Closed CapPow closed 5 years ago
The issue has been isolated to save_output_images
in postProcessing.py
. Where multiple save_worker
workers are spawned (1 per save operation checked in UI). Each eventually calling handle_save_result()
upon completion. The class variable self.save_jobs_running
was an attempt to deal with this but apparently is not functioning. Any fixes should either use or replace (and remove) the self.save_jobs_running
class variable.
temp work around
In case anyone is working on another issue / project and needs a work around: selecting only 1 output option in the Image Locations
tab will avoid the issue.
fixed in commit: 4abd67fb2f406d45c428024661b8b6c90ccae8c2
While doing some speed tests, I attempted to drop many, many images into a watched folder at once. The first few images usually process fine, however as threads get busy there are occasional class variable issues which raise exceptions or assign incorrect class variables to images. Somehow, images are getting the go ahead to start processing out of queue order. This is likely related to
reset_working_variables()
being called too frequently or some issue withself.processing_image
.