Backtracing, Invalidated Bugs and Quality

[Emmet Hikory]

Scott Kitterman wrote:
> By marking incomplete backtrace crash bugs invalid we lose information both
> about circumstances of crashes and frequency.  A bug with 50 dupes and one
> good backtrace is different than one with no dupes.  Reading the dictionary
> definition of 'invalid', I don't think it's correct.  These are real bugs
> that we choose to hide.

    It is also worth noting that no configurable time is likely
correct for all packages.  For packages that get a lot of attention
and updates, bugs that cannot be reproduced within a relatively short
time, yet for which we have good debug traces may be very difficult to
fix, even where someone can construct changes to the version of the
code that crashed from the debug information.  For packages that do
not get as much attention, and may not be updated for years (we still
have many packages not updated since the first import into Ubuntu,
prior to Warty Warthog), a solution may be easy to develop, or the bug
may be trivially reproducible, but this awaits some developer having
sufficient interest to fix the bug, regardless of the age of the bug.

-- 
Emmet HIKORY