dmitrio added a comment.
If I understand correctly what has happened, without this patch the file at destination path existed after the operation was cancelled, and it might even not be corrupted. Still I believe that it is the **source** file that may remain after the cancellation, but not the one existed at the destination path. Try the same test with different files (I mean, with different content) at the source and destination paths. After cancelling an overwrite operation you will have either a copy of the source file or a corrupted file if the source file was sufficiently large (more than 1 GB works in my case, but this probably depends on memory size and disk I/O speed). I agree though that disappearing of the destination file may be confusing to user, so some special handling of this case may be useful. But currently I do not see any sign of this case being already handled by KIO, neither in the source code nor in real tests. REPOSITORY R241 KIO REVISION DETAIL https://phabricator.kde.org/D10663 To: dmitrio, #frameworks, dfaure, ngraham Cc: elvisangelaccio, ngraham, anthonyfieroni, meven, #frameworks, michaelh, bruns
