On feature flags... I'd suggest an xmlrpc/API for evaluating flags.
evaluating flags requires DB access for group membership of a user and
potentially other things like time limited trials, access to opt-in
features and so forth.
The exact API would need to take a user, possibly feature names, and
return a list of the rules whose scopes need to be locally resolved.
Whether this is json or not is immaterial to me - at least for this
patch, the feature evaluating code would load once.
It would be very interesting to do this though, because we could phase
this in per-user.
On feature flags... I'd suggest an xmlrpc/API for evaluating flags.
evaluating flags requires DB access for group membership of a user and
potentially other things like time limited trials, access to opt-in
features and so forth.
The exact API would need to take a user, possibly feature names, and
return a list of the rules whose scopes need to be locally resolved.
Whether this is json or not is immaterial to me - at least for this
patch, the feature evaluating code would load once.
It would be very interesting to do this though, because we could phase
this in per-user.
-Rob