The Run Tests option in the System Administration menu is intended for use by program developers to rerun test jobs. It will repeat all the tasks in the current project (or a selected range of them) in the order they appear in the database. A new project is created as a sub-directory of the current project and the output files are written to that new project. The user can follow progress by watching the Job List window where initially all jobs are reported as 'ON_HOLD'. As each job is run it is updated as 'RUNNING' and 'FINISHED'.
The Run Tests option can be used to set up test runs interactively or with the following command line:
ccp4i -autotest [-project project_name] [-target target_directory] [-first first_job] [-last last_job] [-remote machine_name] [-exit]
Rerun tasks from the project_name project. project_name must be a project alias defined in the current directories.def.
Create a new project and project directory called target_directory. If no name is provided, a name will be constructed from the date (e.g. 10_oct_2000).
Rerun jobs starting at job number first_job.
Rerun jobs finishing at job number last_job.
Rerun the jobs on the remote machine machine_name but note that the controlling CCP4i process will run on the local machine.
Exit the controlling CCP4i process when the tasks are completed.
When rerunning a test job, CCP4i will:
So beware: the order of tasks in the test project may affect the results.
By default a directories.def file (defining the project names and directories) in the directory in which the CCP4i process is started, takes precedence over any other directories.def. This means that a tester should work in a master directory which contains a directories.def defining the test projects, and the test projects themselves preferably as sub-directories of that master directory. It is possible to run through several project directories with a controlling script.