This content has been marked as final. Show 2 replies
just realized, there might be a wiring upate coming up. well. only a name change in the context. see other topic.
that still needs documentation. it would be good if someone could volunteer to read the test suites and the available documentation and creates high quality java docs. with a overall description of the interface in package.html.
Guillaume (the guy who is working on the persistence document) will take that. He will focus on the wiring framework unit test, javadoc and technical diagrams. We, at Bull, definetly need to have a strong knowledge of this module.
Guillaume is a young developer but he is really skilled so i'm confident he will perform a good job. He will be supported by the rest of the team on that.
next on my list will be the pvm and implementing the control flow part of all kinds of node types.
like e.g. sequence that requires child node execution. those tests will serve as examples on how the pvm can be leveraged to implement those node types in the respective languages.
if people are in doubt on how to implement a certain node, we'll do it first on pvm level. only then on process language level. is that ok ?
I agree as well. This is one of my main concerns, specially since the PVM basic execution logic has change. One of the things I proposed to Marc and Charles was to take some of the basic behaviours required in XPDL and BPEL languages implementations and add them as unit test into the PVM (for sure using ProcessBuilder pattern rather than the XML parser).
That way we can improve the PVM by testing basic but concrete execution logic matching with BPEL and XPDL requirements.
This should help us to detect potential bugs at the PVM level before be go to the languages implementations.