We use cookies on this site to enhance your user experience. By using this site, you are giving your consent for us to set cookies.


2022-05-16

Attendees:

Agenda:


Notes:

Paul followed up with oems and will persist


Mkhail and Gunnar met and reviewed VSC.  Mikhail has given some thought to direction forward.

Discussed with legal as to what they can contribute.  Waiting for Magnus and legal to provide path forward.

Once legal done Mikhail will make initial pull request.  

YAML to IST to deployment (needs to be developed) to C++ (common api vsomeip) - deployment configuration

Erik - consider deployment information - might use separate generator

Mikhail - considered the idea of layers - for example layer of some/ip deployment or ... 

Mikhail - adding support to AST (abstract and syntax tree) for extending - need some type of link to deployment 

Mikhail - do we want to merge into one tree or use separate

Erik - this is very similar to the layering discussion in VSS.  For example, data like access control.  How to actually specify is a different topic.  If we plan on specifying your own.  Time will show how it works.

Stephen - So Magnus will not be in call today?

Stephen - How much do specified in AST?

Gunnar - as much as needed - Need to describe enough for the environment.  Whatever info you need for code gen not already in VSC language

Gunnar - everyone has understanding about split - IDL does not contain target info - separate file does meaning code generation

Gunnar - can take specific example - some/ip has numeric identifiers - on the generate side tooling takes care of id is the same.  If boundary between autosar and non-autosar need to specify id.

Gunnar - common uses and common contexts - deployment model contains things that are unique.

Gunnar - VSC definition - check out def - following split done in FRANCA

Erik - as VSC gets more adopted could theoretically start to go deeper into deployment as part of the spec - perhaps kept in layer file

Gunnar - as discussed in VSS if we can have agreement on certain layers they might become part of the spec

Gunnar - perhaps next step is to choose a target and make it real

Mikhail - whatever is needed to be adopted first from the desired architecture

Gunnar - start thinking of what the community needs

Mikhail - likes the discussion about layers and has thought about defining essential layers.  Most of the RPC technologies will need some kind of deployment layers.  We can start making theoretical layers.  For example for some/ip.  From a high level perspective similar to VSS.  Essentially wanted to AST.

Gunnar - depends on data items needed - Mikhail - we know what they are

Stephen - always going to be code gen - understand franca split - and upstream - but thinking downstream - 

Gunnar - Erik's contributions in chat? - 

Erik - walked through pull requests - #27 Introducing DTDL and Protobuf templates - one full week for comment before merge

Gunnar - comment specific code make it's way into code and need to separate

Erik - DTDL has version which VSC currently does not - could be a layer

Erik PR #31 - Restructuring of VSC documentation - keep readme very simple and move syntax documentation to a specific doc.

Erik PR #32 - Refactoring of seat service - align differences between VSC and VSS

Gunnar - re #32 - need to define service status

Paul - How stable do we think VSC language is?

Gunnar - good question has some thoughts

Erik - e.g. how to report screen supplies

Gunnar - who are the stakeholders here - without him

Paul - Do we need a review?

Gunnar - don't need to go through everythign but everyone shoudl go through and ask open issues.

Erik - #30 - VSC Properties #28 


Action items - go through pull requests

Action item - Mikhail and Magnus to perhaps push code








  • No labels