Add additional test for hasher (#6198) [ci fast]

Signed-off-by: Robrecht Cannoodt <rcannood@gmail.com>
Signed-off-by: Ben Sherman <bentshermann@gmail.com>
Co-authored-by: DriesSchaumont <5946712+DriesSchaumont@users.noreply.github.com>
Co-authored-by: Ben Sherman <bentshermann@gmail.com>

Author: 3 people

modules/nf-commons/src/main/nextflow/util/HashBuilder.java

@@ -346,12 +346,15 @@ public FileVisitResult visitFile(Path path, BasicFileAttributes attrs) throws IO
                     try {
                         // the file relative base
                         final String relPath = base.relativize(path).toString();
-                        // compute the file path hash and sum to the result hash
-                        // since the sum is commutative, the traverse order does not matter
-                        sumBytes(resultBytes, hashBytes(relPath, HashMode.STANDARD));
                         // the file content sha-256 checksum
                         final String sha256 = sha256Cache.get(path);
-                        sumBytes(resultBytes, hashBytes(sha256, HashMode.STANDARD));
+                        // compute the file path hash and sum to the result hash
+                        // since the sum is commutative, the traverse order does not matter
+                        // compute a hash of the (file path, file hash) pair.
+                        // since the sum is commutative, the resulting hash in `resultBytes` is invariant to the file traversal order.
+                        // however, the file path and file hash do need to be processed together,
+                        // otherwise this introduces an edge case with directories with similar contents with have the same sha (see nextflow-io/nextflow#6198)
+                        sumBytes(resultBytes, hashBytes(Map.entry(relPath, sha256), HashMode.STANDARD));
                         return FileVisitResult.CONTINUE;
                     }
                     catch (ExecutionException t) {

modules/nf-commons/src/test/nextflow/util/HashBuilderTest.groovy

@@ -131,4 +131,26 @@ class HashBuilderTest extends Specification {
         then:
         hash1.hash() == hash2.hash()
     }
+
+    def 'directories with same content but different structure should yield different hashes'() {
+        given:
+        def folder = TestHelper.createInMemTempDir()
+        folder.resolve('dir1').mkdir()
+        folder.resolve('dir2').mkdir()
+        and:
+        folder.resolve('dir1/foo').text = "I'm foo"
+        folder.resolve('dir1/bar').text = "I'm bar"
+        and:
+        // the content of these files is intentionally swapped
+        folder.resolve('dir2/foo').text = "I'm bar"
+        folder.resolve('dir2/bar').text = "I'm foo"
+
+        when:
+        def hash1 = HashBuilder.hashDirSha256(HashBuilder.defaultHasher(), folder.resolve('dir1'), folder.resolve('dir1'))
+        and:
+        def hash2 = HashBuilder.hashDirSha256(HashBuilder.defaultHasher(), folder.resolve('dir2'), folder.resolve('dir2'))
+
+        then:
+        hash1.hash() != hash2.hash()
+    }
 }