I'm not very familiar with

py-spy

, but is there a way to profile just steps relevant to a backend? I don't really have a technical definition of "relevant" unfortunately, more of a semantic one. Like, would monitoring time spent in the

cc

module be practically good enough? Feels so-so https://www.pantsbuild.org/docs/contributions-debugging#profiling-with-py-spy Idea here is that I want to test having the

cc

backend compile/link natively vs creating Ninja build files, and then having ninja perform the compilation/linking step to see whether that step is a bottleneck (when compared with the other Pants overhead - digests, daemon, dep inference, cache checking, etc).

You should be able to download a Zipkin trace of a build from Toolchain Buildsense.

If I just compile one of my random projects?

the project would have to be uploading telemetry to buildsense

👍 Need to make a decent OSS app I guess

There has been value in having

example-FOO

repositories for the other backends. I imagine that an

example-cc

repo will help now with testing and later with adoption.

For sure, just in my case, it needs to be BIIIIG to test. I think a kitchen sink repo would be great, everything in one monorepo

yeah big repos help out. When we were developing the Java backend, we ended up using the Google Guice repository to test out how well dependency inference was working. Having a non-trivial (but not too complex) repository as a test case was very useful.

LLVM or Chrome 🙂

LLVM would be quite a test case. :) I’d suggest limiting the size of the initial test case though since we tried with Java and Scala to use larger repos, but found that repo-specific integration issues could end up dominating the amount of time needed to make it a viable test case.