Proposal: Add `future::is_ready()` as a synonym for `future::ready()`
Background
Currently member functions providing a boolean state query on a UPC++ library object generally follow the naming pattern {is,has}_[state]
. Examples:
atomic_domain::is_active()
dist_object::is_active()
dist_object::has_value()
global_ptr::is_local()
global_ptr::is_null()
team::is_active()
gpu_device::is_active()
heap_allocator::is_active()
Problem
I believe the only current exceptions to this naming pattern are:
view::empty()
future::ready()
view::empty()
inherits its name from the standard container library (e.g. std::vector::empty()), so should be familiar to novice users and seems unlikely to cause confusion.
However, future::ready()
deviates from this naming pattern for no good reason I can ascertain, aside from historical inertia. Worse, the name ready
could be misread by naive users as a command (e.g. "please ready this future") potentially degrading intuition regarding correct use of UPC++ futures.
Proposal
Specify future::is_ready()
as an alias for future::ready()
, deprecate the old spelling and encourage use of the newer more explicit spelling in new code.
Comments (2)
-
reporter -
reporter - changed status to resolved
- Log in to comment
Proposed resolution in spec PR 102 and impl PR 492