Hi all, We've been thinking it'd possibly be a good idea to update about current progress once a week or two. This should usually be pretty short, but might help some follow ARIA on a high level without having to delve into the JIRA issues etc.
On this week's news: - The main story is the modelling overhaul that's been taking place for some time now - up until recently, we've been using models mostly based on Cloudify's, merged with some TOSCA on the side. Much work's been done as of late to change this and instead have the whole of ARIA use proper TOSCA modelling. This work's not yet completed - redoing the models means changes that propagate to other parts of the system which use these models, which is, well, everything - but a major piece of it has been merged this week, and it seems like the other pieces are also starting to fall into place. - ARIA now has some actual service template examples! Those are the ones that can be found in the TOSCA spec, with some minor adjustments to fix errors in the examples themselves. This also helped us fix some bugs in the parser - which has also been celebrating its first tests being introduced this week. Whoo! - The mechanism for serializing an operation context when using the process executor has been made more generic. ARIA runs operations in subprocesses, and so the operation context - the object through which you use ARIA's operation API - has to find its way into these subprocesses. Before, the serialization mechanism assumed the API interacts specifically with sqlite storage, which is what ARIA uses by default when run as a standalone - Now, however, it should no longer pose a problem for any third parties which might want to use ARIA but with a different type of SQL database. - We've experienced some freezing and other issues on Windows, so we've had the pleasure of much Windows debugging this week. Fun times! See you next week! Unless Windows decides otherwise. Ran
