@zenflowapp turned me on to https://github.com/jpmonettas/flow-storm-debugger#readme. My preference has been, when stuck, to:
• add the obvious print statements;
• run;
• scan the output;
• re-run;
• repeat.
In rare cases I get stumped, and rinse-repeat can go on for a while.
And more than a little work goes into minimizing the output selectively, to make the scan step manageable.
The Flowstorm solution says:
• record everything;
• create an interactive browser for everything.
Is there an ETA on Flowstorm/CLJD?
If not, it would be easy to tell f/mx to record everything MX and f/mx (setState, dispose, etc). It could start as memory data structure for search with CLJD, then grow a UI.
