Ask HN: Do you still trace new codebases manually?
When you drop into a large, unfamiliar repository, what is your process for understanding what components play a role and how?
Do you rely on specific tools, mental mapping techniques, or do you just start dropping breakpoints and print statements? I'd love to hear about your workflow for building a mental model of a new system.
I usually ask an LLM to map out a feature and give me a tour, but they are very bids eye view and I want to know things down to the detail, so I find the entry point and follow calls through the IDE, jump between definitions, search for usages, etc. I try to build the execution flow in my head but end up a bit overwhelmed if there are too many layers. Also asking LLMs is a bit cumbersome because I have to reference what I’m talking about, they give me a 10 paragraph answer and by paragraph 2 I already have a follow up question, so it’s a bit messy.
Are there any tools for this?