Aim to support the use case of a PFA host system updating engine state,
for instance with reference data used during the execution lifecycle. The
current language allowing such functions may be called seemed sufficient
given that I was attempting to limit the scope of changes, but the limitation
around calling functions during the action lifecycle would severely impact
the usefulness of PFA for the business cases I am exploring.
As justification for relaxing this restriction, I note that interacting with
the standard-provided state already has concurrency restrictions
defined in section 3.6 which are sufficient for protecting the integrity of
ongoing calculations, and that anyone who sought to make use of this
functionality would be well aware of the consequences of taking such
actions and presumably finds them acceptable.
-- meta note on the pull request - my github diff view is showing a number of changes I didn't make, but comparing the sources directly across branches show that the changes don't actually exist, so I'm just going to hope that it's a weird web UI bug and git is doing the right thing like it always does.
Aim to support the use case of a PFA host system updating engine state, for instance with reference data used during the execution lifecycle. The current language allowing such functions may be called seemed sufficient given that I was attempting to limit the scope of changes, but the limitation around calling functions during the action lifecycle would severely impact the usefulness of PFA for the business cases I am exploring.
As justification for relaxing this restriction, I note that interacting with the standard-provided state already has concurrency restrictions defined in section 3.6 which are sufficient for protecting the integrity of ongoing calculations, and that anyone who sought to make use of this functionality would be well aware of the consequences of taking such actions and presumably finds them acceptable.
-- meta note on the pull request - my github diff view is showing a number of changes I didn't make, but comparing the sources directly across branches show that the changes don't actually exist, so I'm just going to hope that it's a weird web UI bug and git is doing the right thing like it always does.