Closed nevrome closed 1 year ago
I do like the opt-in idea.
Regarding the renaming, do you mean that we then should not have an update
command anymore? This might alienate some users, might it not?
We should probably discuss the renaming in #164. But yes, I think we should remove the update
command. The name promises something the command can not keep. I don't think this will be a problem for many users. It's pretty obscure. And if we refactor the interface as suggested, then a new name will, in my opinion, actually help to avoid confusion.
OK, I'm on board.
I think it would be very useful to be able to control the behavior of
trident update
more precisely. The idea is a bit similar to #248.At the moment
update
is a catch-all procedure where the user can exclude certain steps (with--noChecksumUpdate
or--ignoreGeno
). I think we should turn this around and make it opt-in. Each aspect of a package that should be updated should also be named explicitly in the call to trident update.That may translate to an interface like this:
This rewrite should be paired with a rename as proposed in #164.