Sebastiaan Huber
Sebastiaan Huber
Note @giovannipizzi that if the repository is not there and the database contains `TrajectoryData` (which I guess is the case for @lekah 's database) the migration is going to fail....
Now that the Django backend has been dropped, it would be good to take a few (big) production databases and run the query provided in the OP. We can then...
> The caching mechanism will also detect the remote folder, if it finds the required remote folder is cleaned This check may not be so trivial to actually implement. Even...
> First, check the remote_folder before calling caching. Sure, but as I said, what are you going to check for? Just checking whether it is empty is not sufficient >...
> Every time we call clean() we change the input remote_folder hash so the process is not cached. The same code can be included in other operations which definitely modified...
See #5189 for adding official API to disable a node for caching
Just wanted to say that any sort of code that is going to try and go _ahead_ in time of the call stack to see what sub processes will be...
> It is perfectly feasible for there to be many 1000s of files generated in a remote folder. If we now have to store all this path information for every...
Thanks for double-checking @unkcpz . Good to know that it isn't the case, because that would have been a bad bug. Before looking into backtracking, I think the more direct...
I am not sure I follow what you are trying to say @unkcpz . Even _if_ it was possible to change how the hash of the `PhCalculation` is computed, that...