BCDA-APS / 3idc-bits / 27653721301

56%

Build Type

push

github

Committed by prjemian

Commit Message

flyscan: suppress post-stop primary events when motor is at rest

Scan 8 (2026-06-16 17:14-17:20) emitted 63 spurious 'primary'
stream events over 2m39s after monitor_loop detected the p_end
crossing and stopped the cam.  Every event reported
array_counter=109 and motor=5.041 -- the same frame and position,
ticking every 2s until cam_stopped_status finally fired.

Reading the code, _emit_pending_frames only emits if
'highest > last_captured'.  Empirically that condition was true
post-stop, meaning the IOC continued to publish num_captured CA
monitor events after the cam was told to stop -- either the cam
finishing its in-flight burst at the time of acquire=0, or the
CA dispatcher re-publishing.  Either way these were not 'new
in-scan frames'; they were noise.

Gate event emission on (NOT acquire_stopped) OR (motor still
moving).  When both are true (post-stop, motor at rest), update
the highest-seen bookkeeping for the loop's exit condition but
suppress bps.create / bps.read / bps.save.

Uses flymotor.motor_done_move (the standard EpicsMotor .DMOV
PV wrapper).  Defaults to True if the attribute is missing or
unreadable, so suppression is conservative on devices that don't
expose .DMOV.

Live test will validate; the nested-closure form of
_emit_pending_frames makes a focused unit test impractical
without a small refactor that I'm deferring.

Coverage Stats

0 of 16 new or added lines in 1 file covered. (0.0%)

2 existing lines in 1 file now uncovered.

1221 of 2189 relevant lines covered (55.78%)

0.56 hits per line