I think we can do something like AllBefore and AllAfter, explicit and gets the point across.

This is possible and will solve the current issue (I guess you mean "before all processors" by the AllBefore).
The only reason not to do it is if we think that in the future a more detailed ordering will be needed, which will in the end require to implement priority mechanism.

Recently I thought about pushing the time normalization to the sink stage.
This is due to the fact that normalization is required for user-facing outputs, and for internal usage we can do something similar to what suggested in #3726.
By moving the normalization to the sink stage, we don't need to worry about processors order or calling the same processor methods more than once (e.g. for signatures/derived events), which will cause "double normalization" with wrong values.

If #3726 is implemented, I don't think there is a need as well.
However, to create such a solution will probably use the processor functions as well...
Either way this will solve the problem if the time change will occur before all other processor functions

If #3726 is implemented, I don't think there is a need as well. However, to create such a solution will probably use the processor functions as well... Either way this will solve the problem if the time change will occur before all other processor functions

Processor functions can use the GetEpoch method of the timestamp (as described in #3726 ) and don't need to care about how the time is represented internally, so there is no need to make any time change before processor functions.