build: Add error_set and audit unicode-ident

I'm a big fan of `error_set` as a library and am prioritizing auditing
it so I can use it in more of my projects guilt-free. I started by
auditing `unicode-ident` because it has zero normal dependencies (only
dev-dependencies) and a reasonably small linecount.

I am not looking forward to auditing `syn`... I am shocked that none of
the registry members appear to have even tried to. I know David Tolnay
is well-respected and is responsible for quite a few foundational
crates, but I figured the sheer influence such crates have in the
ecosystem would have warranted SOMEONE auditing them, even if only as a
formality.

Signed-off-by: hashcatHitman <155700084+hashcatHitman@users.noreply.github.com>

Author: hashcatHitman

Cargo.lock

@@ -20,6 +20,9 @@ publish = false
 [lib]
 path = "src/lib/lib.rs"
 
+[dependencies]
+error_set = "=0.9.1"
+
 [profile.release]
 codegen-units = 1
 strip = true

Cargo.toml

@@ -1,4 +1,17 @@
 
 # cargo-vet audits file
 
-[audits]
+[[audits.unicode-ident]]
+who = "hashcatHitman <155700084+hashcatHitman@users.noreply.github.com>"
+criteria = "safe-to-deploy"
+version = "1.0.22"
+notes = """
+Safety invariants for the two `unsafe` blocks were verified locally using `kani`
+`v0.67.0`. The code doesn't appear to do any concerning filesystem access,
+network access, or code generation. No effort has been made on my part to verify
+the actual correctness of the implementation, but given that the bulk of it
+seems to be generated from a specification and the vast majority of the Rust
+ecosystem depends on this crate transitively, I'm going to give it the benefit
+of the doubt and say it is probably fine, especially since earlier versions have
+already been audited by several registry members such as Google and Mozilla.
+"""