I am trying to automate a workflow using ‘logic-MCP’ and github copilot.
Most of the times, the call to add_analyzer hangs forever:
I can see the capture was started, the analyzer was added, but the call never returns until I manually stop the analyzer in logic2 UI.
After I click “stop” in the UI, the command add_analyzer returns:
{"analyzerId":10230}
{
"analyzerId": 10230
}
This makes it impossible to automate.
Hi @rhobison,
Sorry for the trouble with this.
The quick fix here is to stop the capture using the MCP interface before adding the analyzer.
We should have tested this case ourselves earlier. The design intent was that analyzers are only added through the API after the capture has stopped recording. (This makes sense for our automation API, which the MCP interface is based on. For interactive use of the software, it using analyzers while a recording is in progress is useful)
Sorry for the trouble! It should be a quick fix, just let your agent know that it needs to stop the capture before adding analyzers (or exporting data, or saving a file). In the short term we’ll need to test this and update the add_analyzer method to return a useful error message if the capture is still running.
Just to double check, when you say you click “stop” in the UI, you’re talking about the recording stop button right? If you’re talking about something else, let me know.
Thanks!
Mark
Hi!
Sorry, I missed your message!
Yeah, your recommendation is exactly what I am doing: asking the agent to add analyzers only after stopping the capture.
But it would be nice, indeed, to return an error with some informative message to the agent, so that it could recover from the error.
Yeah, when I said “clicking stop” I was talking about the button to stop the capture in the UI.
Thank you for your response!
Hi @rhobison,
This has been fixed in the latest release, Logic 2.4.48! Logic 2.4.48
It will be in the headless version of the software in its next release.