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
Rob, Matt and I had a talk about how to update the access level to avoid any duplicate work, and to make things simple coding wise. The basic idea is to track the location in the cycle in Core, and then during USetVector, determine if it's time to call the user's access function. This would put all of the logic into one location, which is good for maintainability.
We see 4 access levels
0 - no access
1 - access only at the end of the cycle
2 - access at the end of each iteration
(will happen in practice during the
first F-relax of the next iteration)
3 - access everywhere an F-relax is done
for diagnostic purposes, mainly
Core already tracks where we are in the cycle, to some degree, but it probably requires a few more flags. In the original version of this issue (in roundup), it was mentioned that a couple of other tickets should be finished first. One of them is already finished and the remaining part of the second ticket is in issue4 which doesn't look relevant to this.
The text was updated successfully, but these errors were encountered:
Rob, Matt and I had a talk about how to update the access level to avoid any duplicate work, and to make things simple coding wise. The basic idea is to track the location in the cycle in Core, and then during USetVector, determine if it's time to call the user's access function. This would put all of the logic into one location, which is good for maintainability.
We see 4 access levels
0 - no access
1 - access only at the end of the cycle
2 - access at the end of each iteration
(will happen in practice during the
first F-relax of the next iteration)
3 - access everywhere an F-relax is done
for diagnostic purposes, mainly
Core already tracks where we are in the cycle, to some degree, but it probably requires a few more flags. In the original version of this issue (in roundup), it was mentioned that a couple of other tickets should be finished first. One of them is already finished and the remaining part of the second ticket is in issue4 which doesn't look relevant to this.
The text was updated successfully, but these errors were encountered: