diff -r 689f24ee83e6 -r a0ab96ad65fe doc/migration.dox
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/doc/migration.dox Thu Oct 09 16:20:37 2008 +0100
@@ -0,0 +1,143 @@
+/* -*- mode: C++; indent-tabs-mode: nil; -*-
+ *
+ * This file is a part of LEMON, a generic C++ optimization library.
+ *
+ * Copyright (C) 2003-2008
+ * Egervary Jeno Kombinatorikus Optimalizalasi Kutatocsoport
+ * (Egervary Research Group on Combinatorial Optimization, EGRES).
+ *
+ * Permission to use, modify and distribute this software is granted
+ * provided that this copyright notice appears in all copies. For
+ * precise terms see the accompanying LICENSE file.
+ *
+ * This software is provided "AS IS" with no warranty of any kind,
+ * express or implied, and with no claim as to its suitability for any
+ * purpose.
+ *
+ */
+
+namespace lemon {
+/*!
+
+\page migration Migration from the 0.x Series
+
+This guide gives an in depth description on what has changed compared
+to the 0.x release series.
+
+Many of these changes adjusted automatically by the
+script/lemon-0.x-to-1.x.sh tool. Those requiring manual
+update are typeset boldface.
+
+\section migration-graph Graph Related Name Changes
+
+- \ref concepts::Digraph "Directed graphs" are called \c Digraph and
+ they have Arcs (instead of Edges), while
+ \ref concepts::Graph "undirected graphs" are called \c Graph
+ (instead of \c UGraph) and they have Edges (instead of
+ UEdges). These changes reflected thoroughly everywhere in
+ the library. Namely,
+ - \c Graph -> \c Digraph
+ - \c %ListGraph -> \c ListDigraph, \c %SmartGraph -> \c SmartDigraph etc.
+ - \c UGraph -> \c Graph
+ - \c ListUGraph -> \c ListGraph, \c SmartUGraph -> \c SmartGraph etc.
+ - \c Edge -> \c Arc, \c UEdge -> \c Edge
+ - \c EdgeMap -> \c ArcMap, \c UEdgeMap -> \c EdgeMap
+ - \c EdgeIt -> \c ArcIt, \c UEdgeIt -> \c EdgeIt
+ - Class names and function names containing the words \c graph,
+ \c ugraph, \e edge or \e arc should also be updated.
+- The two endpoints of an (\e undirected) \c Edge can be obtained by the
+ u() and v() member function of the graph
+ (instead of source() and target()). This change
+ must be done by hand.
+ \n Of course, you can still use source() and target()
+ for Arcs (directed edges).
+
+\warning
+The script/lemon-0.x-to-1.x.sh tool replaces all instances of
+the words \c graph, \c digraph, \c edge and \c arc, so it replaces them
+in strings, comments etc. as well as in all identifiers.
+
+\section migration-lgf LGF tools
+ - The \ref lgf-format "LGF file format" has changed,
+ \@nodeset has changed to \@nodes,
+ \@edgeset and \@uedgeset to \@arcs or
+ \@edges, which become completely equivalents. The
+ \@nodes, \@edges and \@uedges sections are
+ removed from the format, the content of them should be
+ the part of \@attributes section. The data fields in
+ the sections must follow a strict format, they must be either character
+ sequences without whitespaces or quoted strings.
+ - The LemonReader and LemonWriter core interfaces
+ are no longer available.
+ - The implementation of the general section readers and writers has changed
+ they are simple functors now. Beside the old
+ stream based section handling, currently line oriented section
+ reading and writing are also supported. In the
+ section readers the lines must be counted manually. The sections
+ should be read and written with the SectionWriter and SectionReader
+ classes.
+ - Instead of the item readers and writers, item converters should be
+ used. The converters are functors, which map the type to
+ std::string or std::string to the type. The converters for standard
+ containers hasn't yet been implemented in the new LEMON. The converters
+ can return strings in any format, because if it is necessary, the LGF
+ writer and reader will quote and unquote the given value.
+ - The DigraphReader and DigraphWriter can used similarly to the
+ 0.x series, however the read or write prefix of
+ the member functions are removed.
+ - The new LEMON supports the function like interface, the \c
+ digraphReader and \c digraphWriter functions are more convenient than
+ using the classes directly.
+
+\section migration-search BFS, DFS and Dijkstra
+- Using the function interface of BFS, DFS and %Dijkstra both source and
+ target nodes can be given as parameters of the run() function
+ (instead of \c bfs(), \c dfs() or \c dijkstra() itself).
+- \ref named-templ-param "Named class template parameters" of \c Bfs,
+ \c Dfs, \c Dijkstra, \c BfsVisit, \c DfsVisit are renamed to start
+ with "Set" instead of "Def". Namely,
+ - \c DefPredMap -> \c SetPredMap
+ - \c DefDistMap -> \c SetDistMap
+ - \c DefReachedMap -> \c SetReachedMap
+ - \c DefProcessedMap -> \c SetProcessedMap
+ - \c DefHeap -> \c SetHeap
+ - \c DefStandardHeap -> \c SetStandardHeap
+ - \c DefOperationTraits -> \c SetOperationTraits
+ - \c DefProcessedMapToBeDefaultMap -> \c SetStandardProcessedMap
+
+\section migration-error Exceptions and Debug tools
+
+The class hierarchy of exceptions has largely been simplified. Now,
+only the i/o related tools may throw exceptions. All other exceptions
+have been replaced with either the \c LEMON_ASSERT or the \c LEMON_DEBUG
+macros.
+
+On the other hand, the parameter order of constructors of the
+exceptions has been changed. See \ref IoError and \ref FormatError for
+more details.
+
+\section migration-other Others
+- The contents of graph_utils.h are moved to core.h
+ and maps.h. core.h is included by all graph types,
+ therefore it usually do not have to be included directly.
+- path_utils.h is merged to \c path.h.
+- The semantic of the assignment operations and copy constructors of maps
+ are still under discussion. So, you must copy them by hand (i.e. copy
+ each entry one-by-one)
+- The parameters of the graph copying tools (i.e. \c GraphCopy,
+ \c DigraphCopy) have to be given in the from-to order.
+- \c copyDigraph() and \c copyGraph() are renamed to \c digraphCopy()
+ and \c graphCopy(), respectively.
+- The interface of \ref DynArcLookUp has changed. It is now the same as
+ of \ref ArcLookUp and \ref AllArcLookUp
+- Some map types should also been renamed. Namely,
+ - \c IntegerMap -> \c RangeMap
+ - \c StdMap -> \c SparseMap
+ - \c FunctorMap -> \c FunctorToMap
+ - \c MapFunctor -> \c MapToFunctor
+ - \c ForkWriteMap -> \c ForkMap
+ - \c StoreBoolMap -> \c LoggerBoolMap
+- \c dim2::BoundingBox -> \c dim2::Box
+
+*/
+}