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
Most likely the function may no longer return false (unless we have poor coverage of it), which is probably a result of switching codecov environment from R 3.6 to R 4.0. We might want to revisit use of anyNamed function.
The text was updated successfully, but these errors were encountered:
Unfortunately, I don't have the time to implement and test a fix but I did have the time to look through the sources and was curious. Maybe this helps.
The function can still return false, I guess, just not when it's called on a non-empty list. But that's currently all that it's used for. Thing is that with reference counting turned on, MAYBE_NAMED checks the reference count and SET_VECTOR_ELT bumps up the reference count of the element it sticks into the list so that it's always at least one. With reference counting turned off, the named attribute doesn't get set. So you'd need to branch for the inner call only based on reference counting yes/no. Or rip out the test one way or another.
According to codecov the follow line is no longer visited:
data.table/src/assign.c
Line 665 in dd7609e
Most likely the function may no longer return false (unless we have poor coverage of it), which is probably a result of switching codecov environment from R 3.6 to R 4.0. We might want to revisit use of
anyNamed
function.The text was updated successfully, but these errors were encountered: