Fix decode strategy when splitting a chunk

[arthurstomp]

This PR fixes the remainig failing test in CI.

The encode function creates its 2x bytes output by merging 2 vectors of x bytes intercalating the values by if the index is even or odd. However, in the decode function splits the same 2 x bytes vector by just cutting it in half. That difference in strategy on splitting those larger vector seems to be the cause of the failing test.

This PR adds new function nibble_chunk that splits a larger vector in 2 smaller vectors using the even/odd strategy from the encode function.

[arthurstomp]

@cfcosta, is this failure of the build being caused by a formatting issue?

Below is the log i get from nix-store when i try nix flake check.

WARN no formatter for path: LICENSE-APACHE
traversed 4 files
emitted 2 files for processing
formatted 2 files (0 changed) in 24ms
2025/02/13 20:43:58 INFO using config file: /nix/store/rpyryr563jjm9li5x8avljprd4haqpkm-treefmt.toml
ERRO file has changed path=src/lib.rs prev_size=7139 prev_mod_time="1970-01-01 00:00:01 +0000 UTC" current_size=7141 current_mod_time="2025-02-13 20:43:58 +0000 UTC"
traversed 1 files
emitted 1 files for processing
formatted 1 files (1 changed) in 30ms
Error: unexpected changes detected, --fail-on-change is enabled

diff --git a/src/lib.rs b/src/lib.rs
index c74f202..264fbe1 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -145,19 +145,18 @@ fn decode_hex_nibble(n: u8x16) -> Result<u8x16, Error> {
 #[inline]
 fn nibble_chunck(chunk: &[u8]) -> (u8x16, u8x16) {
     let parsed_chunk = u8x32::from_slice(chunk);
-    let mut high_bytes_vec: Vec<u8> = vec![];
-    let mut low_bytes_vec: Vec<u8> = vec![];
-    for (index,piece) in parsed_chunk.to_array().iter().enumerate() {
+    let mut high_bytes_vec: Vec<u8> = vec![];
+    let mut low_bytes_vec: Vec<u8> = vec![];
+    for (index, piece) in parsed_chunk.to_array().iter().enumerate() {
         if index % 2 == 0 {
             high_bytes_vec.push(piece.clone())
         } else {
             low_bytes_vec.push(piece.clone())
         }
-
     }
     (
-      u8x16::from_slice(&high_bytes_vec),
-      u8x16::from_slice(&low_bytes_vec),
+        u8x16::from_slice(&high_bytes_vec),
+        u8x16::from_slice(&low_bytes_vec),
     )
 }

[arthurstomp]

Test are passing, though

[cfcosta]

very nice! just a little nitpicks.

[cfcosta]

nibble_chunk

[cfcosta]

Instead of using vec to allocate on the heap, can we just set [u8; X] directly? that way we don't need to allocate memory.