Success message after failing to restore files

[fresheneesz]

• I have searched open and closed issues for duplicates.

---

Environment info

• Duplicati version: 2.0.4.23_beta_2019-07-14
• Operating system: Windows 10
• Backend: local

Description

When restoring files, if the drive containing the backup is not connected, the restore finishes with a success message, even tho no files were actually restored. A separate warning pop up appears that says the backup doesn't exist, but since errors like that happen from time to time, I associate them with backup failures, and it wasn't clear that it was related to the restore failure.

Steps to reproduce

1. Backup something to a drive
2. Disconnect that drive
3. Attempt to restore from that backup

• Actual result:

Duplicati fails to restore any files and yet gives a success message.

• Expected result:

No success message should appear, and instead a failure message should appear.

[fresheneesz]

It seems like there are other situations where the messages given to the user in the main page are not accurate and preempted by a pop-up error. The error system really needs to be revamped to be more integrated with the actual UI. There should never be conflicting messages like "success" in the body of the page and "error" as a popup, or "validating restored files..." in one place and an "error" pop up that actually indicates that validation is no longer happening.

[seantempleton]

Looks like this is ultimately caused by the Runner not rethrowing exceptions when tasks are run by the WorkerThread using commands that were added to the WorkerThread's queue because fromQueue is set to true.

This affects several operations performed by the server: restore, vacuum, verify, compact, repair and others.

Not throwing exceptions is clearly intentional which you can also see from the commit history (20f12f6) where fromQueue used to be called throwEx and was set to false.

[duplicatibot]

This issue has been mentioned on Duplicati. There might be relevant details there:

https://forum.duplicati.com/t/gpg-asymmetric-decryption-should-fail-but-does-not/17536/10