filter-repo: fix pruning of empty commits with blob callbacks

Blob callbacks, either implicit (via e.g. --replace-text) or explicit,
can modify blobs in ways that make them match other blobs, which in turn
can result in some commits becoming empty.  We need to detect such cases
and ensure we prune these empty commits when --prune-empty=auto.

Reported-by: John Gietzen <[email protected]>
Signed-off-by: Elijah Newren <[email protected]>

Author: newren

git-filter-repo

@@ -3348,11 +3348,14 @@ class RepoFilter(object):
       self._filter_files(commit)
 
     # Find out which files were modified by the callbacks.  Such paths could
-    # lead to sebsequent commits being empty (e.g. if removed a line containing
+    # lead to subsequent commits being empty (e.g. if removing a line containing
     # a password from every version of a file that had the password, and some
     # later commit did nothing more than remove that line)
     final_file_changes = set(commit.file_changes)
-    differences = orig_file_changes.symmetric_difference(final_file_changes)
+    if self._args.replace_text or self._blob_callback:
+      differences = orig_file_changes.union(final_file_changes)
+    else:
+      differences = orig_file_changes.symmetric_difference(final_file_changes)
     self._files_tweaked.update(x.filename for x in differences)
 
     # Call the user-defined callback, if any

t/t9390-filter-repo.sh

@@ -1372,6 +1372,27 @@ test_expect_success 'degenerate merge with non-matching filenames' '
 	)
 '
 
+test_expect_success 'Filtering a blob to make it match previous version' '
+	test_create_repo remove_unique_bits_of_blob &&
+	(
+		cd remove_unique_bits_of_blob &&
+
+		test_write_lines foo baz >metasyntactic_names &&
+		git add metasyntactic_names &&
+		git commit -m init &&
+
+		test_write_lines foo bar baz >metasyntactic_names &&
+		git add metasyntactic_names &&
+		git commit -m second &&
+
+		git filter-repo --force --blob-callback "blob.data = blob.data.replace(b\"\\nbar\", b\"\")"
+
+		echo 1 >expect &&
+		git rev-list --count HEAD >actual &&
+		test_cmp expect actual
+	)
+'
+
 test_expect_success 'tweaking just a tag' '
 	test_create_repo tweaking_just_a_tag &&
 	(