Hi, I'm trying to study how the engine internals work and stumped about what I am (not) seeing in the graphviz output. I am running pants test in example-python with

--engine-visualize-to

and rendered rule_graph.test.dot with graphviz. I see the root query and the goal_rule for test (screenshot below), but I can't seem to find a node anywhere in the graph referring to the

setup_pytest_for_target

rule that runs at some point to do the pytest-specific work (I can log or debug there and it def runs). Should I expect to see a node for it here? I am also puzzled as to what the

Reentry

nodes mean.

adjustments to the graph visualization always welcome! i don’t think anyone loves the current display mechanism.

maybe promising ...

ah, neat. yea, that’s a much better visualization for the runtime graph. although it can be tricky since rather than a tree, it is actually a graph (a node might have multiple parents)

yes, so that is a node with multiple clients, right? This should be able to accommodate that. The same node with appear several times.

just know that that can lead to an exponential blowup in the size of the generated data

yes, that would be it.

There was a way to dump as json all the rules and their Gets and which rules got them. Maybe related to the help goal? I'll try to dig around for it sometime this week

slightly more progress, now with lazy loading so full graph can be seen. playing with showing the result type - at least easy to read the function body and see which functions are being called to satisfy the Gets

dumping all the rules and such is just

./pants help-all

, I thought it was something more complicated

thank you, I will need to use that too. My next step is trying to understand the relationship between the rule graph and the runtime graph.