I spent some time on this for a while today and think that the idea of "errors" is overloaded and so it hard to pin things down to a simple implementation. Your comments above highlight this, a user accessing something they aren't authorized to is very different than a NPE in a function that formats a string. Ultimately I couldn't really think of a way to add anything useful to what pathom 2 did.
Hello everyone! Sneak peak on the tracer for Pathom 3!! I was able to re-use the previous timeline, altough the format is a bit different, in Pathom 2 each row was an attribute, now they are tied to resolvers in the graph, getting some cool animations thanks to Cytoscape!
This is phenomenal. In your opinion, how much closer does this push Pathom 3 to “production-ready”? Is this a big step, or small?
feature-wise is close to complete, the tooling should be out this week, so for MVC I say its a small step. production-ready
to me also means the library is in use in production and really tested in real applications. currently I'm developing Pathom 3 without any such app to test it for good, so I will need have people trying it out. since the new planner algorithm is new and quite complex, as the complexity of users graphs grow, I expect early users to find bugs on it (when I was porting repl-tooling I found a few, but I guess there is more out there).
but once a few ups are up and running good, than I say its prod-ready, but also depends on people using it to get there 😉
👍 Gotcha.