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
{{ message }}
This repository has been archived by the owner on Sep 27, 2024. It is now read-only.
Long running acquisitions (like Post Mortem ones), get a timestamp that's very different from when they were actually completed, because these timestamps are propagated to most acquisition data PVs.
The mistake seems to be that the current time is only updated at the start of the acquisition, instead of right before copying the resulting data into the NDArray and running callbacks on it.
This is a workaround for [1]. The timestamp from drvBPM is wrong and
ends up being propagated from the ArrayData records to the final
position ones, which is not desired. This makes it so the final position
records' timestamps are from when they were processed, which is not an
issue for existing clients.
BPMSP (Single Pass) and BPMDsp (used for Monit rate position data)
weren't changed, since the former is no longer used, and timestamps are
relevant for clients of the latter.
[1] #79
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Long running acquisitions (like Post Mortem ones), get a timestamp that's very different from when they were actually completed, because these timestamps are propagated to most acquisition data PVs.
The mistake seems to be that the current time is only updated at the start of the acquisition, instead of right before copying the resulting data into the NDArray and running callbacks on it.
bpm-epics-ioc/BPMApp/src/drvBPM.cpp
Line 2584 in 78b843e
The text was updated successfully, but these errors were encountered: