Merge branch 'error-handling' into 'main'

handle already existing remote files on move

Closes #4

See merge request !3

processor.py

@@ -45,7 +45,7 @@ import tempfile
 
 try:
     import webdav3.client as wc
-    from webdav3.exceptions import RemoteResourceNotFound
+    from webdav3.exceptions import RemoteResourceNotFound, ResponseErrorCode
 except ImportError:
     print(
         "cannot find webdav.client, try `apt install python3-webdavclient` or `pip install webdavclient`",
@@ -372,7 +372,14 @@ class ProcessingClient(wc.Client):
             if not self.dryrun:
                 self.mkdir(folder + "/dangerzone")
                 self.mkdir(folder + "/dangerzone/processing")
-                self.move(remote_path_from=folder + "/" + path, remote_path_to=remote_processing_path)
+                try:
+                    self.move(remote_path_from=folder + "/" + path, remote_path_to=remote_processing_path)
+                except ResponseErrorCode as e:
+                    # https://datatracker.ietf.org/doc/html/rfc7232#section-4.2
+                    # actually used in WebDAV to show the file already exists
+                    if e.code == 412:
+                        logging.warning("file already being processed, skipping")
+                        return
             with tempfile.TemporaryDirectory() as tmpdir:
                 # TODO: sanitize path for local use
                 local_path = tmpdir + "/danger/" + path
@@ -403,10 +410,23 @@ class ProcessingClient(wc.Client):
                     )
                     if not self.dryrun:
                         self.mkdir(folder + "/dangerzone/rejected")
-                        self.move(
-                            remote_path_from=remote_processing_path,
-                            remote_path_to=remote_rejected_path,
-                        )
+                        try:
+                            self.move(
+                                remote_path_from=remote_processing_path,
+                                remote_path_to=remote_rejected_path,
+                            )
+                        except ResponseErrorCode as e:
+                            # https://datatracker.ietf.org/doc/html/rfc7232#section-4.2
+                            # actually used in WebDAV to show the file already exists
+                            if e.code == 412:
+                                # rejected already exists, fall back
+                                # to delete the processing version
+                                # altogether
+                                #
+                                # XXX: we actually lose data here
+                                # which isn't nice. maybe we should
+                                # find a unique filename instead?
+                                self.clean(remote_processing_path)
                     return
 
                 # 6. on success, upload the sanitized file to a safe/
@@ -427,10 +447,23 @@ class ProcessingClient(wc.Client):
                 )
                 if not self.dryrun:
                     self.mkdir(folder + "/dangerzone/processed")
-                    self.move(
-                        remote_path_from=remote_processing_path,
-                        remote_path_to=remote_processed_path,
-                    )
+                    try:
+                        self.move(
+                            remote_path_from=remote_processing_path,
+                            remote_path_to=remote_processed_path,
+                        )
+                    except ResponseErrorCode as e:
+                        # https://datatracker.ietf.org/doc/html/rfc7232#section-4.2
+                        # actually used in WebDAV to show the file already exists
+                        if e.code == 412:
+                            # rejected already exists, fall back
+                            # to delete the processing version
+                            # altogether
+                            #
+                            # XXX: we actually lose data here
+                            # which isn't nice. maybe we should
+                            # find a unique filename instead?
+                            self.clean(remote_processing_path)
 
 
 if __name__ == "__main__":