CORSIKA 8 meeting

Europe/Berlin
Description

Connect to ZOOM meeting room:

https://zoom.us/j/786000579

 

Minutes are on the indico, see also attached here:

- Status of refactory-2020 and related MR was reviewed and discussed.

- Documentation and doxygen are an important aspect of future code quality. But it is not the aim of the code migration to refactroy-2020 to finish and fix all missing documentation. This is a too large burden and delays us too much to make quick progress. Thus, please add markers into files "\todo documentation needs update ..." into the files itself, if you encounter such situation. The "todo" markers are picked up by doxygen, and also many IDEs, and can then later be addressed.

- Remy will demonstrate his work on "conan" packet manager tomorrow. We are very eager to see that, it will work similar to "virtual-environments" in python and provide a clean, well defined environment for corsika. This will replace and fix all external dependencies.

- pythia will need some wrapper for conan,

- corsika-data and cxroot can be extended by ourselves directly.

- other code, like sibyll etc. will remain as integral code contribution in our "modules" directory

- phys_units will be moved to "corsika/framework/...."

- cnpy will eventually be replaced by our regular output format based on parquet, Extra note: we can also use parquet as format in corsika-data -- this will eliminate the explicit use of both libz and also libbz2. Of course compression inside parquet still needs to be handled.

- We also re-discussed boost: we need boost for several reasons, thus, there is no reason to avoid it where it is suited to help us. We will provide boost via conan. For histograms and iterators we will use boost-solution. Still: avoid using boost, if other options are similar in effort, or if the benefit is marginal. We don't want to fully immerse into all boost packages if there is no specific and clear benefit for us.

- Most relevant refactory-2020 MRs are close to final. Augusto's work on the StackInterface will conclude by not later than Tuesday. Following that, we can hopefully seamlessly put into place all other missing pieces of code very transparently.

- A small outlook: finish of code migration is imminent, conan packet manager will revolutionize our dependency management, output format is upcoming. Thus, we are on a trajectory to make a first real "pre-alpha" release in the very next future.

There are minutes attached to this event. Show them.
    • 14:30 14:45
      Welcome and Status 15m
      Speakers: Dr. Antonio Augusto Alves Junior (KIT), Ralf Ulrich (KIT)
    • 14:45 16:00
      Refactoring, status, open questions, discussion 1h 15m

      This is not indented to discuss the scope of the refactory-2020, or implementation/developments/code, or other fundamental aspects of the framework.

      This is to discuss the technical aspects of migrating code/build-system from "master" to "refactory-2020".

      Clarifications of coding styles and ways to most efficiently conclude refactory-2020 will be discussed.

Your browser is out of date!

Update your browser to view this website correctly. Update my browser now

×