Time to sleep and conjure up some answers!
Huh.
Looks like that was easier than anticipated for the most part. In fact... I might be able to keep it all in one process afterall.
Time to sleep and conjure up some answers!
Huh.
Looks like that was easier than anticipated for the most part. In fact... I might be able to keep it all in one process afterall.
All that whining and complaining last night, and after a good night's sleep...
3 files changed, 7 insertions(+), 15 deletions(-
Two files and two insertions are in Cargo.toml and Cargo.lock, so the real change is 5 insertions, 15 deletions, in one file.
But now, a new problem dropped:
> thread 'main' (8588) panicked at /build/cargo-vendor-dir/boring-sys-4.21.1/build/main.rs:672:37:
> called `Result::unwrap()` on an `Err` value: Os { code: 2, kind: NotFound, message: "No such file or directory" }
> note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
That is very helpful, thank you very much.
Said line is...
ensure_patches_applied(&config).unwrap();
Wonderful.
Added git to the buildInputs and all seems to be well now. Why does that work? Because the boring-sys crate applies patches to the source, and apparently, that requires git.
It does a git apply -v --whitespace=fix. Feels like patch would be enough here, but what do I know.
I wonder how fast this thing is now... since the Nix build produces a prod build. I can benchmark that!
bombardiering the static site server directly: ~37k req/sec.
...bombardiering https fails, because it tries to set up the challenge server? again? huh.