Change Details

BUG: 383764 Remove a partially copied file if copyjob was cancelled in the middle of file copying. File is considered to be in the process of being copied after some data block is actually written (or, to be more precise, slotProcessedSize is called). This should help us avoid cleaning up by mistake files that existed before the operation. This also means that in some very rare occasions when user cancels copying in the very beginning of the operation cleaning up procedure may not work — which is, I believe, the best result we can obtain without more significant code changes. Unlike [[ https://phabricator.kde.org/D10635 | this proposal ]], my option does not include handling of the full disk or any other errors. In its current state it applies only to the case when job gets killed by user.

BUG: 383764 Remove a partially copied file if copyjob was cancelled in the middle of file copying. File is considered to be in the process of being copied after some data block is actually written (or, to be more precise, slotProcessedSize is called). This should help us avoid cleaning up by mistake files that existed before the operation. This also means that in some very rare occasions when user cancels copying in the very beginning of the operation cleaning up procedure may not work — which is, I believe, the best result we can obtain without more significant code changes. Unlike D10635, my option does not include handling of the full disk or any other errors. In its current state it applies only to the case when job gets killed by user.

BUG: 383764 Remove a partially copied file if copyjob was cancelled in the middle of file copying. File is considered to be in the process of being copied after some data block is actually written (or, to be more precise, slotProcessedSize is called). This should help us avoid cleaning up by mistake files that existed before the operation. This also means that in some very rare occasions when user cancels copying in the very beginning of the operation cleaning up procedure may not work — which is, I believe, the best result we can obtain without more significant code changes. Unlike [[ https://phabricator.kde.org/D10635 | this proposal ]], my option does not include handling of the full disk or any other errors. In its current state it applies only to the case when job gets killed by user.