Closed FlorentP42 closed 7 months ago
Something seems to have changed about this issue in Fledge 2.3.0: Now the reconfigure function on the filter is called but it is called in a loop infinitely after making a single config change: After that the dispatcher service soon goes unresponsive, until Fledge kills and restarts it (which does not stop the infinite call).
Tested on Fledge 2.3.0-47 nightly, now the observed behavior is as follow:
plugin_reconfigure()
was called from the dispatcher service is printed once.plugin_reconfigure()
is called on a filter which is not the one actually being used by the control pipeline.newConfig
parameter of plugin_reconfigure()
does not contain the new configuration but the one provided at pipeline creation.NB: Adding a Fledge restart after updating the filter configuration does not help either.
It looks like the merge of the fix lost a line of code. This has now be reinstated and has resolved the issue with the repeated logs. This will appear in the next night build. Apologies for this merge issue.
Tested on Fledge 2.3.0-73, this issue seems to be fixed completely in that nightly :)
Describe the bug When a control pipeline with a filter is created, changing the configuration of said filter does not work (the new configuration appears in the GUI, but is not actually used by the filter).
To Reproduce Steps to reproduce the behavior:
plugin_reconfigure()
method.plugin_reconfigure()
never appears after step 5 was executed.Expected behavior Updating the configuration of a filter part of a control pipeline is actually taken into account by said filter.
Additional context Note that configuring FilterA through scripted API calls right after it is created and before it is provided to the control pipeline allows to give it a custom configuration once. Yet it is still impossible to update that configuration after the pipeline creation.