· We did 12 integration loops to finally get all the problems (compilation, conflicts, test errors) solved.
The code which makes the R0 release has been tagged with the CVS tag: ALMA-R0. If you are interested in the souces, you can retrieve them with:
cvs export -rALMA-R0 <list_of_subsystems>
A binary distribution is available on the web, in the IT&S pages under the link "Download" and it is called ALMA-R0. It is a tar-gzipped file.
The download is password protected. Send a mail to alma-sw-it@nrao.edu to know the authentication information.
· No major compilation problems; regarding the written code we discovered naming conflicts and the missing guards in the IDL files.
Most of those problems have been solved by now. The idl files who still do not have a guard are:
./ACS/LGPL/acsBUILD/config/Doxyfile.idl
./ACS/LGPL/CommonSoftware/acstestentities/idl/xmltest.idl
./ACS/LGPL/CommonSoftware/cdb/ws/idl/cdbDAL.idl
./ACS/LGPL/CommonSoftware/define/idl/xmlstore.idl
./ACS/LGPL/CommonSoftware/jcontexmpl/idl/BrightLamp.idl
./ACS/LGPL/CommonSoftware/jcontexmpl/idl/BrightUser.idl
./ACS/LGPL/CommonSoftware/jcontexmpl/idl/GuiComp.idl
./ACS/LGPL/CommonSoftware/jcontexmpl/idl/HelloLamp.idl
./ACS/LGPL/CommonSoftware/jcontexmpl/idl/XmlComponent.idl
./CONTROL/ArrayController/ArrayController/idl/ArrayController.idl
./CONTROL/ArrayController/ArrayMonitor/idl/ArrayMonitor.idl
./CONTROL/ArrayController/ManualArrayController/idl/ManualArrayController.idl
./CONTROL/MasterController/ControlStatus/idl/ControlStatus.idl
./CONTROL/MasterController/ControlSystem/idl/ControlSystem.idl
./PIPELINE/Science/idl/junk/ExampleSciencePipeline.idl
./PIPELINE/Science/idl/junk/NprotoSciencePipeline.idl
./PIPELINE/Science/idl/junk/ProtoSciencePipeline.idl
./PIPELINE/Science/idl/junk/SimpleSciencePipeline.idl
./PIPELINE/Science/idl/junk/XprotoSciencePipeline.idl
./PIPELINE/Science/idl/SciencePipelineTNG.idl
They do not give compilation or run-time problems, nevertheless hopefully they will be fixed for the next integration.
· The test code developed by each Subsystem (with the only exception of the SCHEDULING subsystem) is quite below our ideal
internal threshold of 50% on the entire written code
· Besides our regression tests (based on the modular tests supplied by the developers) we began to prepare integration tests.
We did not manage yet to have a repeatable test suite on the overall system. The different subsystems mostly act as a standalone piece
of software. In this respect, the work is still in progress and we aim to have proper integration and system tests ready for R1.
For this reason, we cannot speak of "validated release" for R0.
|
SUBSYSTEM
|
Comments
|
| ACS | Many modular tests have failed but no work has deliberately been put to improved them because people were busy with ther ACS 2.1 release. The tests have been improved for ACS 2.1. |
| CONTROL | Modular test structure still not existent! |
| CORR | Not all the modules have unit tests but the ones which can be run pass. |
| SCHEDULING | No problem. |
| OBSPREP | The tests are basically OK, but a better filter should be implemented to make them really independent from the machine and the user who runs them. For this reason, two tests still fail. |
| TELCAL | No problem. |
| DRUI | === |
| ARCHIVE | It does not have yet a standard structure for tests. |
| EXEC | The tests should be re-written separating the start and stop of the ACS services from the actual test code. The scripts to start and stop the services should go in the prologue and, resp., epilogue directives supplied by tat. |
| PIPELINE | No problem. |
