The old laptop still runs our read-only FR10 identity check, but the rebuilt one rejects an argument before it contacts the controller. We have several SDK copies and no record of the one in use. I would like to identify that working environment before anyone starts changing controller settings to repair a Python call.
Have the running script report its interpreter and imported module path on both laptops. The folder somebody remembers installing may not be the one Python actually loads.
Old laptop loads a project-local copy. New one loads an installed package. Same script file, different callable signatures at the failing method. That makes the pre-connection error less mysterious.
And find out who owns that copy, because an unidentified working folder is a poor thing to leave the next maintainer when this laptop needs replacing too
Integrator supplied the project archive and can identify its SDK release. I had mistaken the folder's edited date for a version. The actual version is in their release note.
Integrator says unmodified release, and our maintainer is comparing the files before building a clean environment from it. We aren't patching the call to suit whichever import wins.
Keep the intended launch command with that environment too. A shortcut can still select a different interpreter after you have installed the correct package.
Clean environment matches the identified archive. The shortcut now reports the expected interpreter and module path, and the argument check succeeds. Documented read-only connection test is next.
Who keeps those setup notes current when the package changes, David? It would be a shame to replace several mysterious folders with one beautifully documented mystery next year
Check the returned identity and response format when you connect, not only the absence of an exception. The wrapper needs to receive the documented form it expects.
The approved read-only check passed from the normal shortcut, with the expected controller identity and response format. Our software maintainer owns the archived package and rebuild notes. Thanks, Omar, the import-path check found the real mismatch.