parse tables use for
    regression testing.
  * pgen desperately needs a better testsuite. In particular, the
    various filters should be tested at a more fined-grained level. In
    this case, the bug did not popup earlier because of the post-parse
    priority filter that still resolved the ambiguities. For this
    particular issue that is now a test in Stratego/XT, but pgen
    itself should really check this.

Fri Jan 12 15:35:46 CET 2007 <jurgenv@cwi.nl>
 * grammars/Sdf2.baf: introduced a new operator for priorities. The "."
   can stop (full stop) the transitive closure of priorities. Example:
   A "+" A -> A > A "*" A -> A . > A "/" A -> A
                              ^^^
			      This dot makes that + is not bigger than /.
 * src/Sdf2-Normalization.c,src/ksdf2table.c: implemented the semantics
   of the . operator. 
 * src/Sdf2-Normalization.c: the normalizer translates associativities
   to non-transitive priorities now. This solves bug #678 which was
   introduced by the mapping of associativity to priority in specific
   arguments.
 * tests/tables/*: tables contain much less priorities now since the
   associativities are no longer closed under transitivity. This changes
   both the action table and the list of priorities. The regression
   tests for the terms parsed with the new tables did not need any
   updating.

Wed Sep 13 15:54:37 CEST 2006 <economop@cwi.nl>
 * src/Sdf2-Normalization.c: regenerated because the normalisation was
   changed.

Wed May  3 14:00:05 CEST 2006 <jurgenv@cwi.nl>
 * Bumped to 2.4

Sat Apr 22 19:12:02 CEST 2006 <jurgenv@cwi.nl>
 * src/Sdf2-Normalization.c: And yes indeed, see previous logmessage, there was 
   a mistake in the last bugfix and I should have seen this from the
   tables. Fixed the bug now, after regular, priorities are generated
   that need to be normalized. So in fact, priorities depend on regular.
   Added the dependency and the call from r[[.]] to p[[. ]].
 * test/grammars/List.sdf: a test designed to test the dependency between
   r[[.]] and p[[.]]

Wed Apr 19 20:41:30 CEST 2006 <jurgenv@cwi.nl>
 * src/Sdf2-Normalization.c: regenerated again. There were serious
   errors in a number of locations:
     - regular expressions were not expanded from productions generated
       out of context-free priorities sections. This was because p[[.]]
       was applied after r[[.]].
     - the Character table was not complete enough, it still depended on 
       the old type-less implementations of characters to work correctly.
     - literal expansion was not syntax-safe, error were generated by the
       asf-checker.
 * src/Sdf2-Normalization.c: regenerated after fixing a serious bug
   in the normalization of groups of productions with selection
   of a particular argument for a priority.
 * NOTE: the parse tables generated are all different now, but the
   parse trees that come out are the same. I choose to conclude that
   my changes must be correct, but I'm not completely sure. There
   are many factors involved here: bugfixes in the ASF compiler,
   the new structured lexicals features, bugfixes in the SDF Normalizer,
   and so forth.

Wed Apr 19 20:37:59 CEST 2006 <jurgenv@cwi.nl>
 * test/**: added regress-terms.sh.src and all regression tests from
   the sglr package. Regenerated the .org files, since they may be
   different. Here will we test the parse trees with new parse tables
   every time. This is practical for one thing only: if the parse tables
   are not the same anymore, but the parse trees are still the same
   this is a good hint that the change in the parse table was not wrong.
 

Fri Mar 10 11:36:53 CET 2006 <kooiker@cwi.nl>
 * **/Makefile.am:
   - Maintainers use ATerm-dbg library

Tue Feb 28 11:13:30 CET 2006 <kooiker@cwi.nl>
 * {removevarsyntax/src/removevarsyntax, sdf-checker/src/sdfchecker}.c:
   - printing default usage exits with exit value 1

Tue Feb 21 15:07:51 CET 2006 <jurgenv@cwi.nl>
 * pgen/Sdf2-Normalization.c: reactivated promoting productions
   that are part of priority grammars to syntax grammars. This to
   prevent the code duplication that occurs in the C and Java Expression
   grammars. The original reason to turn this off were the fact that
   productions attributes are merged by the normalizer, such that
   sometimes productions that look the same, are not the same. This lead
   to matching failures. Another solution (static checking) needs to be
   introduced to fix that problem.

Tue Feb 21 13:29:51 CET 2006 <jurgenv@cwi.nl>
 * sdfchecker/src/*: closed bug 464, added full support for
   renamings.

Tue Jan 31 17:55:26 CET 2006 <jurgenv@cwi.nl>
 * removevarssyntax: This tool removes the variable definitions
   from an SDF module or definition. It used to be located in the
   package for ASF, but since it is ASF independent I've put it here.

Wed Jan 18 16:29:16 CET 2006 <{jurgenv,kooiker}@cwi.nl>
 * src/Makefile.am,package,configure.in: got rid of transitive
   dependencies of asc-support by using asc-support-config script
   at configure time.

Wed Jan  4 17:42:45 CET 2006 <jurgenv@cwi.nl>
 * normalizer: Implemented case-insensitive literals.
 * tests: added test for ci literals.

Fri Dec 23 10:32:37 CET 2005 <jurgenv@cwi.nl>
 * grammar/Sdf2.baf: added syntax for case-insensitive literals.
   A literal spelled with single quoted, like 'true', generates
   [Tt][Rr][Uu][Ee] -> 'true', such that true can be spelled with
   any combination of lowercase and uppercase characters.
   TODO: regenerate normalizer

Fri Nov 18 16:17:12 CET 2005 <kooiker@cwi.nl>
 * src/sdf2table.src:
   - Output filename does not contain .def. anymore when using definition as 
     input

Thu Aug 18 14:24:25 CEST 2005 <jurgenv@cwi.nl>
 * grammar/Sdf2.baf: regenerated the parse table with a different
   basic/Comments module such that comments and todo's are put in
   their categories (for syntax highlighting)

Tue Aug 16 16:43:30 CEST 2005 <jurgenv@cwi.nl>
 *  sdf-checker/spec/*.{asf,sdf}   The checker accumulates the
    errors and warnings now PER MODULE, such that only the errors for
    a specific module may be extracted, although the entire specification
    had to be analyzed to come up with these errors.


Mon Aug 15 16:37:59 CEST 2005 <jurgenv@cwi.nl>
 * sdf-checker/spec/*.{asf,sdf}: 
   - used utilities/Parsing to map subjects of errors to string. This 
     replaces the manual "to-str" functions.
   - Renamed some error messages, removing "in module" postfixes,
     and adding descriptions of subjects.
   - For all 'double declaration'-like errors and warnings, added
     location for both instances such that the user can click on them.
   - The right-hand side of renamings are now interpreted as 
     definitions of symbols, such that you do not need to give a 
     dummy production to suppress a warning message.

Mon Aug 15 15:08:29 CEST 2005 <jong@cwi.nl>
 * configure.in:
   - add libtool support

Mon Aug 15 15:15:33 CEST 2005 <kooiker@cwi.nl>
 * src/ksdf2table.{c,h}:
   - removed extern declaration of afun_range and afun_char_class
   - added getter function for afun_char_class

Mon Jul  4 10:12:41 CEST 2005 <economop@cwi.nl>
 * Corrected an accidental commit of |first.c|, |first.h| and |goto.c| in the 
   previous commit.
	
Mon Jul  4 10:12:41 CEST 2005 <economop@cwi.nl>
 * Removed -m option that was passed to sglr from |sdf2table.src|.
	
Tue May  3 15:29:20 CEST 2005 <markvdb@cwi.nl>
 * Fixed a problem with respect to converting sorts into strings.

Tue Apr 12 17:05:18 CEST 2005 <markvdb@cwi.nl>
 * Regenerated bootstrap grammar in order to incorporate
   the new prefix function syntax.
 * Upgraded the Sdf-Checker wrt structured lexicals.

Mon Feb 28 16:34:22 CET 2005 <{jurgenv,markvdb}@cwi.nl>
 * Added syntax for chain priorities with specific argument selection.
 * Implemented transitive closure for priorities with argument selection.
 * Regenerated bootstrap grammar.

Wed Jan 26 11:25:22 CET 2005 <jurgenv@cwi.nl>
 * SGLR now produces structured lexicals because of changes in the 
   flattener, also literals are no longer mapped to lit(<str>) trees.
 * Structured lexicals in ASF, means a reimplementation of some of 
   the functionality in the SDF normalizer. Nothing big though.
 * Regression tests succeeed without ANY changes.
 * Bootstrapped SDF2 normalizer and the Sdf2.baf parsetable
 * TODO: reimplement and regenerate SDF-checker!!!!! (sdf2table is now broken)

Wed Dec 29 08:58:03 CET 2004 <jurgenv@cwi.nl>
 * Refactored SDF2 syntax definition:
   - ATerm syntax now uses StrCon and IdCon for literals
   - SDF2 does not reuse ATerm literals, but uses StrCon and IdCon
     where appropriate (such as in labels, prefix functions).
   - Unquoted literal symbols are syntactically disallowed from now on!
   - Introduced new non-terminal: ATermAttribute, such that associativity
     and other attributes (avoid, reject, bracket, prefer)
     can be rejected from it. Nested occurrences (inside an ATerm attribute)
     are now possible.
 * Changed Sdf2-Normalization accordingly: 
   - Quoting literals is not needed anymore.
   - Normalization of literals simplified.

Fri Nov 19 14:22:07 CET 2004 <jurgenv@cwi.nl>
 * Changed dependency on Tide to dependency on small package tide-support.
   (similar to toolbuslib a very small C package)
 * Changed in Sdf-Checker by mark have not been merged yet.

Mon Nov 15 13:18:02 CET 2004 <markvdb@cwi.nl>
 * Merged 1.5 and trunk.

Wed Nov 10 12:05:55 CET 2004 <markvdb@cwi.nl>
 * Improved the bug fix.

Tue Nov  9 20:25:52 CET 2004 <markvdb@cwi.nl>
 * Regenerated the Sdf-Normalizer in order to solve bug #404:
   fixed the manipulation of renamings.

Wed Sep 29 15:31:05 CEST 2004 <markvdb@cwi.nl>
 * Moved the check on cons attributes from the SdfChecker
   to the AsfSdfChecker.

Mon Oct 11 17:42:09 CEST 2004 <jurgenv@cwi.nl>
 * Temporarily added dependency on Tide package. In the near future
   the small library that is needed from tide will be separated out
   in a small package.

Thu Sep 16 11:56:37 CEST 2004 <jurgenv@cwi.nl>
 * Bumped to 2.3

Thu Sep  9 10:27:33 CEST 2004 <markvdb@cwi.nl>
 * Improved the lowel level handling of symbol sets in order
   to improve the efficiency.

Fri Aug 27 14:56:50 CEST 2004 <jong@cwi.nl>
 * refactor to use new PT_yield functions in MEPT API

Thu Aug 19 11:37:20 CEST 2004 <markvdb@cwi.nl>
 * Fixed a bug (366) in the sdf-checker with respect to renaming
   of sorts, renamed sorts are no longer reported as
   being non declared.

Fri Jul 16 15:18:45 CEST 2004 <markvdb@cwi.nl>
 * Added a warning on missing exported start symbols.

Fri Jul  9 11:54:37 CEST 2004 <markvdb@cwi.nl>
 * left-/right-/assoc-/non-assoc-prios have been replaced
   by numbered priorities.
 * Increased the version number of the parse table to 6.
 * Regenerated the tests due to change of version number.

Mon Jul  5 13:01:00 CEST 2004 <markvdb@cwi.nl>
 * Bug fix wrt the translation of non-assoc in number priorities.

Fri Jul  2 17:06:09 CEST 2004 <markvdb@cwi.nl>
 * Fixed a serious bug wrt missing imports.

Thu Jul  1 15:26:09 CEST 2004 <markvdb@cwi.nl>
 * Different translation of (kernel/lexical/context-free) start-symbols
   into production rules of the form X -> <START>

Mon May 17 11:34:26 CEST 2004 <markvdb, jurgenv@cwi.nl>
 * Increased version number in parsetable.
 * Added the arg-gtr-prio in order to be more selective
   in priority definitions.

Fri May 14 11:55:56 CEST 2004 <jurgenv@cwi.nl>
 * Bumped to 2.2

Wed May 12 14:09:17 CEST 2004
 * Fixed bug 297, message wrt to use of CHAR were not clickable.

Thu May  6 21:09:33 CEST 2004 <markvdb@cwi.nl>
 * Fixed the sdf2table wrt the SdfChecker.

Tue May  4 13:23:03 CEST 2004 <markvdb@cwi.nl>
 * Improved the clickable error messages and solved bug 295.

Fri May  7 11:22:40 CEST 2004 <jurgenv@cwi.nl>
 * Bumped to 2.1 because 2.0 was released before

Mon May  3 16:28:32 CEST 2004 <markvdb@cwi.nl>
 * Introduced clickable error message.

Tue Apr 27 16:45:27 CEST 2004 <markvdb@cwi.nl>
* Fixed the erroneous error messages generated in case of circular
  imports, see bug 219.

Fri Apr 23 17:11:13 CEST 2004 <markvdb@cwi.nl>
 * Declared symbols in the sorts section are no longer
   promoted to start symbols. Regenerated the Sdf2-Normalizator.

Thu Apr 15 13:03:25 CEST 2004 <markvdb@cwi.nl>
 * Removed dependency on sglr at build time.

Fri Mar 19 15:17:59 CET 2004 <j{ong,urgenv}@cwi.nl>
 * updated wrt changes in SDFME API

Wed Apr 14 10:53:05 CEST 2004 <markvdb@cwi.nl>
 * The use of CHAR in context-free and lexical syntax productions
   yields a warning.
 * Removed a non-terminating rewriting, fixed bug 125.

Thu Mar 11 16:54:31 CET 2004 <markvdb@cwi.nl>
 * Improved an error message with respect of the definedness of a start
   symbol.

Wed Mar 10 16:57:15 CET 2004 <{jong,kooiker}@cwi.nl>
 * Use new Error API

Tue Feb 10 15:27:49 CET 2004 <m336@cwi.nl>
 * upgrade auto* to: autoconf 2.59, automake 1.7.9, libtool 1.5.2

Mon Feb  2 16:43:59 CET 2004 <{jong,kooiker}@cwi.nl>
 * fix Makefile.am to get "distcheck" working

2004/01/26 (markvdb@cwi.nl)
 * Respecified the Sdf-Checker towards the new error api.

2004/01/07 (jurgenv@cwi.nl)
 * Added dependency on error-support because asc-support needs it.

2003/12/18 (markvdb@cwi.nl)
 * Removed sdf2stats and improved the Sdf-Checker.

2003/12/17 (markvdb@cwi.nl)
 * Adapted the Sdf-Checker to the new Sdf syntax.

2003/12/15 (markvdb@cwi.nl)
 * Added a first version of an Sdf upgrade tool.

2003/12/11 (markvdb@cwi.nl)
 * Introduced the {lexical, context-free} start-symbols in the Sdf2 grammar. 
 * Adapted the Sdf2-Normalization accordingly.
 * Both sorts and start-symbols are now used as start symbols.

Mon Dec  8 14:39:46 CET 2003 <jong@cwi.nl>
 * improved "deprecated-symbol" warning for tuple syntax
 * sdf-checker/spec/*.asf: upgraded condition syntax

Thu Dec  4 15:17:29 CET 2003 <jong@cwi.nl>
 * improved "deprecated-symbol" warning: the symbols aren't deprecated at
   all, it's the notation (unquoted) that is deprecated. The message is
   now: deprecated-unquoted-symbol-notation

2003/10/17 (markvdb@cwi.nl)
 * Added a message on deprecated unquoted literals.
 * Added a message on usage of non Sort symbols in
   right hand side of lexical functions.

2003/10/09 (markvdb@cwi.nl)
 * Added linking of PTMEPT library to parsetablegen

2003/10/09 (markvdb@cwi.nl)
 * Fixed the depricated error message.
 * Regenerated due to change in literals.
 * Added the strategy symbol again.

2003/10/08 (markvdb@cwi.nl)
 * Adapted package dependencies.
 * Bumped into version 2.0.

2003/10/07 (markvdb@cwi.nl)
 * Introduced the new tuple syntax <A,B>.
 * Added a check to the sdf-checker which gives a warning
   when A # B is used.
 * Both variants exist next to each other, for some time!

2003/10/03 (markvdb@cwi.nl)
 * Removed the set and perm symbol in the library.

2003/10/02 (markvdb@cwi.nl)
 * Adapted to the new location of the Sdf2 library.
 * Removed quite a number of warning generated by the asfsdf-checker
   with respect to traversal usage.
 * Added an avoid to the bracket production of CharClass. This
   fixes bug #57.

2003/10/01 (markvdb@cwi.nl)
 * Regenerated the normalizer for the Sdf definition with constructors
   and labels.

2003/09/26 (markvdb@cwi.nl)
 * Regenerated Sdf2-Normalization.c given the new library and
   layout convention.

2003/08/01 (jurgenv@cwi.nl)
 * Set version to 1.7.1

2003/07/24 ({jurgenv,markvdb}@cwi.nl)
 * Added lexical restrictions on ModuleWord and ModuleId.
 * Added a warning when literals are detected in the right hand side
   of productions rules.

2003/07/02 (markvdb@cs.vu.nl)
 * The type (trm/eqs) is now made explicit.
 * Removed extension completely.

2003/07/02 (jurgenv@cwi.nl)
 * Release 1.7

2003/06/26 (jurgenv@cwi.nl)
 * Added package definition file

2003/05/19 ({markvdb,jong}@cwi.nl)
 * Added explicit ATBpack and ATBunpack.

2003/04/29
 * Fixed the annoying bug wrt unnecessary warnings in case of parameters.

2003/02/11
 * Bumped into 1.7beta.
 * Generate Sdf2.def from the asf/library Sdf2 definition.
 * Moved Sdf2.sdf from test/first to test/grammars.

2003/01/24 ({jong,jurgenv}@cwi.nl)
 * Removed obsolete ToolBus requirements

2003/01/23 (vinju@loria.fr)
 * Release 1.6
 * LGPL license

2002/11/05 (vinju@loria.fr)
 * Link and include SDF2PT library 

2002/10/29 (vinju@loria.fr)
 * Removed dependency on ToolBus package by removing implicit build rules
   for ToolBus interfaces

2002/10/04 (vinju@loria.fr)
 * sdfchecker brought up to date with changed pt-support interface

2002/09/26 (vinju@loria.fr)
 * Bumped to 1.6beta

2002/07/24 (jurgenv@cwi.nl)
 * Release 1.5

2002/07/11 (jurgenv@cwi.nl)
 * Moved entire normalizer specification to the asf-library such that 
   (parts of it) can be reused by other tools.

2002/07/02 (jurgenv@cwi.nl)
 * Added a lifting operator to SDF2. Example: `A` is replaced by "A" by
   the normalizer, or `A?` is replaced by ("A" "?") etc. This is a nice
   feature for languages with user-defined syntax.

2002/06/24 (vandenbr@loria.fr)
 * Fixed some bugs in the Sdf-Checker with respect to handling
   of circular import relations.
 * Removed an inefficiency bug revealed by Action Semantic specification.


2002/05/22 (vandenbr@loria.fr)
 * Bumped into version 1.5beta.
 * Based the specification on the Sdf definition in the new asf-library.

2002/04/19 (jurgenv@cwi.nl)
 * Release 1.4

2002/04/18 (vandenbr@cwi.nl)
 * sdfchecker is now "Renaming compliant".
 
2002/03/25 (vandenbr@cwi.nl)
 * Added -s switch to sdf2table to call the sdfchecker.

2002/03/22 (vandenbr@cwi.nl)
 * Improved the handling of aliased symbols.

2002/03/22 (vandenbr@cwi.nl)
 * Added a simple check in the completeness of the definition.
 * Moved the Sdf part of the sdf-checker from meta to pgen.

2002/03/13 (jurgenv@cwi.nl)
 * Fixed another bug in the normalizer. When user-defined lexical
   syntax produces a character class a cyclic injection was introduced.
   For example: "foo" -> [a-z], introduced [a-z] -> [a-z] which is
   a degenerated injection from lex to cf.
 * Fixed a bug in the parsetable generator. User defined productions into
   character classes were ignored at generation time.

2002/03/05 (jurgenv@cwi.nl)
 * Fixed a bug in the normalizer. Two equations:
   [0] <cc -VAR> = cc
   [0] <L  -VAR> = L
   ruined the syntax generated for lexical variables,
   I removed them.

2002/03/04 (vandenbr@loria.fr)
 * Introduced parameterized sorts.
 * Bumped into version 1.4beta.

2002/02/07 (jurgenv@cwi.nl)
 * Release 1.3

2002/01/28 (vandenbr@cwi.ni)
 * Very serious bug fixed in relation with parameterization and
   context-free/lexical restrictions.

2001/12/13 (jurgenv@cwi.nl)
 * Bumped version of parsetablegen to 4.2
 * Added asf-support library linkage (needed for normalizer spec).
 * Added --with-asf-support

2001/11/29 (jurgenv@cwi.nl)
 * Adapted the normalizer such that in the attributes of productions 
   ATerm functions are no longer quoted automatically.
 * Bootstrapped the normalizer and the sdf2 table after that.
 * Regenerated regression tests.
 * Bumped version of parsetablegen to 4.1

2001/11/25 (mdejonge@cwi.nl)
 * src/Makefile.am: added miussing generated .h file to EXTRA_DIST variable
 * package now uses toolbuslib package
 * extended functionality of program-check.m4 macros

2001/11/23 (vandenbr@loria.fr)
 * fixed a bug: the character classes for lexical
   definitions with the same rhs symbol where not merged.
   [a-z]    -> Id
   [a-zA-Z] -> Id
   yields ambiguities.
 * Deactivated this bug fix, left the code in the specification.

2001/11/13 (jong@cwi.nl)
 * released version 1.2

2001/11/13 (markvdb@cwi.nl)
 * Removed unused argument from applyFunctionToTree.

2001/11/07 (markvdb@cwi.nl)
 * Version number introduced in the parse table format.

2001/11/05 ({jurgenv,markvdb}@cwi.nl)
 * Removed module identifiers from productions:
   - adapted the normalization specification.
   - regenerated the regression grammars. 
   - regenerated the bootstrap table.

2001/10/11 (mdejonge@cwi.nl)
 * Added (generated) manual pages to EXTRA_DIST variable because not
   everybody has pod2man available

2001/10/01 (jurgenv@cwi.nl)
 * Bumped version of parsetablegen to 3.1
 * Bumped version of sdf2table to 1.11

2001/09/27 (jurgenv@cwi.nl)
 * Made some important changes in the SDF grammar:
     - added:   ATerm                      -> Attribute 
     - removed: "cons" "(" term:ATerm ")"  -> Attribute 
                "constructor"              -> Attribute 
                "memo"                     -> Attribute 
                "traverse"                 -> Attribute 
     - refactored the ATerm grammar a bit
 * NOTE: These changes affect any tool using SDF or AsFix[2,me].
 * Any ATerm can be an attribute now.
 * Bootstrapped the Sdf2.baf parsetable and the normalizer for the new SDF 
   grammar and the corresponding AsFix attribute format. See the ChangeLog
   of pt-support for details.
 * Due to the changes in SDF, several changes in AsFix have been made as well.
   See the pt-support package for details. Parsetables generated with the new
   SDF/AsFix have not changed, only the format of the productions is 
   slightly changed. 

2001/08/28 (jong@cwi.nl)
 * Added proper #include directives to satisfy gcc-3.0.1

2001/08/22 (mdejonge@cwi.nl)
 * grammar/Makefile.am: added -m switch to sglr to explicitly instruct sglr
   to produce asfix2me as output.

2001/08/22 (jong@cwi.nl)
 * Upgraded assertion that char-ranges fall within Latin-1 set to proper
   error message.

2001/08/08 ({jong,olivierp}@cwi.nl)
 * Fixed some serious memory leaks
 * Fixed int hnr problems (should be unsigned)
 * Added support for compiling and linking against dmalloc

2001/08/07 ({jong,olivierp}@cwi.nl)
 * Added rudimentary support for multiple input files to support memory
   profiling

2001/08/01 ({jong,markvdb,olivierp}@cwi.nl)
 * Improved efficiency some more.

2001/07/27 ({markvdb,jong}@cwi.nl)
 * Fixed three memory leaks
 * Fixed a free/unprotect bug.
 * Fixed an initialization/protect bug.

2001/07/2{4-6} ({markvdb,jong,olivierp}@cwi.nl)
 * Improved pgen efficiency and regenerated parsetables

2001/07/20 ({markvdb,jong,olivierp}@cwi.nl)
 * Switched to new parsetable format
 * Introduced character classes instead of aterm lists in order
   to enable future optimizations.
 * regenerated parsetables.

2001/07/18 (markvdb@cwi.nl, jurgenv@cwi.nl)
 * Fixed a bug with respect to the renaming of labelled symbols.
 * Regenerated normalizer with new compiler, the performance gain
   is significant.

2001/07/17 (jong@cwi.nl)
 * version 1.0 released, bumped to 1.1

2001/07/12 (jong@cwi.nl)
 * patched Makefile.am's to comply with new location of -lATB

2001/07/12 (markvdb@cwi.nl)
 * Added the transitive closure calculation on chain prios.
 * Generated new reference table for the regression tests.

2001/07/11 ({markvdb,olivierp}@cwi.nl)
 * Moved priority uniqueness calculation from spec to c phase
 * Removed reachability analysis from spec

2001/06/07 (markvdb@cwi.nl) 
 * Bumped version to 1.0!
 * Removed C code generated by old compiler.
 * Added C code generated by new compiler. 
 * Adapted flatten code (for LookAheads) and flattenProd replaced by function
   from sdf-support library.
 * Normalization functionality moved to parsetablegen (activitated via "-n").
 * Adapted scripts in order to deal with -n.

2001/04/07 (markvdb@cwi.nl)
 * removed the function filter-attributes which was used
   to "propagate" avoid and prefer over generated injection chains.

2001/05/30
 * Fixed a bug with respect to the Symbol Int+ construction, the
   normalizer still used the old notation {Symbol}Int+.
 * Generated new C code.
 * Bumped version number of parsetablegen into 2.4.

2001/05/23 (markvdb@cwi.nl)
 * Released version 0.8 and bumped into 0.9.

2001/04/27 (markvdb@cwi.nl)
 * Added -m to normalizer.
 * Regenerated some checks due to next point.
 * Context-free functions defined in priority section are no
   longer promoted to ordinary syntax section. Thus it is no
   longer possible to define a (context-free) function by giving
   only the priority declaration.
 * Version number increased to 0.8!

2001/03/15 (markvdb@cwi.nl)
 * Small bug removed in normalizer which causes r[[LAYOUT]] to occur
   as a non normalized term.
 * New test tables generated because the order in the priority section
   differed slightly.

2001/03/08 (jurgenv@cwi.nl)
 * normalizer based on PT and SDF library

2001/02/01 (mdejonge@cwi.nl)
 * reconf and DEVELOPERS-INFO files are now also distributed
 * improved generation of sdf2 parse table process in ./grammar

2001/02/01 (jurgenv@cwi.nl)
 * regress-test moved to test directory

2001/01/15 ({markvdb,jurgenv}@cwi.nl)
 * Bumped version of parsetablegen to 2.3
 * Removed gen-table script. From now on use sdf2table only.
 * Added -m option to specify topmodule for parsetable generation.

2001/01/05 (markvdb@cwi.nl)
 * Bumped version of parsetablegen to 2.2.
 * Added constructor annotations to relevant function in normalization spec.
 * Added a tool to derive some statistics on the original Sdf2 definition.

2000/12/20 (markvdb@cwi.nl)
 * Bumped version to 0.7
 * Removed -sf options from scripts.
 * Added statistics calculation.
 * Added -l option to scripts.

2000/12/14 (markvdb@cwi.nl)
 * Removed the actual parse table generation via the compiled specification.
 * Normalization function moved to other module (Sdf2-Normalization).
 * Bumped version of parsetablegen to 2.1.

2000/11/08 (markvdb@cwi.nl)
 * Bumped version to 0.6
 * Changed implicit priorities of lists slightly.

2000/10/26 (mdejonge@cwi.nl)
 * sdf2table now tries to create unique file names in /var/tmp to prevent
   write permission errors.

2000/10/02 (markvdb@cwi.nl)
 * A small correction with respect to priorities of lists.

2000/4/26 (markvdb@cwi.nl)
 * Code added for dealing with {<symbol>}<nr>+ and {<symbol> <symbol>}<nr>+
   Sdf2 constructions.      

2000/4/25 (markvdb@cwi.nl)
 * upgraded to version 0.4
   destroyed.

2000/4/13 (mdejonge@cwi.nl)
 * no longer perform normalize-normalize-check

2000/4/11 (markvdb@cwi.nl)
 * Added removevars tool

2000/4/10 (mdejonge@cwi.nl)
 * Added SDF2 normalizer to pgen distribution

2000/4/06 (mdejonge@cwi.nl)
 * Fixed include file search path problem by reoredering the search paths

2000/4/05 (mdejonge@cwi.nl)
 * Version is now 0.3
 * Adapted for use with separate asfix distribution (i.e., libasfix is no
   longer distributed as part of the aterm library).

2000/3/31 (mdejonge@cwi.nl)
 * The toolbus is no longer required during the build process of pgen

2000/03/29 (mdejonge@cwi.nl)
 * Some installation instructions written 
 * Added sdf2table and gen-table scripts, and the SDF2 grammar as used by
   pgen
 * Added default value for ASC in configure.in

2000/03/28 (mdejonge@cwi.nl)
 * No longer distribute DEVELOPERS-INFO and reconf

2000/03/03 (mdejonge@cwi.nl)
 * pgen is now a self-cotained component build and configured
   using automake/autoconf
