vle-forge / packages Goto Github PK

When leaving the modeling plugin the data structures should be cleared.

When creating a new dynamic using decision plugin the dynamic and conditions are added to the vpz but not attached to the atomic model

due to the prided new flushbybag option of the File and Console output oov plugin
"vle -P vle.examples test" fail.

A test of the availability of the simulator name was not correctly used.
A question remain, in which case the simulator name is not available?

After successfully installing vle-1.1.0-beta1 on a new Ubuntu 12.04 virtual machine I get the following error when installing the extension packages :

record@record-VirtualBox:~/Documents/packagesvle-1.1.0-beta1$ vle -P vle.output/ configure
-- The CXX compiler identification is GNU
-- The C compiler identification is GNU
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working C compiler: /usr/bin/gcc
-- Check for working C compiler: /usr/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- checking for module 'vle-1.1'
-- found vle-1.1, version 1.1.0
CMake Error at /usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:91 (MESSAGE):
Could NOT find VLE (missing: VLE_LIBRARY_DIRS)
Call Stack (most recent call first):
/usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:252 (_FPHSA_FAILURE_MESSAGE)
cmake/FindVLE.cmake:142 (FIND_PACKAGE_HANDLE_STANDARD_ARGS)
CMakeLists.txt:49 (find_package)

-- Configuring incomplete, errors occurred!

Code should be modified for using Gtk::Builder

This dashboard should enable to configure for each class of entity but also for each entity,

If they are usable, without to be linked explicitly.

If they are observable (in the sens of vle)

If they are visible on the diagram.

A default policy will be furnished.

If a variable is not declared in the map conatining initialisation variables, an exception is thrown. A default initialization to 0 is better.

Variables initialisation should be simplified.
A map "variables" containing for each var a value is expected.
The variable could be initialised, using directly the InitEventList as the container.

The modeling plugin of the Decision extension does not generate an agent that does know where to find the data directory where the plan should be stored.

Providing a way to choose colors but also fonts.

instead of generating a map of initial values of variables.
Directly generating simple doubles parameters ports.
This should be done in to steps.

When compiling, on linux ubuntu 13 (Raring Ringtail), gcc 4.7.3, the stdlib and iostream includes are missing and the compilation fail.

Variables can be linked by input of information flow.

In the current version of the plugin the generated .cpp file is in the /src/ folder with no way to put it in a subfolder
Same goes for "fichierPlan" that is generated in the /data/ folder with no way of putting it in a subfolder

example where these feature could be used :
So far parameters of DifferenceEquation can only be double.
Adding a user part in the class constructor would allow to use other types for parameters (after declaring the object as a private one in the user defined part)

When during a single modeling session, the user goes back to the Forrester plugin, the previous model can still be visible on the background of the current one.

The vpz files unittest-long.vpz and unittest-short.vpz are links and not plain file.

To reduce the size of extension, we need to remove the TARGET_LINK_LIBRARIES of static library.

The extension have a lib directory with binaries. Remove these directory.

In the current version there is a removeActivity method available in KnowledgeBase (but no removeRule), that allow to remove individual activities from their name.

It could be handy to have a method that remove all activities and rules at once.

A use of this would be to load several times the same plan using different relative years. (in the current version the activities and rules names are conflicting because they should be unique)

The time step can not be changed, when the Forrester Model already exist.

providing a new "raw" output file in order to log state changes even a the event resolution and with many outputs at the same time.

The julian day option does not work, the corresponding colum is not generated

The file output use the default ostream precision, ie. a double equal "2.44815e+06". We need to upgrade the output precision.

CellDEVS have several memory leaks.

It would be nice to have an output filter based on temporal resolution.

In my case, a single LotkaVoltera model, where each output port is potentially used, with a RK4 resolution of 0.01 does cost x5 in time computation.

I fail to build vle.extension.differential-equation:

$ vle -P vle.extension.differential-equation clean rclean configure build
-- The CXX compiler identification is GNU 4.8.2
-- The C compiler identification is GNU 4.8.2
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.26") 
-- checking for module 'vle-1.3'
--   found vle-1.3, version 1.3.0
-- Found VLE: /media/Stock/usr/include/vle-1.3;/usr/include/libxml2;/usr/include/glibmm-2.4;/usr/lib/x86_64-linux-gnu/glibmm-2.4/include;/usr/include/sigc++-2.0;/usr/lib/x86_64-linux-gnu/sigc++-2.0/include;/usr/include/glib-2.0;/usr/lib/x86_64-linux-gnu/glib-2.0/include  
CMake Error at cmake/VleUtilsConfig.cmake:222 (string):
  string sub-command REGEX, mode REPLACE needs at least 6 arguments total to
  command.
Call Stack (most recent call first):
  cmake/VleUtilsConfig.cmake:340 (IntVleBuildDepends)
  CMakeLists.txt:58 (VleCheckDependencies)


-- Boost version: 1.55.0
-- Found the following Boost libraries:
--   unit_test_framework
--   date_time
-- Could NOT find Doxygen (missing:  DOXYGEN_EXECUTABLE) 
-- Configuring incomplete, errors occurred!
See also "/media/Stock/devel/vle/packages/vle.extension.differential-equation/buildvle/CMakeFiles/CMakeOutput.log".

The test for dsdevs fails (ubuntu 12.10, for vle and packages dev branch)

either Parameters or Variables can be proposed twice, and this even if the current flow does not depend to the parameter or the variable.

This is due to a wrong criteria.

Using the cloud representation for flow without input or output.

when renaming an output function, it breaks existing input/output ports link in the graph and remove output function of existing activities

A variable dialog box should be at the same level than a compartment. In the way than there is an equation to edit and also dependencies to materialize.

The plugin for configuring vle.output/storage could be improved. The parameters that can be set are number of rows and columns and the increase step associated to. It coud be improved in two ways :

• set default values of these parameters to more realistic values. They are set to 1 which leads to resize the ouptut matrix (eg. number of rows) each time ther is à new observation; which is costly.
• let the possibility to configure the header parameter ("top" means that the header is given on the first row of the matrix)

when saving the model, the relation between observation ports and output views are loset

Error in Multiple.hpp
the following line in Multiple must be public in class Multiple
typedef std::map < std::string, MultipleValues > MultipleValuesMap;
When it's private, glue do not compile

Or glue should not depend on private types of Multiple.

In order to work on an larger canvas, bur also providing a better panning system.

The extension differential-equation takes time to compile due to unit tests. I suggest we move tests to vle.examples.

In the signature of the ack function the 2 parameters (const std::string&/activityname/, const ved::Activity& /activity/) are commented and thus non-available for use in the function

using this terminology « material flow », « information flow », « auxiliary variable », « exogenous variable »

the integration scheme can not be changed when the model already exist.

when re-opening the plugin on a generated model a segfault occurs
(reproductibility not very clear)
after restarting gvle the 1st opening of the plugin works fine but no more after closing the atomic model box

The selection should be easier, and the select area larger

The initialization of variables using createVar(name, value) is erased due tto the fact that initialization relies on experimental conditions. This probem occurs when we initialize variable values into constructor.

when the user don't want any temporal constrain to be used for a given activity, the plugin generate one anyway with a default value not very adapted.
It would be better to completly remove the temporal part of the activity in the "fichierPlan" when the user don't define any

When a Rule is renamed, the plan does keep a rule with the new name, but also a rule with the old name.

The variables init values should be read from the experimental condition, but not from the TAGS

Forrester: providing a light weight documentation system. A way to at list annotate the items.

In order to analyses the result of a simulation, it would be convenient to provide a observation port or more that enable to identify which rule did launch what activity.