Joseph Guillaume

I should have clarified that it's not good practice to export the entire global environment to the worker, which is why parallel$export is meant to be used. However, if both...

No issues with the export to global environment disabled?? > Maybe we just comment it out and add an issue for testing of evalPars with parallelisation. This would then be...

What about if you remove the .export completely? Hopefully all those tests fail?

And actually, would you be up for adding your tests as a PR?

Just realised we also need to address #12, where hydromad options are not exported to workers, so if needed we can just comment out this line of code and then...

@mortisflux - Is it correct that at the moment R CMD check passes without changing this code? If so, I think we close the current PR and work on creating...

Small differences between R and C versions found... > * `x` not equal to `expected`. > Mean relative difference: 0.001979361 > * `x` not equal to `expected`. > Mean relative...

It turns out this is a known property of the Sacramento model. For Milestone 1.0, sacramento.sim should return invalid model

Would have to avoid exporting parallelisation options, so that each worker does not in turn become a master, leading to unexpected behaviour.

Comments should definitely be kept in the source code, and the TODO items can simply stay there for the moment. If there are any that look major, you can open...