The meeting lasted for three days. A lot of this discussion covered changes to LSE-209 and LSE-70, with time for integration testing. The changes which were discussed in the previous meeting were shown, and there was a further discussion on how to clarify the documents.
Integration testing: We initially set up VMs for testing, in case we weren’t able to test over the wireless connections. As it turned out, the wireless wasn’t configured to pass multicast packets, but Dave Mills brought a hub that we used (instead of the VMs). We did an initial connectivity test to make sure that the sending/receiving machines could exchange messages, and that worked. That was using the precompiled SAL software C++-based test programs. Tony Johnson was able to test the SAL Java interface successfully. The Python interface tests did not work. There were a couple of issues. One was that some of the method interfaces weren’t exposed in the Boost-Python binding, and another was that messages were being received, but the payload for the message wasn’t forwarded to Python. While Dave worked on correcting these issues, I looked into how tough a Swig binding would be to implement. This went well, and I was able to send log event and commands from Python on my system to the appropriate C++ receivers on Dave’s machine. Dave was able to fix the issues with the Boost-Python interact. At the end of the meeting he suggested that both binding could be supported.
Story
The meeting lasted for three days. A lot of this discussion covered changes to LSE-209 and LSE-70, with time for integration testing. The changes which were discussed in the previous meeting were shown, and there was a further discussion on how to clarify the documents.
Integration testing: We initially set up VMs for testing, in case we weren’t able to test over the wireless connections. As it turned out, the wireless wasn’t configured to pass multicast packets, but Dave Mills brought a hub that we used (instead of the VMs). We did an initial connectivity test to make sure that the sending/receiving machines could exchange messages, and that worked. That was using the precompiled SAL software C++-based test programs. Tony Johnson was able to test the SAL Java interface successfully. The Python interface tests did not work. There were a couple of issues. One was that some of the method interfaces weren’t exposed in the Boost-Python binding, and another was that messages were being received, but the payload for the message wasn’t forwarded to Python. While Dave worked on correcting these issues, I looked into how tough a Swig binding would be to implement. This went well, and I was able to send log event and commands from Python on my system to the appropriate C++ receivers on Dave’s machine. Dave was able to fix the issues with the Boost-Python interact. At the end of the meeting he suggested that both binding could be supported.