23 Nov
2023
23 Nov
'23
11:31 a.m.
On Wed Nov 22 23:30:15 2023 +0000, Francisco Casas wrote:
I see, that is a better optimization. The disadvantage is that it is at the expense of making the asserts in the following calls to copy_propagation_set_value() and copy_propagation_invalidate_variable() less strong, because the time of the last record may remain smaller that the requested time of the invalidation.
More in general it seems that you can avoid pushing a new record each time it matches the last one, be it statically written or not. It's probably not an essential optimization, but it's an easy one. I'd rather have it, but I can live without it, as you two prefer.