PR #192 (79d755cd) moved rustls, quinn, rcgen and tokio-rustls to aws-lc-rs, but two comments in crates/punktfunk-host/src/identity.rs still credited ring:
The module doc credited the P-256 key to "generated by ring via rcgen". rcgen now selects aws_lc_rs — crates/punktfunk-host/Cargo.toml:135 and crates/punktfunk-core/Cargo.toml:80.
The legacy-RSA fallback claimed "rustls/ring can SERVE an existing RSA cert". The substance still holds under aws-lc-rs; only the provider name was wrong.
4903c9d3 fixed the generate() doc but missed the module doc, whose phrase wraps across two lines (generated by / ring via rcgen), so a line-based grep never matched it.
Why it mattered
The stale comment is load-bearing for anyone reasoning about the crypto backend: it was read as evidence that the P-256 identity is signed by ring, which also affects how you reason about ECDSA nonce generation. Under aws-lc the nonce is hardened with SHA-512(private key ‖ message digest) fed into the RBG as additional data (aws-lc/crypto/fipsmodule/ecdsa/ecdsa.c:280) — the same hedged property, but a different library to go read.
Scope
Comment-only. Every changed line in the diff is a comment, so compilation is unaffected by construction; cargo fmt -p punktfunk-host -- --check is clean and all lines stay within the file's 100-column budget. No behaviour change, no algorithm change — the identity remains ECDSA P-256.
PR #192 (`79d755cd`) moved rustls, quinn, rcgen and tokio-rustls to aws-lc-rs, but two comments in `crates/punktfunk-host/src/identity.rs` still credited `ring`:
- **The module doc** credited the P-256 key to *"generated by ring via rcgen"*. rcgen now selects `aws_lc_rs` — `crates/punktfunk-host/Cargo.toml:135` and `crates/punktfunk-core/Cargo.toml:80`.
- **The legacy-RSA fallback** claimed *"rustls/ring can SERVE an existing RSA cert"*. The substance still holds under aws-lc-rs; only the provider name was wrong.
`4903c9d3` fixed the `generate()` doc but missed the module doc, whose phrase wraps across two lines (`generated by` / `ring via rcgen`), so a line-based grep never matched it.
### Why it mattered
The stale comment is load-bearing for anyone reasoning about the crypto backend: it was read as evidence that the P-256 identity is signed by ring, which also affects how you reason about ECDSA nonce generation. Under aws-lc the nonce is hardened with `SHA-512(private key ‖ message digest)` fed into the RBG as additional data (`aws-lc/crypto/fipsmodule/ecdsa/ecdsa.c:280`) — the same hedged property, but a different library to go read.
### Scope
Comment-only. Every changed line in the diff is a comment, so compilation is unaffected by construction; `cargo fmt -p punktfunk-host -- --check` is clean and all lines stay within the file's 100-column budget. No behaviour change, no algorithm change — the identity remains ECDSA P-256.
PR #192 (79d755cd) moved rustls, quinn, rcgen and tokio-rustls to aws-lc-rs,
but two comments in identity.rs still credited ring:
* the module doc credited the P-256 key to "ring via rcgen" — rcgen now
selects `aws_lc_rs` (punktfunk-host/Cargo.toml:135, punktfunk-core/Cargo.toml:80);
* the legacy-RSA fallback claimed "rustls/ring can SERVE an existing RSA cert".
The substance still holds under aws-lc-rs; only the provider name was wrong.
4903c9d3 fixed the `generate()` doc but missed the module doc, whose phrase wraps
across two lines ("generated by" / "ring via rcgen"), so a line-based grep never
matched it.
Comment-only: every changed line is a comment, cargo fmt clean.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
PR #192 (
79d755cd) moved rustls, quinn, rcgen and tokio-rustls to aws-lc-rs, but two comments incrates/punktfunk-host/src/identity.rsstill creditedring:aws_lc_rs—crates/punktfunk-host/Cargo.toml:135andcrates/punktfunk-core/Cargo.toml:80.4903c9d3fixed thegenerate()doc but missed the module doc, whose phrase wraps across two lines (generated by/ring via rcgen), so a line-based grep never matched it.Why it mattered
The stale comment is load-bearing for anyone reasoning about the crypto backend: it was read as evidence that the P-256 identity is signed by ring, which also affects how you reason about ECDSA nonce generation. Under aws-lc the nonce is hardened with
SHA-512(private key ‖ message digest)fed into the RBG as additional data (aws-lc/crypto/fipsmodule/ecdsa/ecdsa.c:280) — the same hedged property, but a different library to go read.Scope
Comment-only. Every changed line in the diff is a comment, so compilation is unaffected by construction;
cargo fmt -p punktfunk-host -- --checkis clean and all lines stay within the file's 100-column budget. No behaviour change, no algorithm change — the identity remains ECDSA P-256.