apport sends bug reports for situations which are not bugs

Bug #94822 reported by Matthias Klose
2
Affects Status Importance Assigned to Milestone
apport (Ubuntu)
Won't Fix
Undecided
Unassigned

Bug Description

Binary package hint: apport

see bug 89547:

 - it is a simple python backtrace. why is it necessary to send Procmaps
   and Procstatus? IMO it's only misleading. Are there siturations, where
   this information has any value? If not, please don't add it.

- the problem type is labeled "Crash", although the exception is
  OperationalError. If you think that crash reports are necessary
  for unhandled interpreter exceptions, then please analyze the
  exception and only file a report for error conditions. Wrong bug
  reports are not better than a probably missing report.

Revision history for this message
Martin Pitt (pitti) wrote :

> - it is a simple python backtrace. why is it necessary to send Procmaps
> and Procstatus? IMO it's only misleading. Are there siturations, where
> this information has any value? If not, please don't add it.

They are very important for crashes of binary programs, and they will teach the triager about custom libraries, special plugins, etc. It's cheap, it does not need a lot of space, you can ignore it, so I'll just keep it.

> - the problem type is labeled "Crash", although the exception is
> OperationalError. If you think that crash reports are necessary
> for unhandled interpreter exceptions, then please analyze the
> exception and only file a report for error conditions. Wrong bug
> reports are not better than a probably missing report.

OperationalError is not a standard Python exception that we could generally ignore, and even in this case it is not clear to me whether this is a real program error (e. g. in this case it could be a failed macro expansion, or whatever) or an error from the user. However, usually programs should fail more gracefully than spitting an exception into the user's eyes.

How do you propose should this be analyzed merely by looking at the stack trace?

Changed in apport:
status: Unconfirmed → Needs Info
Revision history for this message
Martin Pitt (pitti) wrote :

We are closing this bug report as it lacks the information, described in the previous comments, we need to investigate the problem further. However, please reopen it if you can give us the missing information and don't hesitate to submit bug reports in the future.

Changed in apport:
status: Incomplete → Invalid
Revision history for this message
Matthias Klose (doko) wrote :

reopening. closing as "invalid" sounds wrong; if you don't have a better idea then having a mapping of exceptions to error types, and if that will not be implemented for some reason, then please close as "wontfix".

Changed in apport:
status: Invalid → New
Revision history for this message
Martin Pitt (pitti) wrote :

It is not possible to generally determine whether an exception is a program bug or an user error, and the majority of unhandled exceptions are actual bugs. If programs really want to ignore certain exceptions for themselves, they could ship an apport hook, but preferably they should just be fixed to give a proper error message to the user.

Changed in apport:
status: New → Won't Fix
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.