Thomas Eizinger
Thomas Eizinger
I think it doesn't work because we are using the commandline option to specify the target directory: https://github.com/obi1kenobi/cargo-semver-checks/blob/97f8a9bdc671040542e150bfb602409492356f7b/src/rustdoc_cmd.rs#L78-L79
> I think it doesn't work because we are using the commandline option to specify the target directory: > > https://github.com/obi1kenobi/cargo-semver-checks/blob/97f8a9bdc671040542e150bfb602409492356f7b/src/rustdoc_cmd.rs#L78-L79 If I remove these two, I can set the...
It is a bit of a hack but with this patch, I can successfully reuse most of the build artifacts from my usual build: ```diff Index: src/rustdoc_cmd.rs IDEA additional info:...
With: - a complete debug build of the workspace - the above patch - a cached registry JSON doc A full run of `cargo semver-checks check-release --workspace` takes just over...
> A full run of `cargo semver-checks check-release --workspace` takes just over 3 minutes! This also works if I clean out all the `local-` directories, i.e. the mtime of the...
> * compiling both baseline and current, both with `--all-features` -- IIRC we weren't sure if the two compilations might overwrite something in the regular target dir. We didn't want...
> > A full run of `cargo semver-checks check-release --workspace` takes just over 3 minutes! > > If you have it handy, I'd love to see the timing data for...
> > A full run of `cargo semver-checks check-release --workspace` takes just over 3 minutes! > > If you have it handy, I'd love to see the timing data for...
Friendly bump. Unless I am mistaken, with 0.20 shipped, having to rebuild the current crate on every run is now the dominating factor for the CI runtime of `cargo semver-checks`.
What is necessary to push this forward? With unrelated improvements to our CI, we are now spending about 50% of our time in `cargo semver-checks` of the current crate: https://github.com/libp2p/rust-libp2p/actions/runs/4851553760/jobs/8645517496?pr=3715...