Correction in doc: skeleton/path.h has been moved to the 'skeletons' module.

Minor changes in doc.

Changes in doc.

One more thing to do...

- Clarified Path skeleton.
- setStart() changed to setStartNode()

- Changes in doc
- setGraph(...) finally removed from Dijkstra, Bfs and Dfs.

- Changes in doc
- Some obsolete features has been removed.

Move general graph compilation tests to 'src/test/graph_test.h'.

(none)

(none)

Skeleton for paths.

INSTALL and COPYING are not generated files.

Put a lot of missing header files to src/hugo/Makefile.am

New doxygen module "skeletons" for the skeletons.

- BFD/DFS/Dijkstra compile test is done with skeleton::GraphSkeleton graph
and skeleton::ReadMap.
- 'skeleton::' is explicitely written instead of 'using namespace ...'
in graph_test.cc
- Output messages of type "makeRep(3)..." in unionfind_test.cc have been
changed in order not to confuse compiler output parsers.

Nicer and more documented graph_wrapper.h file.
These are only the first steps for making this file more beautiful.

Formatting: breaking long lines.

- 'minlengpaths_test.cc' added.
- Path tests in 'minlengpaths_test.cc' are swithced out.

.

mincostflow_test is ok.

Tests for the existence of 'KeyType' and 'ValueType' in the graph maps.

Change 'Key' to 'KeyType' (possibly temporarily).

Changes in doc. (New module name for array/vector maps added.)

dimacs.h

(none)

--This line, and those below, will be ignored--

A hugo/sym_map_factory.h
M hugo/list_graph.h
A hugo/array_map_factory.h
A hugo/map_registry.h
M hugo/smart_graph.h
A hugo/map_defines.h
A hugo/extended_pair.h
M hugo/full_graph.h
A hugo/vector_map_factory.h

Changes in the doc.

- DFS class (bfs.h and bfs_test.cc) added
- Bugfixes in Dijkstra and Bfs

correction

.

graph_wrapper.h is ready for hugo 0.2

minlengthpaths_test.cc is already hugo++ comform and is compilable

ResGraphWrapper<Graph> is done, so does dimacs.h.

Merge back the whole branches/hugo++ to trunk.

bug fix, test...

Demo directory...

sg is moved sg is not...

.

.

stGraphWrapper modifications

Revome an unnecessary (but still important) comment.

Add lp_solver_wrapper.h to work doxygen.

Add empty docs in order to make the functions and classes visible in
doxygen.

An experimental LPSolverWrapper class which uses glpk. For a short
demo, max flow problems are solved with it. This demo does not
demonstrates, but the main aims of this class are row and column
generation capabilities, i.e. to be a core for easily
implementable branch-and-cut a column generetion algorithms.

lp_solver_wrapper stuff.

technical corrections

flowValue() is reset to deal with excess[t].

Legal stuff, version 0.2.

New Doxygen module for path/flow algs.

New Doxygen module for path/flow algs.

resetXxx() changed to setXxx().

- A summary of the implemented graph structures.
- Some words on the different (and still nonexisting) graph concepts.

- 'KruskalPairVec' is changed to 'KruskalMapInput'.
- Changes in KruskalMapVec. Still does not work.

Test icc compiler as well.

doc change - one more todo (saveState()).

doc change - one more todo.

- Trimmed in order to work with gcc-3.4
- The number of executions of the tests can be controlled by command arg.

Bugfix in GRAPH_TYPEDEF_FACTORY

bug fixing

(none)

a bug test for preflow with preflow_bug_8 dimacs file

(none)

jacint javitgatott.

Some comments.

Bugfix + working example

- bfs-bench added
- hypercube generators moved to bench-tools.h
- new benchmark script

A very flexible bfs function using named parameters and impicit map types.

It is really obsolete, but containes interesting stuffs.

GraphWrapper::id(const Node&), GraphWrapper::id(const Edge&) function,
'cause I need it.

correction of some bugs pointed by alpar

Several changes in Kruskal alg.
- Input object interface was changed to an STL compatible one.
- template parameters of class KruskalPairVec has been simplified.
- (the most of) the names meet the naming conventions.
- a lot of (but still not enough) documentation has been added.
- class KruskalMapVec has been commented out.

Marci (and the other developpers),
please write exhaustive test code to avoid this kind of bugs...

gcc-3.4 tries to compile the _unused_ members of
an instantiated template class

A remark added.

Check StaticGraphSkeleton, as well.

Skeletons have been simplified.
"Optional features" have been deleted.
Map skeletons have been renamed.

max_flow.h added to Makefile.am

Custom made INSTALL file (will be sometime).

Repeat tests more times.

More documented file.

- Better Makefile.am. (I hope.)
- Some more tests

max_flow.h (wich doesn't use STL
- max_flow.h (which doesn't use STL stack) moved to /src/hugo
- for_each_macros.h was removed from max_flow.h
- (blocking) flow augmenting stuffs was removed.

We do not yet need it.

Do not ignore INSTALL.

Doesn't compile 5 times for the five execution.

For the sake of icc.

Indenting.

MapBase added

Indenting

Ready to run the first test series.

Test EdgeSet/NodeSet as well.

Turn on '-Wall' by default.

trying if without stl stack we are faster

without stl stack we are faster

Oops: "static" and "const" does not like each other.

src/becnhmark gets in the distro.
I hope it works well.

Some tools of common usage was put to bench_tool.h

Put some "static"'s.

docfix

A new benchmark (hcube)
and other minor changes

Warning fixation.

Temporal change: public Edge constructor with given ID.

Unnecessary friend declaration removed.

Added new make target 'benchmark'.

(none)

(none)

(none)

*** empty log message ***

Benchmarks

(none)

Capitalized section title.

Example in the doc is corrected.