Check more information in python backtraces to identify non-duplicates

[mzidek-gh]

Two Python backtraces with the following 'name/messsage' fields were reported as duplicates:

dasbus.error.DBusError: set_user_ssh_key: user tomas does not exis

dasbus.error.DBusError: mount failed: 32

But they are different errors. The backtrace looks almost the same, but the "message" and "exception" fields in the report are different. Maybe we can use them to detect these reports as non-duplicates.

Reference bug in anaconda:
https://bugzilla.redhat.com/show_bug.cgi?id=1812727

[poncovka]

Hi, at this moment, there are about 13 duplicates or detected similar issues of the bug 1812727, but none of them is related to the original bug. Can we do something on the Anaconda's side to stop it?

[ernestask]

It’s entirely the product of the weird threading code in Anaconda.

[poncovka]

The problem is not caused by threads, but processes. Anaconda was split to something like a back-end (represented by DBus services) and a front-end (text and graphical UI). These bugs are caused by back-end errors that are propagated to the front-end. Therefore, the tracebacks start with a DBus call and are not "complete", because we are not able to propagate the first part of the traceback to the UI. I don't think that Anaconda is the only application that works this way.