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
Is your feature request related to a problem? Please describe.
Right now, the documentation for nix_value_force() says
...all functions that return a value call it for you. The only place you will see a NIX_TYPE_THUNK is in the arguments that are passed to a PrimOp function you supplied to nix_alloc_primop
According to @roberth, this is not actually true (anymore?). Having an exhaustive list of places that you can see a NIX_TYPE_THUNK documented in one place is useful, but likely to get out of sync.
Describe the solution you'd like
We should instead document on each function that returns a value whether that value can be a thunk.
Describe alternatives you've considered
Keep the list on nix_value_force() up to date (error-prone)
Generate that list programmatically (seems complicated)
Enforce the documentation as it currently stands is accurate. That every function that returns a value forces it (maybe not desireable?)
Is your feature request related to a problem? Please describe.
Right now, the documentation for
nix_value_force()
saysAccording to @roberth, this is not actually true (anymore?). Having an exhaustive list of places that you can see a
NIX_TYPE_THUNK
documented in one place is useful, but likely to get out of sync.Describe the solution you'd like
We should instead document on each function that returns a value whether that value can be a thunk.
Describe alternatives you've considered
nix_value_force()
up to date (error-prone)Priorities
Add 👍 to issues you find important.
The text was updated successfully, but these errors were encountered: