[QA] re-login hangs with 'no connection configured'

[jnweiger]

Pre-submission Checks

• [X] I checked for similar issues, but could not find any. I also checked the closed issues. I could not contribute additional information to any existing issue.
• [X] I will take the time to fill in all the required fields. I know that the bug report may be dismissed otherwise due to lack of information.

Describe the bug

Seen with 2.10.1-rc.1 (and earlier)

• start a temporary server with oauth2
• Connect client
• Cuit client
• Shut down server
• Restart client (on the next day)
• The account is "signed out" (aka logged out). OK

---

• Click account -> log in
• The message 'No connection configured`is shown. BAD
• Often this message only flashes briefly and gets replaced by Obtaining authorization from browser, but sometimes it hangs. BAD.

---

• The account menu then offers a log out entry. This is odd, we are still logged out.

Expected behaviour

• an error message is shown, when the account -> log in action fails
• The message 'No connection configured`is not shown. Connection information is obviously there, as the log in succeeds on the next attempt. Workaorund
• Click account -> log out, the status message switches back to "signed out"
• Click account -> log in, the login succeeds (hopefully, never saw the failure twice in a row)

Expected behavior

.

Steps to reproduce the issue

.

Screenshots

.

Logs

.

Client version number

.

Desktop environment (Linux only)

.

Client package version and origin (Linux only)

.

Installation path (Windows only)

.

Server information

.

Additional context

.

[TheOneRing]

I don't expect to see that issue with 3.0 and its not a regression in 2.10 its been around for a while. Logging out and back in solves the issue.

[github-actions[bot]]

This issue was marked stale because it has been open for 30 days with no activity. Remove the stale label or comment or this will be closed in 7 days.

[github-actions[bot]]

The issue was marked as stale for 7 days and closed automatically.

[TheOneRing]

@michaelstingl have you seen the issue with a up to date 3.0?

[michaelstingl]

@michaelstingl have you seen the issue with a up to date 3.0?

Yes, I see it reliably at the first restart after a new connection was configured on macOS 12.5.1. @fmoc you had a look at it. what did you find out?

[TheOneRing]

Interesting as I thought we removed the "no connection configured" message

[michaelstingl]

@TheOneRing I was able to reproduce pretty reliably: https://cloud.owncloud.com/index.php/f/5877873 (Screen recording, logs and config)

testpilotcloud 3.0.0.8504-daily20220909 14f97d
Libraries Qt 5.15.5, OpenSSL 1.1.1q 5 Jul 2022
Using virtual files plugin: suffix
osx-21.6.0

[TheOneRing]

Could you retest now that we got the dynamic registration fix merged?

[saw-jan]

This issue still exists and sometimes the connection is not configured within 60 seconds (says the GUI test)

Reports: Report 1, Report 2

[TheOneRing]

@saw-jan could you provide the client logs for this?

[saw-jan]

This is the squish server log of a failing scenario due to this issue: serverlog.log

Full squish server log is here https://cache.owncloud.com/public/owncloud/client/13580/guiReportUpload/serverlog.log

[HanaGemela]

@saw-jan could you please provide log file with the error, screen recording or reliable steps to recreate?

[TheOneRing]

01:18:22:478 AUT stdout: 11-06 01:18:21:652 [ info gui.account.state ]:	AccountState connection status change:  OCC::ConnectionValidator::Undefined -> OCC::ConnectionValidator::CredentialsNotReady
01:18:22:478 AUT stdout: 11-06 01:18:21:652 [ warning sync.credentials.http.legacy ]:	Migrating old keychain entries failed "The name org.freedesktop.secrets was not provided by any .service files"

So I guess its a regression on the tests. This was fixed before.

[saw-jan]

This is the squish server log of a failing scenario due to this issue: serverlog.log

Full squish server log is here https://cache.owncloud.com/public/owncloud/client/13580/guiReportUpload/serverlog.log

These logs were from the GUI tests run against the oC10 server.

could you please provide log file with the error, screen recording or reliable steps to recreate?

I could only grab the logs from the test run and recently I haven't seen this failure (but if fails then this might be due to the reason Hanna mentioned above). Also, I couldn't recreate this issue manually both with oC10 and oCIS servers.

But there is a switch from No connection configured to another message during re-login. With oC10: No connection configured to Connecting to ...

oc10.webm

With oCIS: No connection configured to Reauthorization required.

ocis.webm

[individual-it]

So if the issue is not reproducible anymore I will close it and put it to "done", if it comes back we will have to reopen it and work on it again