Refactor line-specific persistence control flow
- Feb 11, 2021
-
-
Alex Pearce authored
-
Alex Pearce authored
As required for the serialisation.
6957c75d
-
- Feb 10, 2021
-
-
Alex Pearce authoredb571c4e6
-
Alex Pearce authored3d06d9c1
-
Alex Pearce authored
Those under Hlt/Moore/ and doc/ we can do without fallout. Those under Hlt/Hlt2Conf are trickier because lots of folks have active development branches off of master, so a big update here will cause a lot of conflicts. Removing the remaining usages is tracked under #239.
7b0ce27d -
Alex Pearce authored270dcdde
-
Alex Pearce authorede78ed29e
-
Alex Pearce authored
Although the vision for HLT1 is less clear than for HLT2, it is clear that HLT2 persistence is currently considerably more specialised than that for HLT1 (which is largely SelReports-based). Because of this, the HLT2-specific logic is moved to a subclass of DecisionLine (the renamed HltLine class).
8cd32e5d -
Alex Pearce authored
A line runs algorithms to determine the control flow, which may appear in the configuration as single data handles or as algorithms. But the objects to be persisted must be specific outputs of algorithms; either a single output, some, or all.
2948deda -
Alex Pearce authored
This is largely a change in semantics: rather than Moore figuring out which locations should be persisted given the algorithms used in making the line decision, it is the line itself which deduces this.
0e3829fe
-