feat!: Use [u8;4] as chunk id (#279)

This allows to remove the additional string to describe the ids, which
are usually ascii anyway.

git-chunk/src/file/decode.rs

@@ -13,7 +13,7 @@ mod error {
                 display("Sentinel value encountered while still processing chunks.")
             }
             MissingSentinelValue { actual: crate::Kind } {
-                display("Sentinel value wasn't found, saw {:#016x}", actual)
+                display("Sentinel value wasn't found, saw {:?}", std::str::from_utf8(actual.as_ref()).unwrap_or("<non-ascii>"))
             }
             ChunkSizeOutOfBounds { offset: crate::file::Offset, file_length: u64 } {
                 display("The chunk offset {} went past the file of length {} - was it truncated?", offset, file_length)
@@ -22,7 +22,7 @@ mod error {
                 display("All chunk offsets must be incrementing.")
             }
             DuplicateChunk(kind: crate::Kind) {
-                display("The chunk of kind {:#016x} was encountered more than once", kind)
+                display("The chunk of kind {:?} was encountered more than once", std::str::from_utf8(kind.as_ref()).unwrap_or("<non-ascii>"))
             }
             TocTooSmall { actual: usize, expected: usize } {
                 display("The table of contents would be {} bytes, but got only {}", expected, actual)
@@ -58,7 +58,7 @@ impl file::Index {
 
         for _ in 0..num_chunks {
             let (kind, offset) = toc_entry.split_at(4);
-            let kind = be_u32(kind);
+            let kind = to_kind(kind);
             if kind == crate::SENTINEL {
                 return Err(Error::EarlySentinelValue);
             }
@@ -93,7 +93,7 @@ impl file::Index {
             })
         }
 
-        let sentinel = be_u32(&toc_entry[..4]);
+        let sentinel = to_kind(&toc_entry[..4]);
         if sentinel != crate::SENTINEL {
             return Err(Error::MissingSentinelValue { actual: sentinel });
         }
@@ -102,8 +102,8 @@ impl file::Index {
     }
 }
 
-fn be_u32(data: &[u8]) -> u32 {
-    u32::from_be_bytes(data[..4].try_into().unwrap())
+fn to_kind(data: &[u8]) -> crate::Kind {
+    data[..4].try_into().unwrap()
 }
 
 fn be_u64(data: &[u8]) -> u64 {

git-chunk/src/file/index.rs

@@ -10,15 +10,14 @@ pub mod offset_by_kind {
     #[derive(Debug)]
     pub struct Error {
         pub kind: crate::Kind,
-        pub name: &'static str,
     }
 
     impl Display for Error {
         fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
             write!(
                 f,
-                "Chunk named {:?} (id = {}) was not found in chunk file index",
-                self.name, self.kind
+                "Chunk named {:?} was not found in chunk file index",
+                std::str::from_utf8(&self.kind).unwrap_or("<non-ascii>")
             )
         }
     }
@@ -61,25 +60,16 @@ impl Index {
     pub const EMPTY_SIZE: usize = Index::ENTRY_SIZE;
 
     /// Find a chunk of `kind` and return its offset into the data if found
-    pub fn offset_by_kind(
-        &self,
-        kind: crate::Kind,
-        name: &'static str,
-    ) -> Result<Range<crate::file::Offset>, offset_by_kind::Error> {
+    pub fn offset_by_kind(&self, kind: crate::Kind) -> Result<Range<crate::file::Offset>, offset_by_kind::Error> {
         self.chunks
             .iter()
             .find_map(|c| (c.kind == kind).then(|| c.offset.clone()))
-            .ok_or(offset_by_kind::Error { kind, name })
+            .ok_or(offset_by_kind::Error { kind })
     }
 
     /// Find a chunk of `kind` and return its data slice based on its offset.
-    pub fn data_by_kind<'a>(
-        &self,
-        data: &'a [u8],
-        kind: crate::Kind,
-        name: &'static str,
-    ) -> Result<&'a [u8], data_by_kind::Error> {
-        let offset = self.offset_by_kind(kind, name)?;
+    pub fn data_by_kind<'a>(&self, data: &'a [u8], kind: crate::Kind) -> Result<&'a [u8], data_by_kind::Error> {
+        let offset = self.offset_by_kind(kind)?;
         Ok(&data[crate::into_usize_range(offset).ok_or(data_by_kind::Error::FileTooLarge)?])
     }
 

git-chunk/src/lib.rs

@@ -7,11 +7,11 @@
 use std::convert::TryInto;
 use std::ops::Range;
 
-/// An identifier to describe the kind of chunk, unique within a chunk file.
-pub type Kind = u32;
+/// An identifier to describe the kind of chunk, unique within a chunk file, typically in ASCII
+pub type Kind = [u8; 4];
 
 /// A special value denoting the end of the chunk file table of contents.
-pub const SENTINEL: Kind = 0;
+pub const SENTINEL: Kind = [0u8; 4];
 
 /// Turn a u64 Range into a usize range safely, to make chunk ranges useful in memory mapped files.
 pub fn into_usize_range(Range { start, end }: Range<file::Offset>) -> Option<Range<usize>> {