|Date||June 1st 2014|
|Participants||Peter Gorm Larsen, John Fitzgerald, Nick Battle, Shin Sahara, Luis Diogo Couto (minutes)|
Awaiting feedback from Dr. K via Shin on various topics.
The usual dribble of bug fixes since the last meeting. The following is from the VDMJ Release Notes, though the corresponding fixes have also been made to Overture:
2014-05-22 Account for is_ expressions in type checker\
2014-05-21 Set intersect can fail with named types\
2014-05-20 Correct type check of while statement condition\
2014-05-16 Enable type restrictions on UpdatableValues\
2014-05-08 Prevent inheritance of private constructors\
2014-05-08 Correction to listener handling in value assignments\
2014-05-08 Correct guards for static operations and variables\
2014-05-06 Correction for guards with uninitialized field expressions\
2014-05-01 Change SL state record types to reflect any state invariant
The remaining Overture components are not usually discussed under this heading. PGL has a new action (79-1) to discuss moving them over to release planning.
Version 2.0.8 released
Version 2.0.8 was released 19 May, with small bugfixes (grab it at GitHub). Unfortunately, PGL found a regression which means that the combinatorial testing isn’t working. There is also a case where the (beta) code generation fails due to a lack of folders. I expect to include fixes for either 2.1.0 or 2.0.10
These issues highlight the need for some level of UI testing. PGL and NB have a new action (79-2) for creating an initial list of UI tests to be carried out. Automation of these tests is a serious concern.
Also, the fact that no users complained of these bugs once again highlights the need to discuss our user base and community building efforts.
Version 2.1.0 plans
The plan for version 2.1.0 is to have two “banner” features:
The release is nearly ready; first release candidate exposed a small problem with the combinatorial testing; we’ll see nearly next week if it’s fixable in a short timeframe.
We’ve set up a (third) new buildserver at AU, taking over the overture.au.dk/build.overturetool.org address. The previous two build servers have been decommissioned, and the new one appears to be working very well. We will also still continue to use the Travis CI public infrastructure, as it gives fast feedback about any branch without having to configure it on the AU buildserver.
The status of the release process document 1 is steadily improving — JWC is reasonably certain that all of the commands are correct now. Among the things to come is the set of all the places that need to be updated once the release artefacts have been created.
It is still important that someone outside of AU has a handle on what needs to be done to perform a release, but we’re moving in that direction.
See Download stats
The wiki and website discussions were postponed for when we have more people present at the NM.
We once again discussed the topic of community building. While its importance is recognised by all, there is still no consensus on what the best approach is. There we some suggestions made (focusing on improving our tools and methods, targeting students, a wiki clean up) but in general we need a more cohesive strategy and set of goals for community building. The issue will be discussed further at the next workshop.
The SRA needs to be revised and there have been discussions on it. Old materials exit but their usefulness is questionable. Several questions have been raised such as PhD students, research topics for VDM, tying it closer to community building and the ultimate goals of the SRA.
This issue needs to be discussed further and the workshop is one possible venue for it (though there are concerns of whether we’ll have enough time to cover everything). Another possibility is to have a NM dedicated to it.
Updates made to Planned Publications.