I recently noticed a few entries (see below) where it appears we are silently catching an error related to GetFileSizePath().
Does anybody know why it’s being silently eaten and/or why the reason it’s silent isn’t commented in the code?
The first one is suppressed because it happens during the size-scanning.
Any errors here do not affect the backup, but just gives wrong estimates on the size/filecount for the progress output.
The second one is suppressed because we take the length of the actual data being read (if the file changes during read, there is no point in using the stale length). The value read here is only used for the database lookup to check if the filesize has changed. The value -1 means “invalid” and will always trigger a re-scan, thus the effect of that call failing will only force a re-scan of the file.
1 Like