feat: Implement newline normalization.

This should ensure consistent Artifact IDs whether on Windows or
Unix systems.

Signed-off-by: Andrew Lilley Brinker <[email protected]>

gitoid/src/gitoid.rs

@@ -552,6 +552,28 @@ impl<R: BufRead> ForEachChunk for R {
     }
 }
 
+fn digest_with_normalized_newlines<D>(buf: &[u8], digester: &mut D)
+where
+    D: Digest,
+{
+    let not_dos_newline = |char1: &u8, char2: &u8| ((*char1, *char2) == (b'\r', b'\n')).not();
+
+    let chunks = buf.chunk_by(not_dos_newline);
+
+    for chunk in chunks {
+        if chunk.len() == buf.len() {
+            digester.update(chunk);
+            break;
+        }
+
+        if chunk.last() == Some(&b'\r') {
+            digester.update(&chunk[0..(chunk.len() - 1)]);
+        } else {
+            digester.update(chunk)
+        }
+    }
+}
+
 #[cfg(feature = "std")]
 /// Helper function which actually applies the [`GitOid`] construction rules.
 ///
@@ -572,7 +594,8 @@ where
         O::NAME.as_bytes(),
         expected_read_length
     ));
-    let amount_read = BufReader::new(reader).for_each_chunk(|b| digester.update(b))?;
+    let amount_read = BufReader::new(reader)
+        .for_each_chunk(|b| digest_with_normalized_newlines(b, &mut digester))?;
     let hash = digester.finalize();
     Ok((hash, amount_read))
 }
@@ -599,7 +622,7 @@ where
 
     // It's in memory, so we know the exact size up front.
     let amount_read = reader.len();
-    digester.update(reader);
+    digest_with_normalized_newlines(reader, &mut digester);
     let hash = digester.finalize();
     Ok((hash, amount_read))
 }

omnibor-cli/tests/snapshots/test__artifact_id_json.snap

@@ -13,6 +13,6 @@ info:
 success: true
 exit_code: 0
 ----- stdout -----
-{"id":"<GITOID>","path":"tests/data/main.c"}
+{"id":"gitoid:blob:sha256:93561f4501717b4c4a2f3eb5776f03231d32ec2a1f709a611ad3d8dcf931dc1b","path":"tests/data/main.c"}
 
 ----- stderr -----

omnibor-cli/tests/snapshots/test__artifact_id_plain.snap

@@ -13,6 +13,6 @@ info:
 success: true
 exit_code: 0
 ----- stdout -----
-tests/data/main.c => <GITOID>
+tests/data/main.c => gitoid:blob:sha256:93561f4501717b4c4a2f3eb5776f03231d32ec2a1f709a611ad3d8dcf931dc1b
 
 ----- stderr -----

omnibor-cli/tests/snapshots/test__artifact_id_short.snap

@@ -13,6 +13,6 @@ info:
 success: true
 exit_code: 0
 ----- stdout -----
-<GITOID>
+gitoid:blob:sha256:93561f4501717b4c4a2f3eb5776f03231d32ec2a1f709a611ad3d8dcf931dc1b
 
 ----- stderr -----

omnibor-cli/tests/test.rs

@@ -6,7 +6,6 @@ macro_rules! settings {
     ($block:expr) => {
         let mut settings = Settings::clone_current();
         settings.add_filter(r#"omnibor(?:\.exe)?"#, "omnibor");
-        settings.add_filter(r#"gitoid:blob:sha256:[a-f0-9]{64}"#, "<GITOID>");
         settings.bind(|| $block);
     };
 }