ubuntu-bugcontrol team mailing list archive

[Alberto Salvia Novella <es20490446e@xxxxxxxxx>]

Daniel Letzeisen:
> some 'New' bugs can go directly to triaged

The question is what happens when they can't, what is usual: they stay 
shorted in the list on a pile till someone confirms them, making the 
triaging difficult.


Daniel Letzeisen:
> such as requests for packaging/upgrading

Since these are written by developers themselves, we can just ask them 
to set these as confirmed.


Daniel Letzeisen:
> or bugs which have an obvious cause (like a proprietary driver not
> building on a new kernel)

These are confirmed automatically when apport detects it is affecting 
more than one user, aren't they?

If you look through building failures in both nvidia and amd proprietary 
drivers, all of them fall into two categories: they haven't been touched 
for three years; or they have just been reported, no more than a day ago:

- amd: <http://tinyurl.com/p36e7sn>
- nvidia: <http://tinyurl.com/pk3blvr>


Daniel Letzeisen:
> users may mistakenly confirm similar bugs and some bugs are
> automatically confirmed by a bot(s) when only one person experiences
> them, so they're not necessarily reproducible.

Yes, but they will be a minority compared with those shorted in the list 
of new bugs; where most of them are already not fixable, and you need to 
test them by hand every time for realizing that.

Triaging is a straightforward process where you ask for the information 
a manual says it is required. In the worst scenario where the bug cannot 
be reproduced consistently, if all the triaging process has been 
performed, the bug is meant to be set as triaged or invalid; as 
implicitly said in:

 - <https://wiki.ubuntu.com/Bugs/Triage#Confirming>
 - <https://wiki.ubuntu.com/Bugs/Triage#Invalidating>


Daniel Letzeisen:
> I don't personally believe that a list of 'Confirmed' bugs has value
> much greater than looking through 'New' bugs.

My personal experience of being triaging hundreds of bugs for Ubuntu, 
one by one as listed, is new bugs rarely get translated into a fix, 
while confirmed ones do frequently.


Daniel Letzeisen:
> looking through very recent bugs has more value to me, since they're
> less likely to already be fixed upstream.

Agree. This is why I have been looking through bugs that only affect the 
current and future releases only, since I have estimated that nearly all 
from the rest are just End Of Life.

If Launchpad could treat End Of Life bugs automatically I think it will 
be a great success.


Regards.