You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the configure methods for Trigger*Maker accept raw nlohmann::json objects containing their configurations, and anything using these must create the correct structure to configure the algorithm. It would be better if these algorithms could define a schema for their configuration using moo, similar to DUNE DAQModules. This needs to be done such that no dependency on daq-cmake is introduced, so triggeralgs can still be built and used outside of a daq software environment.
The text was updated successfully, but these errors were encountered:
To avoid dependency on daq-cmake I think here is one case where committing the generated files to a repo is okay. Because moo codegen is idempotent, the generated header files will only change when the developer actually modifies their schema (no repo bloat for inconsequential changes). This will not only erase daq-cmake as a build-time dependency but moo itself will not be needed to build the project.
Currently the
configure
methods forTrigger*Maker
accept rawnlohmann::json
objects containing their configurations, and anything using these must create the correct structure to configure the algorithm. It would be better if these algorithms could define a schema for their configuration using moo, similar to DUNE DAQModules. This needs to be done such that no dependency on daq-cmake is introduced, so triggeralgs can still be built and used outside of a daq software environment.The text was updated successfully, but these errors were encountered: