Robin Tweedie

The YAML spec does seem to suggest that the un-indented form is canonical (under 2.3) https://yaml.org/spec/1.2.2/#21-collections

There's quite a few other details I'm not sure whether to include: - to get a useful profile, this was done with `--jobs=1`: it does run faster with `--jobs=0` but...

I tidied up the graph, hope this will help: 

> Looking at your graph it looks like ClassDef.ancestors is taking quite a lot of time. Is there a lot of inheritance in your code ? Maybe this is something...

I was also going to ask: is there a way I could figure out the minimum set of pylint rules to disable in this context to keep my runs fast...

This was fixed in mypy 0.971 🎉 https://mypy-lang.blogspot.com/2022/07/mypy-0971-released.html https://github.com/python/mypy/pull/11143

I probably should have mentioned: the context in which I found this difficult was feeding `callgrind` results to `gprof2dot` and then visualizing with `graphviz` (which just shows the entire path)....

Yep, makes sense. Thanks for an awesome tool!

/remove-lifecycle stale

Still relevant.