Subversion Repositories hepmc

2006-02-17 Lynn Garren

	* HepMC/HeavyIon.h names have been changed to match both HepMC and

	wishes of heavy ion users

2006-01-12 Lynn Garren

	HepPDT 1.26 is available at https://savannah.cern.ch/projects/hepmc/

	* HepMC/HeavyIon.h implements information requested by CMS

	* HepMC/GenParticle.h has pointer (null by default) to HeavyIon

	* code providing interfaces to Fortran common blocks has been moved

	out of libHepMC and into libHepMCfio

	--------------------  HepMC-01-02-26  ---------------------------------

	--------------------  HepMC-01-02-21  ---------------------------------

2005-04-27 Matt Dobbs and Giorgos

	* src/IO_HERWIG.cxx implemented a bug fix suggested by borut,

	which keeps the remapping of daughters/motherrs from going "over

	the end" on herwig events (was noticed in the tauola events for

	rome DC).

2004-08-04  David Quarrie  <David.Quarrie@cern.ch>

	* cmt/requirements: Change dependencies on CERNLIB and CLHEP

        to AtlasCERNLIB and AtlasCLHEP to avoid name clashes

	--------------------  HepMC-01-02-19  ----(gcc 3.2 compatible)---------

2004-04-22  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/IO_HERWIG.cxx (HepMC): fixed bug reported by Hinchliffe in

	IO_HERWIG wherein (for the special case of min bias events only)

	number of entries in HEPEVT was being zero-ed by the IO_HERWIG

	class. Added extra protection in HEPEVT_Wrapper for this as well.

	--------------------  HepMC-01-02-18  ----(gcc 3.2 compatible)---------

2004-03-27  Matt Dobbs  <Matt.Dobbs@Cern.CH>

        Removed the #define statements that allow backwards compatibility

	with gcc 2.95, and specialized the 3.2 compliant iterators to be

	forward iterators such that they work properly with std

	algorithms.

	Thanks to Ulrik Egede for pointing this out.

	* HepMC/GenVertex.h, .cxx HepMC/GenEvent.h, .cxx: changed

	std::forward_iterator<GenParticle*,ptrdiff_t>

	to

	std::iterator<std::forward_iterator_tag,GenParticle*,ptrdiff_t>

	for standard

	compliance, and removed #define statements.

	* HepMC/IO_Ascii.h,.cxx: changed open_mode to openmode for standard

	compliance, and removed #define statements.

2003-12-17  David Rousseau  <droussea@lxplus072.cern.ch>

	* cmt/requirements: remove HepMC_libraries

2003-10-22  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/GenEvent.cxx and lots of other files + latex documentation.

	REmoved any mention of units--- GeV, MeV etc, since HepMC merely

	accepts whatever is put in. I do this because, though HEPEVT has

	used GeV/mm, ATLAS plans to use CLHEP units of MeV/mm.

2003-10-15  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/GenParticle.cxx (HepMC): added a default constructor for

	GenParticle as requested by Giorgos S.

	--------------------  HepMC-01-02-14  ----(gcc 3.2 compatible)---------

2003-04-21  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/IO_HERWIG.cxx:

	minor changes to handling of herwig event record

	--------------------  HepMC-01-02-11  ----(gcc 3.2 compatible)---------

2003-03-10  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* HepMC/IO_BaseClass.h:

	Fixed bug reported by Peter Loch, wherein the virtual destructor

	for IO_BaseClass was missing.

	--------------------  HepMC-01-02-10  ----(gcc 3.2 compatible)---------

2003-03-03  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* HepMC/GenVertex.h: added extra ifdef statement to allow

	compilation on gcc 3.2. (forward_iterator and open_mode problems).

	* HepMC/IO_Ascii.h: as above.

	* src/IO_Ascii.cxx (HepMC): as above.

2003-01-17  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/IO_PDG_ParticleDataTable.cxx (HepMC): if the PDG table does

	not exist, we abort, rather than just returning false.

	--------------------  HepMC-01-02-04  ----(5.0.0)------------

2002-11-04  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/GenEvent.cxx (HepMC): set_barcode(particle, barcode),

	set_barcode(vertex, barcode) Fixed a bug reported by Giorgos S.

	For this bug, a user suggests a barcode for a vertex that is

	already part of an event... the vertex is inserted in the vertex

	map with the new barcode, but HepMC forgets to erase the old

	entry... such that the vertex now appears twice in the map.  This

	is fixed, and the corresponding error for particles is also fixed.

2002-10-31  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* HepMC/IO_HERWIG.h: modified IO_HERWIG to include a switch

	(default m_no_gaps_in_barcodes=on) which removes null entries from

	Herwig HEPEVT and reshuffles the common block

2002-10-22  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* HepMC/HerwigWrapper6_4.h, HepMC/HerwigWrapper.h: new interface

	to HERWIG Version 6.4 is added. There is an example for using it

	at examples/example_MyHerwig.cxx. To use this example you will

	have to download HERWIG version 6.4 and modify the Makefile to

	link it.

	* HepMC/IO_HERWIG.h: New class for reading the Herwig version of

	the HEPEVT common block is added. READ THE COMMENTS IN THE .h file

	carefully before using it!

2002-07-29  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-02-03  ----(4.0.2)------------

	* src/GenVertex.cxx (HepMC): remove_particle()

	* HepMC/GenVertex.h:

	removed confusing comment about the use of GenVertex::remove_particle()

	in .h file, and clarified the comment in .cxx file.

	* src/GenEvent.cxx (HepMC): operator=

	bug discovered by <Malte.Muller@cern.ch>.

	Formerly, the GenEvent::operator= method relied on the particles

	attached to each vertex to always be in the same order. However,

	since the particles are stored in set<>, this is not always true

	-- making the results of the operator= method unreliable.

	This is fixed by mapping the vertices explicitly.

	* src/IO_PDG_ParticleDataTable.cxx (HepMC):

	added a line in read_entry that ignores empty lines (previously

	empty lines in the pdg data table generated a cerr message).

	This change was requested by Davida Costanzo.

2002-03-02  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-02-02  ----(3.0.1)------------

	* src/GenVertex.cxx (HepMC):

	GenVertex::edge_iterator::edge_iterator( .. )

	Extra error protection added in the constructor.

	Sometimes the edge_iterator fails when we have a vertex with no

	in_particles. This never happens in standalone mode, but when ran

	with the ATLAS framework, it does happen. New code identifies the

	scenario explicitly, and catches the error, which has to do with

	the past-the-end value of the edge_iterator.

	* test/test_iterators.cxx: new test program added for iterators,

	to test for the above problems.

2002-02-19  Matt Dobbs  <Matt.Dobbs@Cern.CH>

        CHANGES TO ALLOW HEPMC TO COMPILE WITH SOLARIS CC 5.2

	* cmt/requirements: added the line

	macro_append cppflags "" Solaris      " -D__SUNPRO_CC "

	to handle Solaris CC 5.2 features.

	* HepMC/ParticleDataTable.h:

	turns on MISSING_FORWARD_ITERATOR when __SUNPRO_CC is defined.

	This defines forward_iterator as iterator

	* HepMC/GenVertex.h:

	same as ParticleDataTable.h, but also turns on

	NEED_SOLARIS_FRIEND_FEATURE when __SUNPRO_CC is defined.

	* src/Polarization.cxx (HepMC):

	replaced abs() with theta = ( theta>0 ? theta : -theta );

	for compatibility with Solaris.

2002-02-15  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-02-01  ----(3.0.0)------------

	* HepMC/GenEvent.h: bug discovered by <srinir@bnl.gov> with help

	from Paolo and Hong.

	The GenEvent::particle_iterator, const_particle_iterator,

	              vertex_iterator, const_vertex_iterator

	had methods like:

	  bool operator !=(const particle_iterator& a) const

	  {  return !(**this == *a); }

	which doesn't work because you are not allowed to de-reference

	the end() iterator  [the above coding is a relic of the old

	              walking iterators which did not inherit from

	              forward_iterstor, and for which *end() was well

	              defined.]

	The correct usage is:

	{ return m_map_iterator == a.m_map_iterator; }

	Note: this does NOT affect the GenVertex::***_iterator's, since

	for those iterators *end() is well defined.

2002-01-23  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-02-00  --------------------

	                  ( Public Version 1.2 )

	User manual, webpage, and public version updated.

	The CLHEP names branch, called HepMC-00-00-96-CLHEP, has

	been moved to the head of the main trunk in cvs, and commited.

	Thus the main trunk is the only development branch.

2002-01-22  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-12  --------------------

	* src/GenEvent.cxx (HepMC): delete_all_vertices()

	modified the increment of the vertex iterator when looping over

	vertices for deletion.

	* src/GenVertex.cxx (HepMC): delete_adopted_particles()

	modified the increment of the particle iterator when looping over

	particles for deletion.

2002-01-21  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-07  --------------------

	* HepMC/GenEvent.cxx (delete_all_vertices):

	Possible bug reported here by ATLfast group. Added extra error

	checking, and changed the deleting of the vertices such that each

	one is explicitly erased from the map.

2002-01-18  Matt Dobbs  <Matt.Dobbs@Cern.CH>

  More changes for Windoes MS Visual C++ compatibility:

	* HepMC/PythiaWrapper6_2.h, HepMC/PythiaWrapper6_152.h:

	included new wrapper methods like call_pyinit, which hide the

	funny syntax necessary to call fortran routines from C++.

	* examples/example_MyPythiaWithEventSelection.cxx:

	* examples/example_MyPythiaOnlyToHepMC.cxx:

	* examples/example_MyPythia.cxx:

	* examples/Benchmarks_f77/example_PythiaToLCWrite.cxx:

	modified to use the call_*** pythia methods.

2002-01-15  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-06  --------------------

	* Makefile.standalone: updated to use cernlib 2001 and pythia6152

	* examples/example_UsingIterators.cxx:

	The input file was from the old non-barcodes version, and so

	didn't work. Replaced it with a proper input file.

These changes suggested by Witold Pokorski <Witold.Pokorski@Cern.Ch> from

LHCb to allow for Windows Visual C++ compliance. All of these changes are

fully backwards compatible(!).

	* HepMC/GenEvent.h, HepMC/GenVertex.h: change forward_iterator to

	simply iterator (which in the STL standard apparently implies

	forward_iterator)

	In Visual c++, a for (int i ... ) {} statement does not scope out

	the i variable ... so had to modify the code to ensure no

	variables were used twice in the same method:

	* src/IO_Ascii.cxx (HepMC):

	  in lines 155, 161, 175 "int i" is repeated

	  in lines 293, 314, 321 "int i" is repeated

          in lines 293, 314, 321 "p" is repeated

	* src/IO_HEPEVT.cxx (HepMC):

	  in lines 59,65,88 "i" is repeated

          in lines 118, 129 "p" is repeated

	* src/IO_PDG_ParticleDataTable.cxx (HepMC):

	  in lines 80, 103, 124 "id_i" is repeated

	* src/GenVertex.cxx (HepMC):

	  in lines: 64 and 72; 168 and 178; 195 and 201; 260 and 273;

	            322 and 331   variable "part" is repeated

	* have verified all STL calls for set, map, less, greater, cout,

	endl, list, cerr, ios, iostream, flush are properly prefixed with

	std::

	* HepMC/GenEvent.h: bug fix:

	added a std::greater<int> qualifier to the m_map_iterator of the

	vertex iterators.

	* HepMC/GenVertex.h: bug fix:

	made the GenVertex::particle_iterator a friend of GenVertex.

	* HepMC/GenEvent.h, HepMC/GenVertex.h:

	move all of the friend class declarations

	into statements separate from the class declaration, i.e.

	  class edge_iterator { ... };

	  friend class edge_iterator;

	instead of

	  friend class edge_iterator{ ... };

	Change all   std::ptrdiff_t   to  ptrdiff_t

	* HepMC/IO_Ascii.h: ios::openmode changed to ios::open_mode in the

	constructor, for compatibility with windows Visual C++.

	* Change #include <iostream.h> to #include <iostream>

	Note: this will spoil compatibility with HPUX 10.2 CC.

2001-11-29  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-05  --------------------

	* Fix typo in user manual "particle barcodes are positive"

2001-11  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-04  --------------------

	* update user manual to explain barcodes. NO changes to code.

2001-11-10  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-03  --------------------

	* cmt/requirements: jetset74 requirement removed, as requested by

	I.Hinchliffe.

2001-11-04  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-02  --------------------

	* HepMC/GenVertex.h: removed merge_vertex() method.

	* HepMC/IO_HEPEVT.h:

	removed the merge_vertices possibility, but added a

	build_production_vertex possibility to build_end_vertex.

	--------------------  HepMC-01-01-01  --------------------

	* src/HEPEVT_Wrapper.cxx (HepMC):

	new method check_hepevt_consistency() added.

	new method zero_everything() added.

2001-11-02  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-01-00  --------------------

	* src/IO_HEPEVT.cxx (HepMC):

	Modified such that the HEPEVT indices are used for the particle

	barcodes.

	* HepMC/IO_Ascii.h:

	Modifed to properly read/write the barcodes and the

	GenEvent new data members.

        Unique reference numbers called "barcodes" are added as data

	members to GenParticles and GenVertices.

	They are cross-referenced in GenEvent via a map<int,GenXXX>.

	The GenEvent::m_vertices container of vertices is then no longer

	necessary, so it is removed.

	The GenEvent::particle_iterator and GenEvent::vertex_iterator

	are updated to walk over the maps. The iterator change should be

	transparent to the user.

	Constant versions of both iterators are included.

	The old GenEvent::vertex_iterator was a typedef of the set:

	  typedef std::set<GenVertex*>::iterator vertex_iterator;

	The maps get filled via the:

		set_barcode( GenXXX* x, int suggested_barcode )

	methods, which in turn are called only by the vertex and particle:

	GenVertex::set_parent_event_()

	GenPartcle::set_production_vertex_()

	GenPartcle::set_end_vertex_(}

	* HepMC/GenEvent.h:

	New data members and access methods are added for consistency with

	the HepUP standard, see hep-ph/0109068

	double GenEvent::m_event_scale

	double GenEvent::m_alphaQCD

	double GenEvent::m_alphaQED

	* src/IO_HEPEVT.cxx (HepMC): Fixed some const warnings with the

	 GenEvent::vertex_iterator usage.

2001-10-11  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	--------------------  HepMC-01-00-01  --------------------

	* cmt/requirements: changed use CERNLIB v2001 External  to  use

	CERNLIB CERNLIB-01-* External   as per r.d.'s request

2001-10-10  Matt Dobbs  <Matt.Dobbs@Cern.CH>

        --------------------  HepMC-01-00-00  --------------------

	Identical to HepMC-00-03-08.  This tag name follows atlas

	naming conventions ... increment the major tag when code is NOT

        backwards compatible. Since Atlas is moving from tag

	HepMC-00-01-12 (old atlas names) to HepMC-00-03-08 (clhep names)

	this change is not backwards compatible.

2001-09-05  Matt Dobbs  <Matt.Dobbs@Cern.CH>

        ------- below this line is tag HepMC-00-03-08 ----------

	(clhep names version to be used in atlas. The changes have not

	been made in the HepMC-00-01-** branch)

	* src/IO_HEPEVT.h (HepMC):

	* src/IO_HEPEVT.cxx (HepMC): When inconsistent info between

	mother and daughters is found in the

	IO_HEPEVT::build_production_vertex method, IO_HEPEVT will no

	longer merge_vertex (the 2001-03-29 "fix") by default. In order to

	get this behaviour you have to explicitly ask for it with the

	IO_HEPEVT::set_merge_vertex_switch() method.

2001-09-05  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* cmt/requirements: added to the CLHEP names branch (its just a

	direct copy of what is in the trunk ATLAS names).

2001-07-17  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	(note: HepMC-00-01-24 is an accidental tag which should be ignored)

	----- below this line is  HepMC-00-01-23, HepMC-00-03-07 -----

	* src/GenVertex.cxx (HepMC):

	GenVertex::vertex_iterator::follow_edge_()  added extra error

	checking for the very special-rare case where a particle might

	point to the same vertex for both production and end.

	Meant to handle Ian H.'s bug of 02 Jul 2001

2001-06-28  Matt Dobbs  <Matt.Dobbs@Cern.CH>

        --------------------  HepMC-00-03-06  --------------------

	* examples/GNUmakefile.in:

	updated the makefile to link to  cernlib 2001 version of pythia.

	This involves a change -lpythia6136 to -lpythia6152 -lpythiad

	Note that for xample_MyPythiaWithEventSelection.cxx 1/100 or

	36/1000 events pass cuts (with 6.136 it was 3/100 or 37/1000).

	* HepMC/GenParticle.h: now has a parent_event() access method,

	which returns the parent_event of the particle's container

	vertex's event.

2001-06-27  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	---> Make the Vertex know which event it in in. <---

	* HepMC/GenVertex.h: added private data member m_event, public

	access method parent_event() and protected access method

	set_parent_event, which is only to be used by GenEvent. Thus

	GenEvent is made to be a friend of GenVertex.

	* HepMC/GenEvent.h, src/GenEvent.cxx (HepMC): added logic to

	GenEvent::remove_vertex and GenEvent::add_vertex. The vtx's

	m_event pointer is set to point back to the event.

	* HepMC/ParticleData.h:

	* HepMC/GenEvent.h:

	* HepMC/GenParticle.h:

	* HepMC/GenVertex.h: made method counter() protected. It was never

	intended as public, merely as a method for detecting memory

	leaks. If no-one complains (they shouldn't, it was commented

	"temporary for debugging") I'll remove it entirely later.

2001-03-29  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* src/IO_HEPEVT.cxx (HepMC):

	When m_trust_mothers_before_daughters=0, the vertex position is

	now filled. In ISAJET sometimes disjoint vertex structures exist,

	in this case a merger of vertices is done, which accounts for it.

	* HepMC/GenVertex.h:

	* src/GenVertex.cxx (HepMC):

	Added a            void merge_vertex_contents( GenVertex* v_in );

	method which takes all the contents of v_in and places it in

	the vertex. It does not delete v_in, nor does it remove

	v_in from an event.

	It will mainly be used by IO_HEPEVT, but it is forseen that users

	may want to use it when reducing the overall size of the event

	(deleting intermediate vertices) such as will be done by atlas.

2001-02-28  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* Makefile.standalone:

	using

	  HepMCdir = $(shell pwd)

	instead of

	  HepMCdir = /afs/cern.ch/user/m/mdobbs/HepMC/myCheckoutV1.01

	for better portability.

	* src/GenVertex.cxx (HepMC): added extra protection to the

	GenVertex::edge_iterator::edge_iterator constructor to give

	sensible result for the special case when a vertex has incoming

	particles, but none outgoing and the user requests a children

	iterator.  The old version would have

	m_set_iter points to m_particles_in.begin() and so the first

	dereference might be wrong.

	* HepMC/IO_HEPEVT.h: Two new switches are added to give more

	flexibility in handling the possible inconsistencies arrising

	from the bi-directional pointers in HEPEVT. The switches are:

		bool m_trust_mothers_before_daughters;

		bool m_print_inconsistency_errors;

	which are described by comments in the code. Also makes minor

	changes to src/IO_HEPEVT.cxx

	* HepMC/Typedef_Version0names.h:

	This header file added. It contains typedef statements which

	should allow the new version 1.XX class names to be used with code

	written with the old version 0.XX class names.

	If the user wants to take advantage of these typedefs he needs to

	define type variable HEPMC_SHORT_NAMES in his compiler options as

	follows: -DHEPMC_SHORT_NAMES

	or in his code using #define HEPMC_SHORT_NAMES

	Note that since the header file names also changed, this ALSO

	requires changing the include statements in the user code.

	* doc/latex_user_manual/physicist_visualization.eps

	fixed a mis-labelling in this figure... the labels were all

	shifted and intermingled... its now right. Unfortunately

	the mislabelled version appears in the CPC publication.

	also affects doc/latex_user_manual/HepMC_user_manual.ps

******************************************************

Everything below this line corresponds to Version 1.01

******************************************************

2001-01-11  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* examples/example_UsingIterators.cxx:

	Changed the class IsFinalState such that a particle must have no

	end vertex AND have status==1 to be considered stable.

	Does not effect package, only this specific example.

	* HepMC/IO_BaseClass.h:

	Bug reported by Lynn Garren

	"There appears to be a problem with const-ness in HepMC/IO_BaseClass.h.

	I have made the following changes in order to compile HepMC 1.0

	with g++.

	virtual const GenEvent*& operator<<( GenEvent*& );

		becomes

	virtual       GenEvent*& operator<<( GenEvent*& );

	virtual const ParticleDataTable*& operator<<( ParticleDataTable*& )

	        becomes

	virtual       ParticleDataTable*& operator<<( ParticleDataTable*& )

	In other words, I have made the const-ness match."

	...

	These changes are implemented by Matt in the atlas repository.

	* HepMC/HEPEVT_Wrapper.h:

	Bug reported by Andreas Dell'Acqua

		HEPEVT_Wrapper::last_child

		HEPEVT_Wrapper::last_parent

	There methods were returning 0 when exactly one parent exists. This

	caused the number_children and number_parent methods to return 0

	every time there was exactly 1 parent/child. This error caused

	major problems in the IO_HEPEVT class, with the end result that

	the HEPEVT graphs were interpretted incorrectly [all particles are

	correctly interpretted, but their relationships are incorrect for

	the case where a particle has exactly ONE mother: in this case the

	particles appeared as orphans.]

	This bug was

	introduced in version HepMC-00-00-94, so it has been present for

	some time. This fix is for HEPEVT_Wrapper::last_child():

	// Returns the Index of the LAST child in the HEPEVT record

	// for particle with Index index.

	// If there is only one child, the last child is forced to

	// be the same as the first child.

	// If there are no children for this particle, both the first_child

	// and the last_child with return 0.

	// Error checking is done to ensure the child is always

	// within range ( 0 <= parent <= nhep )

	* src/IO_HEPEVT.cxx (HepMC):

	only comments in IO_HEPEVT::build_particle have changed. No change

	to code.

******************************************************

Everything below this line corresponds to Version 1.0

******************************************************

2000-11-10  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	MAJOR NAME CHANGES AFFECT ALMOST ALL FILES:

	GeneratorEvent --> GenEvent

	Particle       --> GenParticle

	Vertex         --> GenVertex

	user manual and online documentaion are updated to reflect this.

	* HepMC/PythiaWrapper6_152.h:

	A new wrapper for Pythia 6.1 is created to replace the old Pythia

	5.7 wrapper. It works with the versions of Pythia in both

	CERNlib2000 and CERNlib2001 (versions 6.136 and 6.152 and probably

	any Pythia 6).

	A pointer file:

	* HepMC/PythiaWrapper.h:

	is created so the used doesn't have to worry about pythia version

	numbers. A test file is located at:

	* test/test_PythiaWrapper.cxx

	All examples are updated to use Pythia 6.

2000-11-09  Matt Dobbs  <Matt.Dobbs@Cern.CH>

	* HepMC/WeightContainer.h:

	Created this class to house the weights in the vertex and event

	classes. It is just an interface to std::vertex<double>.

	The vertex class is modifid to use this. Required changes to:

	* HepMC/Vertex.h:

	* src/Vertex.cxx:

	* src/IO_Ascii.cxx (HepMC):

******************************************************

Everything below this line corresponds to Version 0.94

******************************************************

2000-07-05  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/HEPEVT_Wrapper.h: Completely new wrapper for HEPEVT_Common

	is implemented. This wrapper treats the common as a series of

	bytes and interprets them "on the fly". The user can set the size

	(in bytes) of the integer and real numbers to be interpretted from

	HEPEVT, and also the total number of entries in HEPEVT. This

	allows the user to interface to two applications which use

	different HEPEVT definitions.

	Benchmarks: 1000 events pythia only 34 seconds.

	            1000 events Pythia+HepMC(using this new wrapper) 40s

        on my particular machine. (extra time is 6 seconds.) Compare this

	to the extra 5 seconds reported in the HepMC user manual for the

	old HEPEVT wrapper ... so the difference is small compared to the

	old wrapper considering the added functionality.

	(all examples have been tested and they Run!)

	* HepMC/HEPEVT_Common.h: Deleted, this common is now defined

	inside HEPEVT_Wrapper.h

	* src/HEPEVT_Wrapper.cxx (HepMC): This file is created since it is

	necessary to instantiate the static members of HEPEVT_Wrapper in a

	.cxx file.

	* Makefile.Standalone: modified to build src/HEPEVT_Wrapper.cxx into

	the HepMC library

	* HepMC/IO_HEPEVT.h: Comment modified only so as to treat

	HEPEVT_Wrapper consistently.

	* examples/example_MyPythiaWithEventSelection.cxx: Modified to use new

	HEPEVT_Wrapper

	* examples/example_MyPythia.cxx: Modified to use new

	HEPEVT_Wrapper

	* examples/example_MyPythiaOnlyToHepMC.cxx (main): Modified to use new

	HEPEVT_Wrapper

	* PACKAGE: added linkset HepMC -lib:HepMC

	(Modified GNUmakefile.in to use libHepMC.so as HepMC linkset)

2000-07-04  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* examples/GNUmakefile.in: $(top_srcdir) replaced by

	$(srcdir)

	* GNUmakefile.in (libHepMC.so_SRC): The location of src .cxx files

	is now explicitly specified with $(top_srcdir)/src/*.cxx

	This makefile now builds oth a shared library libHepMC.so and

	libHepMC.a (requested by Ian H.)

	* the initpydata.f routine is no longer compiled into the the

	HepMC library, but is still packaged with the other routines.

	The user needs to link to it explicitly. See the GNUmakefile.in

	in the examples directory for an example of how to do this.

	(requested by Ian H.)

	* ./src is replaced with $(top_srcdir)/src everywhere in the

	GNUmakefile.in

	Files changed this time:

	GNUmakefile.in, examples/GNUmakefile.in, Makefile.standalone,

	Changelog

2000-05-18  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/IO_BaseClass.h: There is now a fill_next_event method.

	- bool fill_next_event( GeneratorEvent* evt ) is a new abstract method

	  which fills the passed event, and also returns true if successful.

	- GeneratorEvent* read_next_event() is now make concrete. it

	  creates a new event, fills it using the above abstract method,

	  and returns the event

	This change is backwards compatible since the method

	read_next_event() has identical behavior.

	Identical changes are made for the methods:

	- bool fill_particle_data_table( ParticleDataTable* pdt )

	- ParticleDataTable* read_particle_data_table()

	The methods are updated [ i.e. GeneratorEvent* read_next_event()

	is changed to bool fill_next_event( GeneratorEvent* evt ), etc. ]

	in:

	- IO_HEPEVT.h

	- IO_Ascii.h

	- IO_PDG_ParticleDataTable.h

	- IO_HEPEVT.cxx

	- IO_Ascii.cxx

	- IO_PDG_ParticleDataTable.cxx

	(This change requested by Ian Hinchliffe and Marjorie Shapiro.)

2000-04-24  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/HEPEVT_Wrapper.h: set_momentum and set_position now

	properly cast input doubles as type HEPEVT_Precision,

	previously they were hardwired as cast to double, this

	meant writing to real*4 hepevt would fail.

******************************************************

Everything below this line corresponds to Version 0.91

******************************************************

2000-04-11  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* doc/latex_user_manual/HepMC_user_manual.tex: User manual is

	updated to replect changes since V0.9

	- HEPEVT and position in [mm]

2000-04-06  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* examples/example_MyPythia.cxx

	* examples/example_MyPythiaOnlyToHepMC.cxx

	* examples/example_MyPythiaWithEventSelection.cxx:

	By commenting out the initpydata call (which may be required on

	some systems to initialize the Pythia PYDATA block data as

	external) the MyPythia examples now run on HPUX (and still on

	Linux). Have not tested other platforms.

	* Makefile: HPUX with gcc requires the -lf library which contains

	the fortran intrinsic function.

2000-04-05  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	The changes below will have no effect on the user or package, they

	just serve to remove warnings when compiling.

	* HepMC/Vertex.h: 313:342:364: warning: unused parameter `enum

	::HepMC::IteratorRange dummy_range' (all dummy range parameters

	now appear as particles_end( IteratorRange /* dummy_range */ ) )

	* src/IO_Ascii.cxx (HepMC): 506: warning: ANSI C++ forbids

	variable-size array `c' (the array is now declared as char* c =

	new char[key_length +1]; and deleted before returning)

	* HepMC/IO_PDG_ParticleDataTable.h: 53: warning: control reaches

	end of non-void function read_next_event()

	(this is a dummy function anyway -- return 0)

	* HepMC/IO_HEPEVT.h:55: (as above)

	* src/Flow.cxx (HepMC): 26: warning: default argument given for

	parameter 1 of `void ::HepMC::Flow::print(class ostream & = cout)

	const' (done)

	* HepMC/IO_HEPEVT.h:57: warning: base class `class

	::HepMC::IO_BaseClass' should be explicitly initialized in the

	copy constructor (done)

	* HepMC/IO_Ascii.h: 98: (as above)

	* HepMC/IO_PDG_ParticleDataTable.h: 56: (as above)

	* HepMC/ParticleDataTable.h: 155: warning: suggest parentheses

	around assignment used as truth value (done)

	* src/IO_Ascii.cxx (HepMC): 37: warning: default argument given

	(removed default argument)

	* HepMC/Flow.h:21: warning: multiline `//' comment

	(added space after \ to remove compiler warning)

	* HepMC/Particle.h: removed =0 in the inline declaration

	inline void Particle::set_flow( int code_index, int code = 0 )

	(caused warning).

2000-04-04  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/HEPEVT_Wrapper.h:  HepMC/HEPEVT_Common.h:

	replace the classes HEPEVT_BaseClass.h HEPEVT_Double2000.h

	HEPEVT_Double4000.h HEPEVT_Real2000.h HEPEVT_Real4000.h

	Using different classes for each of the HEPEVT common block

	varieties proved very troublesome - users had trouble and 4

	similar classes were being maintained. The new version uses only

	static methods and cannot be instantiated (logical since it has

	no data members). The HEPEVT common block is contained in the

	HepMC/HEPEVT_Common.h header. The precision and number of entries

	is specified using #define commands, default is double precision,

	2000 entries. The user will see only one change - the IO_HEPEVT

	constructor now has no arguments, so he need not define the

	any HEPEVT_XXX class. He may need to set the precision and number

	of entries. See the header for instructions, and

	examples/example_MyPythia.cxx for an example.

	example_MyPythia.cxx, example_MyPythiaOnlyToHepMC.cxx,

	example_MyPythiaWithEventSelection.cxx are modified to take this

	into account.

	* HepMC/IO_HEPEVT.h: This class now extracts information from the

	fortran HEPEVT common using HEPEVT_Wrapper instead of having to

	pass it a class which inherits from the obsolete HEPEVT_BaseClass.

2000-04-03  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/PythiaWrapper5_720.h: Credit for writing the wrapper is

	to Silvia Resconi.

	Inclusion of HEPEVT wrapper header file is not necessary --- hence

	removed.

2000-03-31  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* HepMC/ParticleData.h: clifetime is now represented in [mm]

	instead of [cm] ... previously HepMC had internal inconsistencies

	in this --- it claimed to use [cm], but sometimes used [mm] (as

	when it read from HEPEVT, and othertimes used [cm] (as when it

	transformed a width into a lifetime. Now all is consitent and [mm]

	is used everywhere as it should be.

	* HepMC/Vertex.h: src/ParticleData.cxx:

	Changed comments which say clifetime is in [cm], to say

	clifetime is in [mm].

	* Makefile (CLHEPdir):

	Changed

	CLHEPdir      = /afs/cern.ch/sw/lhcxx/specific/Linux/CLHEP/pro

	to

	CLHEPdir      = /afs/cern.ch/sw/lhcxx/specific/@sys/CLHEP/dev

	to make it platform independent for CERN users.

	And some cosmetic changes to the Makefile.

	* HepMC/IO_BaseClass.h:

	Using the compiler options -ansi -pedantic cause this error:

	no match for `::HepMC::IO_Ascii & << ::HepMC::GeneratorEvent *&'

	Even though it recognizes:

	operator <<(::HepMC::IO_BaseClass &,

	            const ::HepMC::GeneratorEvent *&) <near match>

	as the "best <near match>".

	To fix this I make operator<<, operator>> members of the

	IO_BaseClass rather than friends, and I add versions where the

	argument of operator<< is not constant.

	This will not affect users in any way (unless they were using

	these compiler options, in which case HepMC didn't work anyway...)

2000-02-16  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

	* GeneratorEvent.h: set_signal_process_vertex() modified to ensure

	that any vertex specified as signal process is ALSO in the

	m_vertices set.

	* This change log documents changes since Version 9 of HepMC was

	"released" February 11, 2000

pre 2000-02-16  Matt Dobbs  <Matthew.Adam.Dobbs@Cern.CH>

Changes Before Version 0.9 (i.e. changes in going from V0.1 -->0.9)

	    HepMC improvements:

	        +backup V0.1  (done)

	        +save particle data as id in Particle  (done)

		+update IO strategies  (done)

		+test  (done)

		+backup  (done)

	     -change set<Vertex*> in event to include all vertices (done)

		+change iterators to reflect this (done)

		+update IO strategies (done)

		+test (done)

	     -fix all operator= to first destruct dependents. (done)

	     -access functions to iterate over all flows  (done)

	     -check for stl namespace  (done)

	      grep -n cout HepMC/*.h src/*.cxx | grep -v std::cout |

	        less

             -add Hepevt:write event  (done)

	     -implement lujet strategy (not done!)

	     -update documentation -- enumerating recent changes.  (done)

	     -user manual written