As my notebooks get more complex, I sometimes there are some unexpected dataflow events in my notebooks that cause bugs and unnecessary reevaluation.
While the notebook visualizer is useful in many situations, sometimes the bug is about dynamic behaviour. So for these kinds of bugs I made
Been using this for a while. Quite useful. I just upgraded it to include the dependency graph as an overlay. Also added anon variables
Some of the trickier problems span notebooks, so now the debugger visualizes cell updates across ALL notebooks in the main runtime (not just the open notebook). I’ve also added more tools to allow you to zoom right in when paused. I needed this to fix a glitching cycle across dependancies.
The following graph is of a reactivity cycle, you can see the regular pattern. The actual notebook had > 1000 cells so being able to identify the 10 involved in the self-triggering cycle was helpful, although it was still hard to figure out. Also note how zoomed in I am, you can see the aliasing of the animation frames. There are 18 cycles in 900ms, thats 50ms per glitch and you can see detail within those 50ms periods, so this is now capable of really zooming in to sub second timing issues. You can also get a sense of how long an async cell update takes by the distance between a blue and green bar.
