Merge branch 'data_storage' of git+ssh://scm.gforge.inria.fr//gitroot/batsim/batsim into data_storage

These instance IDs can be used in the YAML scripts,
to generate filenames for example. This would allow
to bypass the maximum filename limitation.

These instance IDs are also logged in a CSV file,
located in base_output_dir/instances/instances_info.csv.
This would allow simple and generic merge of information
in the postprocessing.

Merge branch 'data_storage' of git+ssh://scm.gforge.inria.fr//gitroot/batsim/batsim into data_storage

Merge branch 'data_storage' of git+ssh://scm.gforge.inria.fr/gitroot/batsim/batsim into data_storage

One script has been added to generate compute-only jobs.
The two scripts should now sort the output JSON by key, to
simplify diff checkings.

Added a "redundant edge check" for workflow loading

all workoad jobs after the last workflow, if any as completed.

This is useful for doing workflow-related research so as to cut down
on simulation time once the workflows have completed.

Merge branch 'data_storage' of git+ssh://scm.gforge.inria.fr//gitroot/batsim/batsim into data_storage

  The workload_name field of the MainArguments::WorkflowDescription was no longer initialized when
parsing command-line argument '-W'!!

Buffers' sizes have been increased.
Furthermore, every snprintf output value is now checked,
to assert the buffer's end has not been reached.

Indeed, some information could have been lost, which
led to invalid output files

If the processes were killed too fast (when the other one ended/crashed),
their output information might have been lost. Increasing the timer from
1 second to 30 seconds should solve the problem in most cases

Submission times can now be automatically translated
via a command-line option