Document the onerror workaround with its removal condition

fsspec's AbstractFileSystem.mv() passed the typo'd "onerror" keyword
(instead of "on_error", which copy() consumes) until it was fixed in
fsspec 2026.6.0, so it leaks through copy(**kwargs) into cp_file on
older releases and must not reach the S3 API. Replace the stale TODO
with the actual mechanism and the fsspec version at which the pop can
be removed.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: laughingman7743

pyathena/filesystem/s3.py

@@ -1011,10 +1011,12 @@ def cp_file(
             to optimize performance for large files. The copy operation is
             performed entirely on the S3 service without data transfer.
         """
-        # TODO: Delete the value that seems to be a typo, onerror=false.
+        # fsspec < 2026.6.0: AbstractFileSystem.mv() passed the typo'd
+        # "onerror" keyword (instead of "on_error", which copy() consumes),
+        # so it leaked through copy(**kwargs) into cp_file and must not
+        # reach the S3 API. Remove this once the fsspec requirement is
+        # >= 2026.6.0, where mv() passes on_error correctly.
         # https://github.com/fsspec/filesystem_spec/commit/346a589fef9308550ffa3d0d510f2db67281bb05
-        # https://github.com/fsspec/filesystem_spec/blob/2024.10.0/fsspec/spec.py#L1185
-        # https://github.com/fsspec/filesystem_spec/blob/2024.10.0/fsspec/spec.py#L1077
         kwargs.pop("onerror", None)
         bucket1, key1, version_id1 = self.parse_path(path1)
         bucket2, key2, version_id2 = self.parse_path(path2)

pyathena/filesystem/s3_async.py

@@ -191,6 +191,8 @@ async def _delete_chunk(chunk: list[dict[str, Any]]) -> None:
 
     async def _cp_file(self, path1: str, path2: str, **kwargs) -> None:
         """Copy an S3 object, using async parallel multipart upload for large files."""
+        # fsspec < 2026.6.0 leaks the typo'd "onerror" keyword from mv();
+        # see S3FileSystem.cp_file.
         kwargs.pop("onerror", None)
         bucket1, key1, version_id1 = self.parse_path(path1)
         bucket2, key2, version_id2 = self.parse_path(path2)