libreoffice team mailing list archive

[Brian Murray <brian@xxxxxxxxxx>]

On Thu, Dec 27, 2012 at 08:05:06PM -0600, Christopher M. Penalver wrote:
> Martin Pitt/LibreOffice Packaging/Ubuntu Bug Control:
> 
> Hello. I am e-mailing you after battling to triage
> https://bugs.launchpad.net/bugs/1074834 , to get your input on the
> following issue. The original reporter attached a crash file from
> Quantal with LO 3.6.2~rc2-0ubuntu4:
> https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1074834/+attachment/3469304/+files/1074834_1-3.6.2%7Erc2-0ubuntu4.crash.zip
> 
> that when he attempts to file using ubuntu-bug:
> ubuntu-bug _usr_lib_libreoffice_program_soffice.bin.1000.crash
> 
> it notes this is a duplicate of
> https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1062900 .
> However, bug 1062900 is Expired, Private (unviewable to the Original
> Reporter of bug 1074834), and was filed against an outdated version of
> LO (3.6.2~rc2-0ubuntu1).
> 
> Hence, it seems prudent that Apport Retracing Service?! does not mark
> new LO crash reports duplicates of Expired bug reports, and/or ones
> with outdated versions of LO. This is not the first time I've seen new
> LO crash reports dup'ed to Expired LO reports.

The goal of marking bugs duplicate of other bugs is to prevent there
from being thousands of bug reports about the same issue, if we do not
automatically mark duplicates (or prevent reporting as in your specific
example) there would be lots of bug reports that we would need to
manually duplicate or that would be left open when one of the crashes
was fixed.

Addressing the individual issues you've brought up:

The master bug was Expired - when a new duplicate comes or a reporter is
pointed to a master bug with this status the hope is that they will read
the comment asking for more information and provide that information.

The master bug is Private - part of this issue is tracked in
http://launchpad.net/bugs/764414.  Barring a fix for that, what we as
bug control members should do is try to make all private bugs public
(provided they can be) so that people do not run into this situation.

The master bug is about an outdated version of LibreOffice - apport has
determined that the crash address signature is the same for these two
crashes which means that the crashes are the same regardless of the
package version.

You can learn a bit more about how client side duplicate detection,
which is what the reporter of the duplicate bug encountered, at:

http://www.piware.de/2011/11/apport-1-90-client-side-duplicate-checking/

--
Brian Murray

Attachment: signature.asc
Description: Digital signature