Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Minutes

...

Zoom link
meeting ID  : 993 7365 2562
 password: Vhal!21
local dial-in numbers: https://zoom.us/u/aqC5ms8E7


...

Tuesday December 7 – 5:00 PM CET
Participants

  • Kyrylo Skidanov (EPAM)
  • Stephen (Renesas)
  • Stefan Wysocki
  • Gunnar (COVESA)
  • Alexander Domin


Minutes

Kirill:  Procuring a development board.  Apparently graphics subsystem not starting on ARM64 (Google confirmed this...)  Being investigated.
  x86 emulation might be an option.  More disk space coming soon to prepare for builds.

Alexander: 

Further discussions about challenges of mapping VSS/different environments.
The Franca mapping file is very flexible - can map to Franca attribute (including subscription of changes, or to just a broadcast
or to an explicit getter function exposed on SOME/IP.  Similarly for writing, but it is defined independently => in theory different mappings for reading vs writing.

Stefan:  At the moment interested in Vehicle-Properties translation primarily, but it is interesting to see this other concept progress.
...Good to see that Microsoft's C++ implemented GraphQL is mature enough.  Apollo tried before which meant putting Java into the HAL layer, not ideal.

Compiling vspec2aaproperties code

Stefan: I tried also this but little progress (challenges with diferent computers, environment, disk space).







---


Tuesday November 30 – 5:00 PM CET
Participants

  • Kyrylo Skidanov (EPAM)
  • Johan (Melco)
  • Stephen (Renesas)
  • Manu (Bittium)
  • Philippe Robin (COVESA)
  • Gunnar (COVESA)
  • Alexander Domin (BMW)


Minutes

  • Updates on vspec2aaproperties code generator.  Generated code through compilation with minor changes.
    Need to add generation of getProperty... function for fetching dependent properties needed for complex calculation.  This function is currently not included by reference to the implementation. 
  • Manu: .... prefers generating these functions (variants depending on data type) according to if they are needed, which was done similarly with the type-conversion functions.
  • Discussion about setting up a repeatable build environment.  Apparently the standard way of building AOSP is "the whole system" (although rebuilding in the same directory structure is of course incremental and relatively fast).  How is AOSP builds normally set up for CI systems???  There ought to be a structured way to reuse a cache of already built artifacts.
  • Explanation from Alexander about the set of tooling (mentioned last week's minutes).  A demonstration of the technology planned for the next CVII Tech Stack meeting instead, since it is also more widely applicable beyond Android.


Tuesday November 23 – 5:00 PM CET

...